Merge remote-tracking branch 'origin/gfx950' into andriy/lwpck-2682

0ef27d53 · Andriy Roshchenko · 6778c318 · 74a743e2 · 0ef27d53 · 0ef27d53
Commit 0ef27d53 authored Jan 21, 2025 by Andriy Roshchenko
20 changed files
--- a/library/include/ck/library/tensor_operation_instance/gpu/grouped_conv_bwd_weight/device_grouped_conv_bwd_weight_two_stage_xdl_instance.hpp
+++ b/library/include/ck/library/tensor_operation_instance/gpu/grouped_conv_bwd_weight/device_grouped_conv_bwd_weight_two_stage_xdl_instance.hpp
 // SPDX-License-Identifier: MIT
-// Copyright (c) 2024, Advanced Micro Devices, Inc. All rights reserved.
+// Copyright (c) 2024-2025, Advanced Micro Devices, Inc. All rights reserved.

 #include "ck/ck.hpp"
 #include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
@@ -81,6 +81,28 @@ using device_grouped_conv_bwd_weight_two_stage_nhwgc_xdl_c_shuffle_f16_instances
    // clang-format on
    >;

+template <ck::index_t NDimSpatial,
+          typename ALayout,
+          typename BLayout,
+          typename ELayout,
+          ConvolutionBackwardWeightSpecialization ConvSpec,
+          BlockGemmPipelineScheduler Scheduler,
+          BlockGemmPipelineVersion PipelineVersion>
+using device_grouped_conv_bwd_weight_two_stage_nhwgc_xdl_c_shuffle_f16_irregular_instances =
+    std::tuple<
+        // clang-format off
+        //#########################################|              Num| InLayout| WeiLayout| OutLayout| InData| WeiData| OutData| AccData|          In|         Wei|         Out|              ConvBackward| Block|  MPer|  NPer| K0Per| K1| MPer| NPer| MXdl| NXdl|  ABlockTransfer|   ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds|  BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds|    CShuffle|    CShuffle|   CBlockTransfer|  CBlockTransfer| BlockGemm| BlockGemm| NumGroups|
+        //#########################################|              Dim|         |          |          |   Type|    Type|    Type|    Type| Elementwise| Elementwise| Elementwise|                    Weight|  Size| Block| Block| Block|   |  XDL|  XDL|  Per|  Per|   ThreadCluster|    ThreadCluster| SrcAccessOrder|   SrcVectorDim|      SrcScalar|      DstScalar| AddExtraM|   ThreadCluster|  ThreadCluster| SrcAccessOrder|  SrcVectorDim|      SrcScalar|      DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave|   ClusterLengths| ScalarPerVector|  Pipeline|  Pipeline|   ToMerge|
+        //#########################################|          Spatial|         |          |          |       |        |        |        |   Operation|   Operation|   Operation|            Specialization|      |      |      |      |   |     |     | Wave| Wave| Lengths_K0_M_K1|     ArrangeOrder|               |               |      PerVector|   PerVector_K1|          | Lengths_K0_N_K1|   ArrangeOrder|               |              |      PerVector|   PerVector_K1|          |  PerShuffle|  PerShuffle| MBlock_MPerBlock|    NWaveNPerXdl| Scheduler|   Version|          |
+        //#########################################|                 |         |          |          |       |        |        |        |            |            |            |                          |      |      |      |      |   |     |     |     |     |                |                 |               |               |               |               |          |                |               |               |              |               |               |          |            |            | NBlock_NPerBlock|                |          |          |          |
+        DeviceGroupedConvBwdWeightTwoStage_Xdl_CShuffle< NDimSpatial,  ALayout,   BLayout,   ELayout,   F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,                  ConvSpec,    64,    48,    64,     32,   8,   16,   16,    3,    4,  S<4, 16, 1>,  S<2, 0, 1>,  S<2, 0, 1>,                  1,              3,              4,      false,  S<4, 16,  1>,  S<2, 0, 1>,  S<2, 0, 1>,                1,              4,              4,      false,           1,           1,   S<1, 8, 1, 8>,                  1, Scheduler, PipelineVersion, 1>,
+        DeviceGroupedConvBwdWeightTwoStage_Xdl_CShuffle< NDimSpatial,  ALayout,   BLayout,   ELayout,   F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,                  ConvSpec,    64,    64,    48,     32,   8,   16,   16,    4,    3,  S<4, 16, 1>,  S<2, 0, 1>,  S<2, 0, 1>,                  1,              4,              4,      false,  S<4, 16,  1>,  S<2, 0, 1>,  S<2, 0, 1>,                1,              3,              4,      false,           1,           1,   S<1, 8, 1, 8>,                  1, Scheduler, PipelineVersion, 1>,
+        DeviceGroupedConvBwdWeightTwoStage_Xdl_CShuffle< NDimSpatial,  ALayout,   BLayout,   ELayout,   F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,                  ConvSpec,    64,    64,    80,     32,   8,   16,   16,    4,    5,  S<4, 16, 1>,  S<2, 0, 1>,  S<2, 0, 1>,                  1,              4,              4,      false,  S<4, 16,  1>,  S<2, 0, 1>,  S<2, 0, 1>,                1,              5,              4,      false,           1,           1,   S<1, 8, 1, 8>,                  1, Scheduler, PipelineVersion, 1>,
+        DeviceGroupedConvBwdWeightTwoStage_Xdl_CShuffle< NDimSpatial,  ALayout,   BLayout,   ELayout,   F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,                  ConvSpec,    64,    64,   112,     32,   8,   16,   16,    4,    7,  S<4, 16, 1>,  S<2, 0, 1>,  S<2, 0, 1>,                  1,              4,              4,      false,  S<4, 16,  1>,  S<2, 0, 1>,  S<2, 0, 1>,                1,              7,              4,      false,           1,           1,   S<1, 8, 1, 8>,                  1, Scheduler, PipelineVersion, 1>,
+        DeviceGroupedConvBwdWeightTwoStage_Xdl_CShuffle< NDimSpatial,  ALayout,   BLayout,   ELayout,   F16,     F16,     F16,     F32, PassThrough, PassThrough, PassThrough,                  ConvSpec,    64,    64,   208,     32,   8,   16,   16,    4,   13,  S<4, 16, 1>,  S<2, 0, 1>,  S<2, 0, 1>,                  1,              4,              4,      false,  S<4, 16,  1>,  S<2, 0, 1>,  S<2, 0, 1>,                1,             13,              4,      false,           1,           1,   S<1, 8, 1, 8>,                  1, Scheduler, PipelineVersion, 1>
+        // clang-format on
+        >;
+
 template <ck::index_t NDimSpatial,
          typename ALayout,
          typename BLayout,
@@ -130,6 +152,28 @@ using device_grouped_conv_bwd_weight_two_stage_nhwgc_xdl_c_shuffle_bf16_instance
    // clang-format on
    >;

+template <ck::index_t NDimSpatial,
+          typename ALayout,
+          typename BLayout,
+          typename ELayout,
+          ConvolutionBackwardWeightSpecialization ConvSpec,
+          BlockGemmPipelineScheduler Scheduler,
+          BlockGemmPipelineVersion PipelineVersion>
+using device_grouped_conv_bwd_weight_two_stage_nhwgc_xdl_c_shuffle_bf16_irregular_instances =
+    std::tuple<
+        // clang-format off
+        //#########################################|              Num| InLayout| WeiLayout| OutLayout| InData| WeiData| OutData| AccData|          In|         Wei|         Out|              ConvBackward| Block|  MPer|  NPer| K0Per| K1| MPer| NPer| MXdl| NXdl|  ABlockTransfer|   ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds|  BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds|    CShuffle|    CShuffle|   CBlockTransfer|  CBlockTransfer| BlockGemm| BlockGemm| NumGroups|
+        //#########################################|              Dim|         |          |          |   Type|    Type|    Type|    Type| Elementwise| Elementwise| Elementwise|                    Weight|  Size| Block| Block| Block|   |  XDL|  XDL|  Per|  Per|   ThreadCluster|    ThreadCluster| SrcAccessOrder|   SrcVectorDim|      SrcScalar|      DstScalar| AddExtraM|   ThreadCluster|  ThreadCluster| SrcAccessOrder|  SrcVectorDim|      SrcScalar|      DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave|   ClusterLengths| ScalarPerVector|  Pipeline|  Pipeline|   ToMerge|
+        //#########################################|          Spatial|         |          |          |       |        |        |        |   Operation|   Operation|   Operation|            Specialization|      |      |      |      |   |     |     | Wave| Wave| Lengths_K0_M_K1|     ArrangeOrder|               |               |      PerVector|   PerVector_K1|          | Lengths_K0_N_K1|   ArrangeOrder|               |              |      PerVector|   PerVector_K1|          |  PerShuffle|  PerShuffle| MBlock_MPerBlock|    NWaveNPerXdl| Scheduler|   Version|          |
+        //#########################################|                 |         |          |          |       |        |        |        |            |            |            |                          |      |      |      |      |   |     |     |     |     |                |                 |               |               |               |               |          |                |               |               |              |               |               |          |            |            | NBlock_NPerBlock|                |          |          |          |
+        DeviceGroupedConvBwdWeightTwoStage_Xdl_CShuffle< NDimSpatial,  ALayout,   BLayout,   ELayout,   BF16,     BF16,     BF16,     F32, PassThrough, PassThrough, PassThrough,                  ConvSpec,    64,    48,    64,     32,   8,   16,   16,    3,    4,  S<4, 16, 1>,  S<2, 0, 1>,  S<2, 0, 1>,                  1,              3,              4,      false,  S<4, 16,  1>,  S<2, 0, 1>,  S<2, 0, 1>,                1,              4,              4,      false,           1,           1,   S<1, 8, 1, 8>,                  1, Scheduler, PipelineVersion, 1>,
+        DeviceGroupedConvBwdWeightTwoStage_Xdl_CShuffle< NDimSpatial,  ALayout,   BLayout,   ELayout,   BF16,     BF16,     BF16,     F32, PassThrough, PassThrough, PassThrough,                  ConvSpec,    64,    64,    48,     32,   8,   16,   16,    4,    3,  S<4, 16, 1>,  S<2, 0, 1>,  S<2, 0, 1>,                  1,              4,              4,      false,  S<4, 16,  1>,  S<2, 0, 1>,  S<2, 0, 1>,                1,              3,              4,      false,           1,           1,   S<1, 8, 1, 8>,                  1, Scheduler, PipelineVersion, 1>,
+        DeviceGroupedConvBwdWeightTwoStage_Xdl_CShuffle< NDimSpatial,  ALayout,   BLayout,   ELayout,   BF16,     BF16,     BF16,     F32, PassThrough, PassThrough, PassThrough,                  ConvSpec,    64,    64,    80,     32,   8,   16,   16,    4,    5,  S<4, 16, 1>,  S<2, 0, 1>,  S<2, 0, 1>,                  1,              4,              4,      false,  S<4, 16,  1>,  S<2, 0, 1>,  S<2, 0, 1>,                1,              5,              4,      false,           1,           1,   S<1, 8, 1, 8>,                  1, Scheduler, PipelineVersion, 1>,
+        DeviceGroupedConvBwdWeightTwoStage_Xdl_CShuffle< NDimSpatial,  ALayout,   BLayout,   ELayout,   BF16,     BF16,     BF16,     F32, PassThrough, PassThrough, PassThrough,                  ConvSpec,    64,    64,   112,     32,   8,   16,   16,    4,    7,  S<4, 16, 1>,  S<2, 0, 1>,  S<2, 0, 1>,                  1,              4,              4,      false,  S<4, 16,  1>,  S<2, 0, 1>,  S<2, 0, 1>,                1,              7,              4,      false,           1,           1,   S<1, 8, 1, 8>,                  1, Scheduler, PipelineVersion, 1>,
+        DeviceGroupedConvBwdWeightTwoStage_Xdl_CShuffle< NDimSpatial,  ALayout,   BLayout,   ELayout,   BF16,     BF16,     BF16,     F32, PassThrough, PassThrough, PassThrough,                  ConvSpec,    64,    64,   208,     32,   8,   16,   16,    4,   13,  S<4, 16, 1>,  S<2, 0, 1>,  S<2, 0, 1>,                  1,              4,              4,      false,  S<4, 16,  1>,  S<2, 0, 1>,  S<2, 0, 1>,                1,             13,              4,      false,           1,           1,   S<1, 8, 1, 8>,                  1, Scheduler, PipelineVersion, 1>
+        // clang-format on
+        >;
+
 template <ck::index_t NDimSpatial,
          typename ALayout,
          typename BLayout,

--- a/library/include/ck/library/tensor_operation_instance/gpu/grouped_convolution_backward_weight.hpp
+++ b/library/include/ck/library/tensor_operation_instance/gpu/grouped_convolution_backward_weight.hpp
 // SPDX-License-Identifier: MIT
-// Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
+// Copyright (c) 2018-2025, Advanced Micro Devices, Inc. All rights reserved.

 #pragma once

@@ -358,6 +358,10 @@ struct DeviceOperationInstanceFactory<ck::tensor_operation::device::DeviceGroupe
                        op_ptrs);
                    add_device_grouped_conv2d_bwd_weight_two_stage_xdl_nhwgc_gkyxc_nhwgk_f16_pipev5_instances(
                        op_ptrs);
+                    add_device_grouped_conv2d_bwd_weight_two_stage_xdl_nhwgc_gkyxc_nhwgk_f16_pipev2_irregular_instances(
+                        op_ptrs);
+                    add_device_grouped_conv2d_bwd_weight_two_stage_xdl_nhwgc_gkyxc_nhwgk_f16_pipev5_irregular_instances(
+                        op_ptrs);
                }
 #endif
 #ifdef CK_ENABLE_BF16
@@ -383,6 +387,10 @@ struct DeviceOperationInstanceFactory<ck::tensor_operation::device::DeviceGroupe
                        op_ptrs);
                    add_device_grouped_conv2d_bwd_weight_two_stage_xdl_nhwgc_gkyxc_nhwgk_bf16_pipev5_instances(
                        op_ptrs);
+                    add_device_grouped_conv2d_bwd_weight_two_stage_xdl_nhwgc_gkyxc_nhwgk_bf16_pipev2_irregular_instances(
+                        op_ptrs);
+                    add_device_grouped_conv2d_bwd_weight_two_stage_xdl_nhwgc_gkyxc_nhwgk_bf16_pipev5_irregular_instances(
+                        op_ptrs);
                }
 #endif
            }
@@ -478,6 +486,10 @@ struct DeviceOperationInstanceFactory<ck::tensor_operation::device::DeviceGroupe
                        op_ptrs);
                    add_device_grouped_conv3d_bwd_weight_two_stage_xdl_ndhwgc_gkzyxc_ndhwgk_f16_pipev5_instances(
                        op_ptrs);
+                    add_device_grouped_conv3d_bwd_weight_two_stage_xdl_ndhwgc_gkzyxc_ndhwgk_f16_pipev2_irregular_instances(
+                        op_ptrs);
+                    add_device_grouped_conv3d_bwd_weight_two_stage_xdl_ndhwgc_gkzyxc_ndhwgk_f16_pipev5_irregular_instances(
+                        op_ptrs);
                }
 #endif
 #ifdef CK_ENABLE_BF16
@@ -503,6 +515,10 @@ struct DeviceOperationInstanceFactory<ck::tensor_operation::device::DeviceGroupe
                        op_ptrs);
                    add_device_grouped_conv3d_bwd_weight_two_stage_xdl_ndhwgc_gkzyxc_ndhwgk_bf16_pipev5_instances(
                        op_ptrs);
+                    add_device_grouped_conv3d_bwd_weight_two_stage_xdl_ndhwgc_gkzyxc_ndhwgk_bf16_pipev2_irregular_instances(
+                        op_ptrs);
+                    add_device_grouped_conv3d_bwd_weight_two_stage_xdl_ndhwgc_gkzyxc_ndhwgk_bf16_pipev5_irregular_instances(
+                        op_ptrs);
                }
 #endif
 #if defined CK_ENABLE_FP16 && defined CK_ENABLE_FP8 && defined CK_ENABLE_BF8

--- a/library/include/ck/library/tensor_operation_instance/gpu/grouped_convolution_backward_weight_xdl.inc
+++ b/library/include/ck/library/tensor_operation_instance/gpu/grouped_convolution_backward_weight_xdl.inc
 // SPDX-License-Identifier: MIT
-// Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
+// Copyright (c) 2018-2025, Advanced Micro Devices, Inc. All rights reserved.

 #pragma once

@@ -149,6 +149,30 @@ void add_device_grouped_conv2d_bwd_weight_two_stage_xdl_nhwgc_gkyxc_nhwgk_bf16_p
                                                           PassThrough,
                                                           PassThrough>>>& instances);

+void add_device_grouped_conv2d_bwd_weight_two_stage_xdl_nhwgc_gkyxc_nhwgk_bf16_pipev2_irregular_instances(
+    std::vector<std::unique_ptr<DeviceGroupedConvBwdWeight<2,
+                                                           NHWGC,
+                                                           GKYXC,
+                                                           NHWGK,
+                                                           BF16,
+                                                           BF16,
+                                                           BF16,
+                                                           PassThrough,
+                                                           PassThrough,
+                                                           PassThrough>>>& instances);
+
+void add_device_grouped_conv2d_bwd_weight_two_stage_xdl_nhwgc_gkyxc_nhwgk_bf16_pipev5_irregular_instances(
+    std::vector<std::unique_ptr<DeviceGroupedConvBwdWeight<2,
+                                                           NHWGC,
+                                                           GKYXC,
+                                                           NHWGK,
+                                                           BF16,
+                                                           BF16,
+                                                           BF16,
+                                                           PassThrough,
+                                                           PassThrough,
+                                                           PassThrough>>>& instances);
+
 void add_device_grouped_conv2d_bwd_weight_two_stage_xdl_ngchw_gkyxc_ngkhw_bf16_pipev1_instances(
    std::vector<std::unique_ptr<DeviceGroupedConvBwdWeight<2,
                                                           NGCHW,
@@ -234,6 +258,30 @@ void add_device_grouped_conv2d_bwd_weight_two_stage_xdl_nhwgc_gkyxc_nhwgk_f16_pi
                                                           PassThrough,
                                                           PassThrough>>>& instances);

+void add_device_grouped_conv2d_bwd_weight_two_stage_xdl_nhwgc_gkyxc_nhwgk_f16_pipev2_irregular_instances(
+    std::vector<std::unique_ptr<DeviceGroupedConvBwdWeight<2,
+                                                           NHWGC,
+                                                           GKYXC,
+                                                           NHWGK,
+                                                           F16,
+                                                           F16,
+                                                           F16,
+                                                           PassThrough,
+                                                           PassThrough,
+                                                           PassThrough>>>& instances);
+
+void add_device_grouped_conv2d_bwd_weight_two_stage_xdl_nhwgc_gkyxc_nhwgk_f16_pipev5_irregular_instances(
+    std::vector<std::unique_ptr<DeviceGroupedConvBwdWeight<2,
+                                                           NHWGC,
+                                                           GKYXC,
+                                                           NHWGK,
+                                                           F16,
+                                                           F16,
+                                                           F16,
+                                                           PassThrough,
+                                                           PassThrough,
+                                                           PassThrough>>>& instances);
+
 void add_device_grouped_conv2d_bwd_weight_two_stage_xdl_ngchw_gkyxc_ngkhw_f16_pipev1_instances(
    std::vector<std::unique_ptr<DeviceGroupedConvBwdWeight<2,
                                                           NGCHW,
@@ -384,6 +432,30 @@ void add_device_grouped_conv3d_bwd_weight_two_stage_xdl_ndhwgc_gkzyxc_ndhwgk_bf1
                                                           PassThrough,
                                                           PassThrough>>>& instances);

+void add_device_grouped_conv3d_bwd_weight_two_stage_xdl_ndhwgc_gkzyxc_ndhwgk_bf16_pipev2_irregular_instances(
+    std::vector<std::unique_ptr<DeviceGroupedConvBwdWeight<3,
+                                                           NDHWGC,
+                                                           GKZYXC,
+                                                           NDHWGK,
+                                                           BF16,
+                                                           BF16,
+                                                           BF16,
+                                                           PassThrough,
+                                                           PassThrough,
+                                                           PassThrough>>>& instances);
+
+void add_device_grouped_conv3d_bwd_weight_two_stage_xdl_ndhwgc_gkzyxc_ndhwgk_bf16_pipev5_irregular_instances(
+    std::vector<std::unique_ptr<DeviceGroupedConvBwdWeight<3,
+                                                           NDHWGC,
+                                                           GKZYXC,
+                                                           NDHWGK,
+                                                           BF16,
+                                                           BF16,
+                                                           BF16,
+                                                           PassThrough,
+                                                           PassThrough,
+                                                           PassThrough>>>& instances);
+
 void add_device_grouped_conv3d_bwd_weight_two_stage_xdl_ngcdhw_gkzyxc_ngkdhw_bf16_pipev1_instances(
    std::vector<std::unique_ptr<DeviceGroupedConvBwdWeight<3,
                                                           NGCDHW,
@@ -469,6 +541,30 @@ void add_device_grouped_conv3d_bwd_weight_two_stage_xdl_ndhwgc_gkzyxc_ndhwgk_f16
                                                           PassThrough,
                                                           PassThrough>>>& instances);

+void add_device_grouped_conv3d_bwd_weight_two_stage_xdl_ndhwgc_gkzyxc_ndhwgk_f16_pipev2_irregular_instances(
+    std::vector<std::unique_ptr<DeviceGroupedConvBwdWeight<3,
+                                                           NDHWGC,
+                                                           GKZYXC,
+                                                           NDHWGK,
+                                                           F16,
+                                                           F16,
+                                                           F16,
+                                                           PassThrough,
+                                                           PassThrough,
+                                                           PassThrough>>>& instances);
+
+void add_device_grouped_conv3d_bwd_weight_two_stage_xdl_ndhwgc_gkzyxc_ndhwgk_f16_pipev5_irregular_instances(
+    std::vector<std::unique_ptr<DeviceGroupedConvBwdWeight<3,
+                                                           NDHWGC,
+                                                           GKZYXC,
+                                                           NDHWGK,
+                                                           F16,
+                                                           F16,
+                                                           F16,
+                                                           PassThrough,
+                                                           PassThrough,
+                                                           PassThrough>>>& instances);
+
 void add_device_grouped_conv3d_bwd_weight_two_stage_xdl_ngcdhw_gkzyxc_ngkdhw_f16_pipev1_instances(
    std::vector<std::unique_ptr<DeviceGroupedConvBwdWeight<3,
                                                           NGCDHW,

--- a/library/include/ck/library/tensor_operation_instance/gpu/grouped_convolution_forward.hpp
+++ b/library/include/ck/library/tensor_operation_instance/gpu/grouped_convolution_forward.hpp
@@ -304,7 +304,23 @@ struct DeviceOperationInstanceFactory<ck::tensor_operation::device::DeviceGroupe
                    op_ptrs);
            }
 #endif
-
+#ifdef CK_ENABLE_BF16
+            if constexpr(is_same_v<InDataType, ck::bhalf_t> &&
+                         is_same_v<WeiDataType, ck::bhalf_t> &&
+                         is_same_v<OutDataType, ck::bhalf_t> &&
+                         is_same_v<AComputeType, ck::bhalf_t> &&
+                         is_same_v<BComputeType, ck::bhalf_t>)
+            {
+                add_device_grouped_conv2d_fwd_xdl_merged_groups_ngchw_gkyxc_ngkhw_bf16_instances(
+                    op_ptrs);
+                add_device_grouped_conv2d_fwd_xdl_ngchw_gkyxc_ngkhw_bf16_instances(op_ptrs);
+                add_device_grouped_conv2d_fwd_xdl_ngchw_gkyxc_ngkhw_bf16_comp_instances(op_ptrs);
+                add_device_grouped_conv2d_fwd_xdl_ngchw_gkyxc_ngkhw_bf16_mem_intra_instances(
+                    op_ptrs);
+                add_device_grouped_conv2d_fwd_xdl_ngchw_gkyxc_ngkhw_bf16_mem_inter_instances(
+                    op_ptrs);
+            }
+#endif
 #ifdef CK_ENABLE_INT8
            if constexpr(is_same_v<InDataType, int8_t> && is_same_v<WeiDataType, int8_t> &&
                         is_same_v<OutDataType, int8_t> && is_same_v<AComputeType, int8_t> &&

--- a/library/include/ck/library/tensor_operation_instance/gpu/grouped_convolution_forward_comp_xdl.inc
+++ b/library/include/ck/library/tensor_operation_instance/gpu/grouped_convolution_forward_comp_xdl.inc
@@ -90,6 +90,22 @@ void add_device_grouped_conv2d_fwd_xdl_ngchw_gkyxc_ngkhw_f16_comp_instances(
                                                                PassThrough>>>& instances);
 #endif

+#ifdef CK_ENABLE_BF16
+void add_device_grouped_conv2d_fwd_xdl_ngchw_gkyxc_ngkhw_bf16_comp_instances(
+    std::vector<std::unique_ptr<DeviceGroupedConvFwdMultipleABD<2,
+                                                                NGCHW,
+                                                                GKYXC,
+                                                                Empty_Tuple,
+                                                                NGKHW,
+                                                                BF16,
+                                                                BF16,
+                                                                Empty_Tuple,
+                                                                BF16,
+                                                                PassThrough,
+                                                                PassThrough,
+                                                                PassThrough>>>& instances);
+#endif
+
 #ifdef CK_ENABLE_FP32
 void add_device_grouped_conv2d_fwd_xdl_ngchw_gkyxc_ngkhw_f32_comp_instances(
    std::vector<std::unique_ptr<DeviceGroupedConvFwdMultipleABD<2,

--- a/library/include/ck/library/tensor_operation_instance/gpu/grouped_convolution_forward_mem_inter_xdl.inc
+++ b/library/include/ck/library/tensor_operation_instance/gpu/grouped_convolution_forward_mem_inter_xdl.inc
@@ -90,6 +90,22 @@ void add_device_grouped_conv2d_fwd_xdl_ngchw_gkyxc_ngkhw_f16_mem_inter_instances
                                                                PassThrough>>>& instances);
 #endif

+#ifdef CK_ENABLE_BF16
+void add_device_grouped_conv2d_fwd_xdl_ngchw_gkyxc_ngkhw_bf16_mem_inter_instances(
+    std::vector<std::unique_ptr<DeviceGroupedConvFwdMultipleABD<2,
+                                                                NGCHW,
+                                                                GKYXC,
+                                                                Empty_Tuple,
+                                                                NGKHW,
+                                                                BF16,
+                                                                BF16,
+                                                                Empty_Tuple,
+                                                                BF16,
+                                                                PassThrough,
+                                                                PassThrough,
+                                                                PassThrough>>>& instances);
+#endif
+
 #ifdef CK_ENABLE_FP32
 void add_device_grouped_conv2d_fwd_xdl_ngchw_gkyxc_ngkhw_f32_mem_inter_instances(
    std::vector<std::unique_ptr<DeviceGroupedConvFwdMultipleABD<2,

--- a/library/include/ck/library/tensor_operation_instance/gpu/grouped_convolution_forward_mem_intra_xdl.inc
+++ b/library/include/ck/library/tensor_operation_instance/gpu/grouped_convolution_forward_mem_intra_xdl.inc
@@ -90,6 +90,22 @@ void add_device_grouped_conv2d_fwd_xdl_ngchw_gkyxc_ngkhw_f16_mem_intra_instances
                                                                PassThrough>>>& instances);
 #endif

+#ifdef CK_ENABLE_BF16
+void add_device_grouped_conv2d_fwd_xdl_ngchw_gkyxc_ngkhw_bf16_mem_intra_instances(
+    std::vector<std::unique_ptr<DeviceGroupedConvFwdMultipleABD<2,
+                                                                NGCHW,
+                                                                GKYXC,
+                                                                Empty_Tuple,
+                                                                NGKHW,
+                                                                BF16,
+                                                                BF16,
+                                                                Empty_Tuple,
+                                                                BF16,
+                                                                PassThrough,
+                                                                PassThrough,
+                                                                PassThrough>>>& instances);
+#endif
+
 #ifdef CK_ENABLE_FP32
 void add_device_grouped_conv2d_fwd_xdl_ngchw_gkyxc_ngkhw_f32_mem_intra_instances(
    std::vector<std::unique_ptr<DeviceGroupedConvFwdMultipleABD<2,

--- a/library/include/ck/library/tensor_operation_instance/gpu/grouped_convolution_forward_xdl.inc
+++ b/library/include/ck/library/tensor_operation_instance/gpu/grouped_convolution_forward_xdl.inc
@@ -204,6 +204,22 @@ void add_device_grouped_conv2d_fwd_xdl_ngchw_gkyxc_ngkhw_f16_instances(
                                                                PassThrough>>>& instances);
 #endif

+#ifdef CK_ENABLE_BF16
+void add_device_grouped_conv2d_fwd_xdl_ngchw_gkyxc_ngkhw_bf16_instances(
+    std::vector<std::unique_ptr<DeviceGroupedConvFwdMultipleABD<2,
+                                                                NGCHW,
+                                                                GKYXC,
+                                                                Empty_Tuple,
+                                                                NGKHW,
+                                                                BF16,
+                                                                BF16,
+                                                                Empty_Tuple,
+                                                                BF16,
+                                                                PassThrough,
+                                                                PassThrough,
+                                                                PassThrough>>>& instances);
+#endif
+
 #ifdef CK_ENABLE_FP32
 void add_device_grouped_conv2d_fwd_xdl_ngchw_gkyxc_ngkhw_f32_instances(
    std::vector<std::unique_ptr<DeviceGroupedConvFwdMultipleABD<2,

--- a/library/include/ck/library/tensor_operation_instance/gpu/grouped_convolution_forward_xdl_merged_groups.inc
+++ b/library/include/ck/library/tensor_operation_instance/gpu/grouped_convolution_forward_xdl_merged_groups.inc
@@ -23,6 +23,20 @@ void add_device_grouped_conv2d_fwd_xdl_merged_groups_nhwgc_gkyxc_nhwgk_bf16_inst
                                                                PassThrough,
                                                                PassThrough,
                                                                PassThrough>>>& instances);
+
+void add_device_grouped_conv2d_fwd_xdl_merged_groups_ngchw_gkyxc_ngkhw_bf16_instances(
+    std::vector<std::unique_ptr<DeviceGroupedConvFwdMultipleABD<2,
+                                                                NGCHW,
+                                                                GKYXC,
+                                                                Empty_Tuple,
+                                                                NGKHW,
+                                                                BF16,
+                                                                BF16,
+                                                                Empty_Tuple,
+                                                                BF16,
+                                                                PassThrough,
+                                                                PassThrough,
+                                                                PassThrough>>>& instances);
 #endif

 #ifdef CK_ENABLE_FP16

--- a/library/src/tensor_operation_instance/gpu/CMakeLists.txt
+++ b/library/src/tensor_operation_instance/gpu/CMakeLists.txt
@@ -39,6 +39,13 @@ function(add_instance_library INSTANCE_NAME)

    set(INST_TARGETS ${SUPPORTED_GPU_TARGETS})

+    # Do not build DPP instances if DPP_KERNELS macro is not set
+    foreach(source IN LISTS ARGN)
+        if(NOT DEFINED DPP_KERNELS AND source MATCHES "_dpp")
+            message("removing dpp instance ${source} ")
+            list(REMOVE_ITEM ARGN "${source}")
+        endif()
+    endforeach()
    # Do not build DL instances if DL_KERNELS macro is not set
    foreach(source IN LISTS ARGN)
        if(NOT DEFINED DL_KERNELS AND source MATCHES "_dl")
@@ -183,6 +190,10 @@ FOREACH(subdir_path ${dir_list})
            message("bf8 instance found!")
            set(add_inst 1)
        endif()
+        if(("${cmake_instance}" MATCHES "_bf16" OR "${cmake_instance}" MATCHES "_b16") AND DTYPES MATCHES "bf16")
+            message("bf16 instance found!")
+            set(add_inst 1)
+        endif()
        if(("${cmake_instance}" MATCHES "_fp16" OR "${cmake_instance}" MATCHES "_f16") AND DTYPES MATCHES "fp16")
            message("fp16 instance found!")
            set(add_inst 1)
@@ -195,10 +206,6 @@ FOREACH(subdir_path ${dir_list})
            message("fp64 instance found!")
            set(add_inst 1)
        endif()
-        if("${cmake_instance}" MATCHES "_bf16" AND DTYPES MATCHES "bf16")
-            message("bf16 instance found!")
-            set(add_inst 1)
-        endif()
        if(("${cmake_instance}" MATCHES "_int8" OR "${cmake_instance}" MATCHES "_i8") AND DTYPES MATCHES "int8")
            message("int8 instance found!")
            set(add_inst 1)

--- a/library/src/tensor_operation_instance/gpu/gemm_b_scale/CMakeLists.txt
+++ b/library/src/tensor_operation_instance/gpu/gemm_b_scale/CMakeLists.txt
+# ONLY XDL_KERNELS
+set(GEMM_B_SCALE_INSTANCES)
+
+list(APPEND GEMM_B_SCALE_INSTANCES 
+        device_gemm_b_scale_xdl_f16_i4_f16/device_gemm_b_scale_xdl_f16_i4_f16_mk_nk_mn_mem_v2_default_instance.cpp
+        )
+
+set_source_files_properties(device_gemm_b_scale_xdl_f16_i4_f16/device_gemm_b_scale_xdl_f16_i4_f16_mk_nk_mn_mem_v2_default_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1")
+
+add_instance_library(device_gemm_b_scale_instance ${GEMM_B_SCALE_INSTANCES})
\ No newline at end of file
--- a/library/src/tensor_operation_instance/gpu/gemm_b_scale/device_gemm_b_scale_xdl_f16_i4_f16/device_gemm_b_scale_xdl_f16_i4_f16_mk_nk_mn.hpp
+++ b/library/src/tensor_operation_instance/gpu/gemm_b_scale/device_gemm_b_scale_xdl_f16_i4_f16/device_gemm_b_scale_xdl_f16_i4_f16_mk_nk_mn.hpp
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
+
+#include "ck/ck.hpp"
+#include "ck/tensor_operation/gpu/device/gemm_specialization.hpp"
+#include "ck/tensor_operation/gpu/device/impl/device_gemm_xdl_cshuffle_v3_b_scale.hpp"
+#include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
+
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace instance {
+
+using I4  = pk_i4_t;
+using F16 = half_t;
+using F32 = float;
+
+using Row = tensor_layout::gemm::RowMajor;
+using Col = tensor_layout::gemm::ColumnMajor;
+
+template <index_t... Is>
+using S = Sequence<Is...>;
+
+using PassThrough = element_wise::PassThrough;
+
+static constexpr auto GemmDefault    = GemmSpecialization::Default;
+static constexpr auto GemmKPadding   = GemmSpecialization::KPadding;
+static constexpr auto GemmMNPadding  = GemmSpecialization::MNPadding;
+static constexpr auto GemmMNKPadding = GemmSpecialization::MNKPadding;
+
+static constexpr auto Intrawave = BlockGemmPipelineScheduler::Intrawave;
+static constexpr auto Interwave = BlockGemmPipelineScheduler::Interwave;
+
+#if 0
+template <GemmSpecialization GemmSpec>
+using device_gemm_xdl_b_scale_f16_i4_f16_mk_nk_mn_comp_instances = std::tuple<
+
+#endif
+
+template <BlockGemmPipelineScheduler BlkGemmPipeSched, GemmSpecialization GemmSpec>
+using device_gemm_b_scale_xdl_f16_i4_f16_mk_nk_mn_mem_instances = std::tuple<
+    // clang-format off
+        //#########################| ALayout| BLayout| CLayout|AData| BData| BScale| CData| AccData| Cshuffle|           A|           B|           C|          GEMM| Block| Scale| Scale|  MPer|  NPer|  KPer| AK1| BK1|MPer| NPer| MXdl| NXdl|  ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds|  BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds|    CShuffle|    CShuffle|     CBlockTransferClusterLengths|  CBlockTransfer|    Block-wiseGemm|               Block-wiseGemm|
+        //#########################|        |        |        | Type|  Type|   Data|  Type|    Type|     Type| Elementwise| Elementwise| Elementwise|Specialization|  Size| Block| Block| Block| Block| Block|    |    | XDL|  XDL|  Per|  Per|   ThreadCluster|  ThreadCluster| SrcAccessOrder|   SrcVectorDim|      SrcScalar|      DstScalar| AddExtraM|   ThreadCluster|  ThreadCluster| SrcAccessOrder|  SrcVectorDim|      SrcScalar|      DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MXdlPerWave_MWaveMPerXdl| ScalarPerVector|          Pipeline|                     Pipeline|
+        //#########################|        |        |        |     |      |   Type|      |        |         |   Operation|   Operation|   Operation|              |      |     N|     K|      |      |      |    |    |Wave| Wave|     |     | Lengths_K0_M_K1|   ArrangeOrder|               |               |      PerVector|   PerVector_K1|          | Lengths_K0_N_K1|   ArrangeOrder|               |              |      PerVector|   PerVector_K1|          |  PerShuffle|  PerShuffle| _NBlock_NXdlPerWave_NWaveNPerXdl|   _NWaveNPerXdl|         Scheduler|                     Verision|
+        //#########################|        |        |        |     |      |       |      |        |         |            |            |            |              |      |      |      |      |      |      |    |    |    |     |     |     |                |               |               |               |               |               |          |                |               |               |              |               |               |          |            |            |                                 |                |                  |                             |
+        
+        //Compute friendly
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   256,    1,   128,   128,  128,   128,   8,   32,  32,   32,    2,    2,     S<16, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,             2,              8,              8,          0,    S<4, 64, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             32,             32,          0,          1,           1,                   S<1, 32, 1, 8>,               8,  BlkGemmPipeSched, BlockGemmPipelineVersion::v3, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   256,    1,   128,   128,  128,    64,   8,   32,  32,   32,    2,    2,     S<8, 32, 1>,      S<1, 0, 2>,    S<1, 0, 2>,             2,              8,              8,          0,    S<2, 128, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             32,             32,          0,          1,           1,                   S<1, 32, 1, 8>,               8,  BlkGemmPipeSched, BlockGemmPipelineVersion::v4, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   256,    1,   128,   128,  128,   128,   8,   32,  32,   32,    2,    2,     S<16, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,             2,              8,              8,          0,    S<4, 64, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             32,             32,          0,          1,           1,                   S<1, 32, 1, 8>,               8,  BlkGemmPipeSched, BlockGemmPipelineVersion::v3, half_t, half_t, false, false>,
+ 
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   256,    1,   128,   128,  128,    64,   8,   32,  32,   32,    2,    2,     S<8, 32, 1>,      S<1, 0, 2>,    S<1, 0, 2>,             2,              8,              8,          0,    S<2, 128, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,               8,  BlkGemmPipeSched, BlockGemmPipelineVersion::v4, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   256,    1,   128,   128,  128,    64,   8,   32,  32,   32,    2,    2,     S<8, 32, 1>,      S<1, 0, 2>,    S<1, 0, 2>,             2,              8,              8,          0,    S<2, 128, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,               8,  BlkGemmPipeSched, BlockGemmPipelineVersion::v3, half_t, half_t, false, false>,
+        
+        //Latency friendly
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   128,    1,   128,    32,   16,   128,   8,   16,  16,   16,    1,    1,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,               2,  BlkGemmPipeSched, BlockGemmPipelineVersion::v3, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,    64,    1,   128,    16,   16,   128,   8,   16,  16,   16,    1,    1,     S<16, 4, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<8, 8, 1>,      S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 4>,               4,  BlkGemmPipeSched, BlockGemmPipelineVersion::v3, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,    64,    1,   128,    16,   16,   128,   8,   16,  16,   16,    1,    1,     S<8,  8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<8, 8, 1>,      S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 4>,               4,  BlkGemmPipeSched, BlockGemmPipelineVersion::v3, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   128,    1,   128,    16,   32,   128,   8,   32,  16,   16,    1,    1,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<4, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             32,             32,          0,          1,           1,                   S<1, 16, 1, 8>,               4,  BlkGemmPipeSched, BlockGemmPipelineVersion::v3, half_t, half_t, false, false>,
+        
+        // Memory friendly v3
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   128,    1,   128,   128,   32,   128,   8,   32,  32,   32,    2,    1,     S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<4, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             32,             32,          0,          1,           1,                   S<1, 16, 1, 8>,               4,  BlkGemmPipeSched, BlockGemmPipelineVersion::v3, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   128,    1,   128,   128,   16,   128,   8,   16,  16,   16,    4,    1,     S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,               2,  BlkGemmPipeSched, BlockGemmPipelineVersion::v3, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   128,    1,   128,    64,   32,   128,   8,   32,  32,   32,    1,    1,     S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<4, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             32,             32,          0,          1,           1,                   S<1, 16, 1, 8>,               4,  BlkGemmPipeSched, BlockGemmPipelineVersion::v3, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   128,    1,   128,    64,   16,   128,   8,   16,  16,   16,    2,    1,     S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,               2,  BlkGemmPipeSched, BlockGemmPipelineVersion::v3, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   128,    1,   128,    32,   16,   128,   8,   16,  16,   16,    1,    1,     S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,               2,  BlkGemmPipeSched, BlockGemmPipelineVersion::v3, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,    64,    1,   128,    16,   16,   128,   8,   16,  16,   16,    1,    1,     S<16, 4, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<8, 8, 1>,      S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 4>,               4,  BlkGemmPipeSched, BlockGemmPipelineVersion::v3, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,    64,    1,   128,    16,   16,   128,   8,   16,  16,   16,    1,    1,     S<16, 4, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<8, 8, 1>,      S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 4>,               4,  BlkGemmPipeSched, BlockGemmPipelineVersion::v3, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   128,    1,   128,    16,   32,   128,   8,   32,  16,   16,    1,    1,     S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<4, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             32,             32,          0,          1,           1,                   S<1, 16, 1, 8>,               4,  BlkGemmPipeSched, BlockGemmPipelineVersion::v3, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   128,    1,   128,    16,   64,   128,   8,   32,  16,   16,    1,    2,     S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<4, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             32,             32,          0,          1,           1,                   S<1, 16, 1, 8>,               4,  BlkGemmPipeSched, BlockGemmPipelineVersion::v3, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   128,    1,   128,    32,   64,   128,   8,   32,  32,   32,    1,    1,     S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<4, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             32,             32,          0,          1,           1,                   S<1, 16, 1, 8>,               8,  BlkGemmPipeSched, BlockGemmPipelineVersion::v3, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   128,    1,   128,    16,  128,   128,   8,   32,  16,   16,    1,    4,     S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<4, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             32,             32,          0,          1,           1,                   S<1, 16, 1, 8>,               4,  BlkGemmPipeSched, BlockGemmPipelineVersion::v3, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   128,    1,   128,    32,  128,   128,   8,   32,  32,   32,    1,    2,     S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<4, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             32,             32,          0,          1,           1,                   S<1, 16, 1, 8>,               8,  BlkGemmPipeSched, BlockGemmPipelineVersion::v3, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   256,    1,   128,    16,  256,   128,   8,   32,  16,   16,    1,    4,     S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<4, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             32,             32,          0,          1,           1,                   S<1, 16, 1, 16>,              4,  BlkGemmPipeSched, BlockGemmPipelineVersion::v3, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   256,    1,   128,    32,  256,   128,   8,   32,  32,   32,    1,    2,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<4, 64, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             32,             32,          0,          1,           1,                   S<1, 16, 1, 16>,              8,  BlkGemmPipeSched, BlockGemmPipelineVersion::v3, half_t, half_t, false, false>,
+
+        // Memory friendly v4
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   128,    1,   128,    64,   32,   128,   8,   32,  32,   32,    1,    1,     S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<4, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             32,             32,          0,          1,           1,                   S<1, 16, 1, 8>,               4,  BlkGemmPipeSched, BlockGemmPipelineVersion::v4, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   128,    1,   128,    64,   16,   128,   8,   16,  16,   16,    2,    1,     S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,               2,  BlkGemmPipeSched, BlockGemmPipelineVersion::v4, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   128,    1,   128,    32,   16,   128,   8,   16,  16,   16,    1,    1,     S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,               2,  BlkGemmPipeSched, BlockGemmPipelineVersion::v4, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,    64,    1,   128,    16,   16,   128,   8,   16,  16,   16,    1,    1,     S<16, 4, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<8, 8, 1>,      S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 4>,               4,  BlkGemmPipeSched, BlockGemmPipelineVersion::v4, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,    64,    1,   128,    16,   16,   128,   8,   16,  16,   16,    1,    1,     S<16, 4, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<8, 8, 1>,      S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 4>,               4,  BlkGemmPipeSched, BlockGemmPipelineVersion::v4, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   128,    1,   128,    16,   32,   128,   8,   32,  16,   16,    1,    1,     S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<4, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             32,             32,          0,          1,           1,                   S<1, 16, 1, 8>,               4,  BlkGemmPipeSched, BlockGemmPipelineVersion::v4, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   128,    1,   128,    16,   64,   128,   8,   32,  16,   16,    1,    2,     S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<4, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             32,             32,          0,          1,           1,                   S<1, 16, 1, 8>,               4,  BlkGemmPipeSched, BlockGemmPipelineVersion::v4, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   128,    1,   128,    32,   64,   128,   8,   32,  32,   32,    1,    1,     S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<4, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             32,             32,          0,          1,           1,                   S<1, 16, 1, 8>,               8,  BlkGemmPipeSched, BlockGemmPipelineVersion::v4, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   128,    1,   128,    16,  128,   128,   8,   32,  16,   16,    1,    4,     S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<4, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             32,             32,          0,          1,           1,                   S<1, 16, 1, 8>,               4,  BlkGemmPipeSched, BlockGemmPipelineVersion::v4, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   128,    1,   128,    32,  128,   128,   8,   32,  32,   32,    1,    2,     S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<4, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             32,             32,          0,          1,           1,                   S<1, 16, 1, 8>,               8,  BlkGemmPipeSched, BlockGemmPipelineVersion::v4, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   256,    1,   128,    16,  256,   128,   8,   32,  16,   16,    1,    4,     S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<4, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             32,             32,          0,          1,           1,                   S<1, 16, 1, 16>,              4,  BlkGemmPipeSched, BlockGemmPipelineVersion::v4, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   256,    1,   128,    32,  256,   128,   8,   32,  32,   32,    1,    2,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<4, 64, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             32,             32,          0,          1,           1,                   S<1, 16, 1, 16>,              8,  BlkGemmPipeSched, BlockGemmPipelineVersion::v4, half_t, half_t, false, false>,
+
+        //new Compute friendly kernel
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   256,    1,   128,   128,   128,   64,   8,   32,  32,   32,    2,    2,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<2, 128, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             32,             32,          0,          1,           1,                   S<1, 32, 1, 8>,               8,  BlkGemmPipeSched, BlockGemmPipelineVersion::v3, half_t, half_t, false, false>,
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   256,    1,   128,   128,   128,   64,   8,   32,  32,   32,    4,    1,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<2, 128, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             32,             32,          0,          1,           1,                   S<1, 32, 1, 8>,               8,  BlkGemmPipeSched, BlockGemmPipelineVersion::v3, half_t, half_t, false, false>,
+
+        //new Memory friendly kernel
+        DeviceGemm_Xdl_CShuffleV3<  Row,     Col,     Row,     F16,    I4,   F16,   F16,   F32,     F16,      PassThrough, PassThrough, PassThrough,       GemmSpec,   256,    1,   128,   16,    64,   256,   8,   32,  16,   16,    1,    1,     S<32, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,              2,              8,              8,          0,    S<8, 32, 1>,      S<1, 0, 2>,    S<1, 0, 2>,               2,             32,             32,          0,          1,           1,                   S<1, 16, 1, 8>,               8,  BlkGemmPipeSched, BlockGemmPipelineVersion::v3, half_t, half_t, false, false>
+    // clang-format on
+    >;
+} // namespace instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck
--- a/library/src/tensor_operation_instance/gpu/gemm_b_scale/device_gemm_b_scale_xdl_f16_i4_f16/device_gemm_b_scale_xdl_f16_i4_f16_mk_nk_mn_mem_v2_default_instance.cpp
+++ b/library/src/tensor_operation_instance/gpu/gemm_b_scale/device_gemm_b_scale_xdl_f16_i4_f16/device_gemm_b_scale_xdl_f16_i4_f16_mk_nk_mn_mem_v2_default_instance.cpp
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
+
+#include "device_gemm_b_scale_xdl_f16_i4_f16_mk_nk_mn.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace instance {
+void add_device_gemm_b_scale_xdl_f16_i4_f16_mk_nk_mn_mem_v2_default_instances(
+    std::vector<std::unique_ptr<DeviceGemmV2BScale<Row,
+                                                   Col,
+                                                   Row,
+                                                   F16,
+                                                   I4,
+                                                   F16,
+                                                   F16,
+                                                   1,
+                                                   128,
+                                                   PassThrough,
+                                                   PassThrough,
+                                                   PassThrough>>>& instances)
+{
+    add_device_operation_instances(
+        instances,
+        device_gemm_b_scale_xdl_f16_i4_f16_mk_nk_mn_mem_instances<Intrawave, GemmDefault>{});
+}
+
+} // namespace instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck
--- a/library/src/tensor_operation_instance/gpu/gemm_multiply_multiply/CMakeLists.txt
+++ b/library/src/tensor_operation_instance/gpu/gemm_multiply_multiply/CMakeLists.txt
@@ -9,6 +9,13 @@ list(APPEND GEMM_MULTIPLY_MULTIPLY_INSTANCES
        device_gemm_multiply_multiply_xdl_f8_f8_bf16/device_gemm_multiply_multiply_xdl_f8_f8_bf16_mk_nk_mn_mem_v2_default_instance.cpp
        device_gemm_multiply_multiply_xdl_f8_f8_bf16/device_gemm_multiply_multiply_xdl_f8_f8_bf16_mk_nk_mn_mem_v2_kpadding_instance.cpp

+        device_gemm_multiply_multiply_xdl_f8_f8_f16/device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn_comp_default_instance.cpp
+        device_gemm_multiply_multiply_xdl_f8_f8_f16/device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn_comp_kpadding_instance.cpp
+        device_gemm_multiply_multiply_xdl_f8_f8_f16/device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn_mem_v1_default_instance.cpp
+        device_gemm_multiply_multiply_xdl_f8_f8_f16/device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn_mem_v1_kpadding_instance.cpp
+        device_gemm_multiply_multiply_xdl_f8_f8_f16/device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn_mem_v2_default_instance.cpp
+        device_gemm_multiply_multiply_xdl_f8_f8_f16/device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn_mem_v2_kpadding_instance.cpp
+
        device_gemm_multiply_multiply_xdl_i8_i8_bf16/device_gemm_multiply_multiply_xdl_i8_i8_bf16_mk_nk_mn_comp_default_instance.cpp
        device_gemm_multiply_multiply_xdl_i8_i8_bf16/device_gemm_multiply_multiply_xdl_i8_i8_bf16_mk_nk_mn_comp_kpadding_instance.cpp
        device_gemm_multiply_multiply_xdl_i8_i8_bf16/device_gemm_multiply_multiply_xdl_i8_i8_bf16_mk_nk_mn_mem_v1_default_instance.cpp
@@ -20,6 +27,9 @@ list(APPEND GEMM_MULTIPLY_MULTIPLY_INSTANCES
 set_source_files_properties(device_gemm_multiply_multiply_xdl_f8_f8_bf16/device_gemm_multiply_multiply_xdl_f8_f8_bf16_mk_nk_mn_comp_default_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1")
 set_source_files_properties(device_gemm_multiply_multiply_xdl_f8_f8_bf16/device_gemm_multiply_multiply_xdl_f8_f8_bf16_mk_nk_mn_comp_kpadding_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1")

+set_source_files_properties(device_gemm_multiply_multiply_xdl_f8_f8_f16/device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn_comp_default_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1")
+set_source_files_properties(device_gemm_multiply_multiply_xdl_f8_f8_f16/device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn_comp_kpadding_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1")
+
 set_source_files_properties(device_gemm_multiply_multiply_xdl_i8_i8_bf16/device_gemm_multiply_multiply_xdl_i8_i8_bf16_mk_nk_mn_comp_default_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1")
 set_source_files_properties(device_gemm_multiply_multiply_xdl_i8_i8_bf16/device_gemm_multiply_multiply_xdl_i8_i8_bf16_mk_nk_mn_comp_kpadding_instance.cpp PROPERTIES COMPILE_OPTIONS ";-mllvm;-greedy-reverse-local-assignment=1")


--- a/library/src/tensor_operation_instance/gpu/gemm_multiply_multiply/device_gemm_multiply_multiply_xdl_f8_f8_bf16/device_gemm_multiply_multiply_xdl_f8_f8_bf16_mk_nk_mn.hpp
+++ b/library/src/tensor_operation_instance/gpu/gemm_multiply_multiply/device_gemm_multiply_multiply_xdl_f8_f8_bf16/device_gemm_multiply_multiply_xdl_f8_f8_bf16_mk_nk_mn.hpp
@@ -46,8 +46,49 @@ using device_gemm_multiply_multiply_xdl_f8_f8_bf16_mk_nk_mn_comp_instances = std
        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,   256,    64,  16,  16,  32,   32,    4,    4,     S<4, 64, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<4, 64, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v4, F8>,
        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   128,   128,   128,  16,  16,  32,   32,    2,    2,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v4, F8>,
        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   128,   128,    64,  16,  16,  32,   32,    2,    2,     S<4, 64, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<4, 64, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v4, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,   256,    64,  16,  16,  32,   32,    4,    4,     S<4, 64, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<4, 64, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v4, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,   256,   128,  16,  16,  32,   32,    4,    4,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,   224,   128,  16,  16,  32,   32,    2,    7,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          2,           1,                   S<1, 64, 1, 4>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,   192,   128,  16,  16,  32,   32,    4,    3,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,   160,   128,  16,  16,  32,   32,    2,    5,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          2,           1,                   S<1, 64, 1, 4>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,   128,   128,  16,  16,  32,   32,    4,    2,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,    96,   128,  16,  16,  32,   32,    2,    3,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          2,           1,                   S<1, 64, 1, 4>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,    64,   128,  16,  16,  32,   32,    4,    1,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   128,   256,   128,  16,  16,  32,   32,    2,    4,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   128,   224,   128,  16,  16,  32,   32,    1,    7,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 64, 1, 4>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   128,   192,   128,  16,  16,  32,   32,    2,    3,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   128,   160,   128,  16,  16,  32,   32,    1,    5,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 64, 1, 4>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   128,   128,   256,  16,  16,  32,   32,    2,    2,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   128,   128,   128,  16,  16,  32,   32,    2,    2,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   128,    96,   256,  16,  16,  32,   32,    1,    3,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 64, 1, 4>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        // Compute friendly
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    64,   256,   128,  16,  16,  32,   32,    1,    4,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    64,   224,   128,  16,  16,  16,   16,    2,    7,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          2,           1,                   S<1, 64, 1, 4>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    64,   192,   256,  16,  16,  32,   32,    1,    3,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    64,   192,   128,  16,  16,  32,   32,    1,    3,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    64,   160,   256,  16,  16,  16,   16,    2,    5,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          2,           1,                   S<1, 64, 1, 4>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    64,   128,   256,  16,  16,  32,   32,    1,    2,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    64,    96,   256,  16,  16,  16,   16,    2,    3,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          2,           1,                   S<1, 64, 1, 4>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    64,    64,   512,  16,  16,  32,   32,    1,    1,     S<32, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<32, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    32,   256,   128,  16,  16,  32,   32,    1,    2,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    32,   224,   256,  16,  16,  16,   16,    1,    7,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<4, 4, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    32,   192,   256,  16,  16,  16,   16,    1,    6,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           2,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    32,   160,   256,  16,  16,  16,   16,    1,    5,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<4, 4, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    32,   128,   256,  16,  16,  32,   32,    1,    1,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    32,    96,   256,  16,  16,  16,   16,    1,    3,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<4, 4, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    32,    64,   512,  16,  16,  16,   16,    1,    2,     S<32, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<32, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           2,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    16,   256,   128,   8,  16,  16,   16,    1,    4,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,              8,              8,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           2,                   S<1, 16, 1, 16>,    S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    16,   192,   256,  16,  16,  16,   16,    1,    3,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 16>,    S<4, 4, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    16,   128,   256,  16,  16,  16,   16,    1,    2,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           2,                   S<1, 16, 1, 16>,    S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    16,    64,   512,  16,  16,  16,   16,    1,    1,     S<32, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<32, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 16>,    S<4, 4, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,   256,   128,  16,  16,  16,   16,    8,    8,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           2,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,   256,    64,  16,  16,  16,   16,    8,    8,     S<4, 64, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<4, 64, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           2,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,   192,   128,  16,  16,  16,   16,    8,    6,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           2,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,   160,   128,  16,  16,  16,   16,    8,    5,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          2,           1,                   S<1, 64, 1, 4>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,   128,   128,  16,  16,  16,   16,    8,    4,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           2,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,    96,   128,  16,  16,  16,   16,    8,    3,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          2,           1,                   S<1, 64, 1, 4>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,    64,   128,  16,  16,  16,   16,    8,    2,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           2,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   224,   256,    128, 16,  16,  16,   16,    7,    8,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           2,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,   224,    128, 16,  16,  16,   16,    8,    7,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          2,           1,                   S<1, 64, 1, 4>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, BF16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   128,   128,    128, 16,  16,  32,   32,    2,    2,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
@@ -74,23 +115,35 @@ using device_gemm_multiply_multiply_xdl_f8_f8_bf16_mk_nk_mn_mem_instances = std:
        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    32,   16,    128, 16,  16,  16,   16,    1,    1,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,      S<2, 2, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v1, F8>,
        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,    64,    16,   16,    128, 16,  16,  16,   16,    1,    1,     S<8,  8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8,  8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 4>,      S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v1, F8>,
        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    16,   32,    128, 16,  16,  16,   16,    1,    1,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,      S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v1, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,    64,   16,    256, 16,  16,  16,   16,    1,    1,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 64, 1, 4>,      S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v1, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    32,   16,    256, 16,  16,  16,   16,    1,    1,     S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 4>,      S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v1, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,    64,    16,   16,    256, 16,  16,  16,   16,    1,    1,     S<16, 4, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 4, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 4>,      S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v1, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    16,   32,    256, 16,  16,  16,   16,    1,    1,     S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,      S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v1, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,    16,   64,    256, 16,  16,  16,   16,    1,    1,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 16>,     S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v1, F8>,
        // Memory friendly
-        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,   256,   32,    128, 16,  16,  32,   32,    2,    1,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
-        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,   256,   16,    128, 16,  16,  16,   16,    4,    1,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<2, 2, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
-        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,   128,   32,    128, 16,  16,  32,   32,    2,    1,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,     S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
-        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,   128,   16,    128, 16,  16,  16,   16,    4,    1,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,     S<2, 2, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,   128,   32,    128, 16,  16,  32,   32,    1,    1,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,   128,   16,    128, 16,   8,  16,   16,    2,    1,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,              8,              8,          0,          1,           1,                   S<1, 64, 1, 4>,     S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    64,   32,    128, 16,  16,  32,   32,    1,    1,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,     S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    64,   16,    128, 16,  16,  16,   16,    2,    1,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,     S<2, 2, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    32,   16,    128, 16,  16,  16,   16,    1,    1,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,     S<2, 2, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
-        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,    64,    16,   16,     64, 16,  16,  16,   16,    1,    1,     S<4, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<4, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 4>,     S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,    64,    16,   16,    128, 16,  16,  16,   16,    1,    1,     S<8,  8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8,  8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 4>,     S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    16,   32,    128, 16,  16,  16,   16,    1,    1,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,     S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
-        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    16,   64,    128, 16,  16,  16,   16,    1,    2,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,     S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,    16,   64,    128,  8,  16,  16,   16,    1,    1,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,              8,              8,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 16>,    S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    32,   64,    128, 16,  16,  32,   32,    1,    1,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,     S<8, 8, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
-        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    16,  128,    128, 16,  16,  16,   16,    1,    4,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,     S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
-        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    32,  128,    128, 16,  16,  32,   32,    1,    2,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,     S<8, 8, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,    16,  128,    128,  8,  16,  16,   16,    1,    2,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,              8,              8,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           2,                   S<1, 16, 1, 16>,    S<8, 8, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,    32,  128,    128, 16,  16,  32,   32,    1,    1,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 16>,    S<8, 8, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,    64,   16,    512, 16,  16,  16,   16,    1,    1,     S<32, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<32, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 64, 1, 4>,     S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    32,   16,    512, 16,  16,  16,   16,    1,    1,     S<32, 4, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<32, 4, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 4>,     S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,    64,    16,   16,    512, 16,  16,  16,   16,    1,    1,     S<32, 2, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<32, 2, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 4>,     S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    16,   32,    512, 16,  16,  16,   16,    1,    1,     S<32, 4, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<32, 4, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,     S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,    16,  256,    128, 16,  16,  16,   16,    1,    4,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 16>,    S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
-        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,    32,  256,    128, 16,  16,  32,   32,    1,    2,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 16>,    S<8, 8, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,    32,  256,    128, 16,  16,  32,   32,    1,    2,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 16>,    S<8, 8, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,    32,  128,    128, 16,  16,  32,   32,    1,    1,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,      S<8, 8, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v1, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,    32,  128,    256, 16,  16,  32,   32,    1,    1,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,      S<8, 8, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v1, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    32,  128,    128, 16,  16,  32,   32,    1,    2,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,      S<8, 8, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v1, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    32,  128,    128, 16,  16,  32,   32,    1,    2,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,      S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,    32,  128,    128, 16,  16,  32,   32,    1,    1,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,      S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, BF16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,    32,  128,    256, 16,  16,  32,   32,    1,    1,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,      S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>
 #endif
    // clang-format on
    >;

--- a/library/src/tensor_operation_instance/gpu/gemm_multiply_multiply/device_gemm_multiply_multiply_xdl_f8_f8_f16/device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn.hpp
+++ b/library/src/tensor_operation_instance/gpu/gemm_multiply_multiply/device_gemm_multiply_multiply_xdl_f8_f8_f16/device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn.hpp
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2024, Advanced Micro Devices, Inc. All rights reserved.
+
+#include "ck/ck.hpp"
+#include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
+#include "ck/tensor_operation/gpu/device/gemm_specialization.hpp"
+#include "ck/tensor_operation/gpu/device/impl/device_gemm_multiple_d_xdl_cshuffle_v3.hpp"
+
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace instance {
+
+using F8  = f8_t;
+using F16 = half_t;
+using F32 = float;
+
+using Row = tensor_layout::gemm::RowMajor;
+using Col = tensor_layout::gemm::ColumnMajor;
+
+template <index_t... Is>
+using S = Sequence<Is...>;
+
+using PassThrough      = element_wise::PassThrough;
+using MultiplyMultiply = element_wise::MultiplyMultiply;
+
+static constexpr auto GemmDefault    = GemmSpecialization::Default;
+static constexpr auto GemmKPadding   = GemmSpecialization::KPadding;
+static constexpr auto GemmMNPadding  = GemmSpecialization::MNPadding;
+static constexpr auto GemmMNKPadding = GemmSpecialization::MNKPadding;
+
+static constexpr auto Intrawave = BlockGemmPipelineScheduler::Intrawave;
+static constexpr auto Interwave = BlockGemmPipelineScheduler::Interwave;
+
+template <GemmSpecialization GemmSpec>
+using device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn_comp_instances = std::tuple<
+// clang-format off
+        //################################| ALayout| BLayout|         DsLayout| ELayout|AData| BData|          DsData| EData| AccData| Cshuffle|           A|           B|              C|          GEMM| Block|  MPer|  NPer|  KPer| AK1| BK1|MPer| NPer| MXdl| NXdl|  ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds|  BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds|    CShuffle|    CShuffle|     CBlockTransferClusterLengths|  CBlockTransfer|                         Block-wiseGemm|               Block-wiseGemm|
+        //################################|        |        |                 |        | Type|  Type|            Type|  Type|    Type|     Type| Elementwise| Elementwise|    Elementwise|Specialization|  Size| Block| Block| Block|    |    | XDL|  XDL|  Per|  Per|   ThreadCluster|  ThreadCluster| SrcAccessOrder|   SrcVectorDim|      SrcScalar|      DstScalar| AddExtraM|   ThreadCluster|  ThreadCluster| SrcAccessOrder|  SrcVectorDim|      SrcScalar|      DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MXdlPerWave_MWaveMPerXdl| ScalarPerVector|                               Pipeline|                     Pipeline|
+        //################################|        |        |                 |        |     |      |                |      |        |         |   Operation|   Operation|      Operation|              |      |      |      |      |    |    |    |     | Wave| Wave| Lengths_K0_M_K1|   ArrangeOrder|               |               |      PerVector|   PerVector_K1|          | Lengths_K0_N_K1|   ArrangeOrder|               |              |      PerVector|   PerVector_K1|          |  PerShuffle|  PerShuffle| _NBlock_NXdlPerWave_NWaveNPerXdl|   _NWaveNPerXdl|                              Scheduler|                     Verision|
+        //################################|        |        |                 |        |     |      |                |      |        |         |            |            |               |              |      |      |      |      |    |    |    |     |     |     |                |               |               |               |               |               |          |                |               |               |              |               |               |          |            |            |                                 |                |                                       |                             |
+#if defined(__gfx94__) || defined(CK_USE_GFX94) || defined(CK_USE_FP8_ON_UNSUPPORTED_ARCH)
+        // Compute friendly
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,   256,    64,  16,  16,  32,   32,    4,    4,     S<4, 64, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<4, 64, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v4, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   128,   128,   128,  16,  16,  32,   32,    2,    2,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v4, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   128,   128,    64,  16,  16,  32,   32,    2,    2,     S<4, 64, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<4, 64, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v4, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,   256,    64,  16,  16,  32,   32,    4,    4,     S<4, 64, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<4, 64, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v4, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,   256,   128,  16,  16,  32,   32,    4,    4,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,   224,   128,  16,  16,  32,   32,    2,    7,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          2,           1,                   S<1, 64, 1, 4>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,   192,   128,  16,  16,  32,   32,    4,    3,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,   160,   128,  16,  16,  32,   32,    2,    5,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          2,           1,                   S<1, 64, 1, 4>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,   128,   128,  16,  16,  32,   32,    4,    2,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,    96,   128,  16,  16,  32,   32,    2,    3,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          2,           1,                   S<1, 64, 1, 4>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,    64,   128,  16,  16,  32,   32,    4,    1,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   128,   256,   128,  16,  16,  32,   32,    2,    4,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   128,   224,   128,  16,  16,  32,   32,    1,    7,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 64, 1, 4>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   128,   192,   128,  16,  16,  32,   32,    2,    3,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   128,   160,   128,  16,  16,  32,   32,    1,    5,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 64, 1, 4>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   128,   128,   256,  16,  16,  32,   32,    2,    2,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   128,   128,   128,  16,  16,  32,   32,    2,    2,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   128,    96,   256,  16,  16,  32,   32,    1,    3,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 64, 1, 4>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        // Compute friendly
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    64,   256,   128,  16,  16,  32,   32,    1,    4,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    64,   224,   128,  16,  16,  16,   16,    2,    7,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          2,           1,                   S<1, 64, 1, 4>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    64,   192,   256,  16,  16,  32,   32,    1,    3,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    64,   192,   128,  16,  16,  32,   32,    1,    3,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    64,   160,   256,  16,  16,  16,   16,    2,    5,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          2,           1,                   S<1, 64, 1, 4>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    64,   128,   256,  16,  16,  32,   32,    1,    2,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    64,    96,   256,  16,  16,  16,   16,    2,    3,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          2,           1,                   S<1, 64, 1, 4>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    64,    64,   512,  16,  16,  32,   32,    1,    1,     S<32, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<32, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    32,   256,   128,  16,  16,  32,   32,    1,    2,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    32,   224,   256,  16,  16,  16,   16,    1,    7,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<4, 4, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    32,   192,   256,  16,  16,  16,   16,    1,    6,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           2,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    32,   160,   256,  16,  16,  16,   16,    1,    5,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<4, 4, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    32,   128,   256,  16,  16,  32,   32,    1,    1,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    32,    96,   256,  16,  16,  16,   16,    1,    3,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<4, 4, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    32,    64,   512,  16,  16,  16,   16,    1,    2,     S<32, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<32, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           2,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    16,   256,   128,   8,  16,  16,   16,    1,    4,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,              8,              8,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           2,                   S<1, 16, 1, 16>,    S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    16,   192,   256,  16,  16,  16,   16,    1,    3,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 16>,    S<4, 4, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    16,   128,   256,  16,  16,  16,   16,    1,    2,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           2,                   S<1, 16, 1, 16>,    S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    16,    64,   512,  16,  16,  16,   16,    1,    1,     S<32, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<32, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 16>,    S<4, 4, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,   256,   128,  16,  16,  16,   16,    8,    8,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           2,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,   256,    64,  16,  16,  16,   16,    8,    8,     S<4, 64, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<4, 64, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           2,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,   192,   128,  16,  16,  16,   16,    8,    6,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           2,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,   160,   128,  16,  16,  16,   16,    8,    5,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          2,           1,                   S<1, 64, 1, 4>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,   128,   128,  16,  16,  16,   16,    8,    4,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           2,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,    96,   128,  16,  16,  16,   16,    8,    3,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          2,           1,                   S<1, 64, 1, 4>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,    64,   128,  16,  16,  16,   16,    8,    2,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           2,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   224,   256,    128, 16,  16,  16,   16,    7,    8,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           2,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,   224,    128, 16,  16,  16,   16,    8,    7,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          2,           1,                   S<1, 64, 1, 4>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   128,   128,    128, 16,  16,  32,   32,    2,    2,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   128,   128,    128, 16,  16,  32,   32,    2,    2,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v5, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   128,   256,    64,  16,  16,  32,   32,    2,    4,     S<4, 64, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<4, 64, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Interwave, BlockGemmPipelineVersion::v1, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   256,   128,    64,  16,  16,  32,   32,    4,    2,     S<4, 64, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<4, 64, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Interwave, BlockGemmPipelineVersion::v1, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   128,   128,    128, 16,  16,  32,   32,    2,    2,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Interwave, BlockGemmPipelineVersion::v1, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,   128,    64,    128, 16,  16,  32,   32,    2,    1,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    64,   128,    128, 16,  16,  32,   32,    1,    2,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>,  Row,    F8,    F8,    Tuple<F32, F32>, F16,  F32,     F32,     PassThrough, PassThrough, MultiplyMultiply,    GemmSpec,   256,    64,    64,    128, 16,  16,  32,   32,    1,    1,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>
+#endif
+    // clang-format on
+    >;
+
+template <BlockGemmPipelineScheduler BlkGemmPipeSched, GemmSpecialization GemmSpec>
+using device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn_mem_instances = std::tuple<
+// clang-format off
+        //################################| ALayout| BLayout|         DsLayout| ELayout|AData| BData|          DsData| EData| AccData| Cshuffle|           A|           B|               C|          GEMM| Block|  MPer|  NPer|  KPer| AK1| BK1|MPer| NPer| MXdl| NXdl|  ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds|  BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds|    CShuffle|    CShuffle|     CBlockTransferClusterLengths|  CBlockTransfer|    Block-wiseGemm|               Block-wiseGemm|
+        //################################|        |        |                 |        | Type|  Type|            Type|  Type|    Type|     Type| Elementwise| Elementwise|     Elementwise|Specialization|  Size| Block| Block| Block|    |    | XDL|  XDL|  Per|  Per|   ThreadCluster|  ThreadCluster| SrcAccessOrder|   SrcVectorDim|      SrcScalar|      DstScalar| AddExtraM|   ThreadCluster|  ThreadCluster| SrcAccessOrder|  SrcVectorDim|      SrcScalar|      DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MXdlPerWave_MWaveMPerXdl| ScalarPerVector|          Pipeline|                     Pipeline|
+        //################################|        |        |                 |        |     |      |                |      |        |         |   Operation|   Operation|       Operation|              |      |      |      |      |    |    |    |     | Wave| Wave| Lengths_K0_M_K1|   ArrangeOrder|               |               |      PerVector|   PerVector_K1|          | Lengths_K0_N_K1|   ArrangeOrder|               |              |      PerVector|   PerVector_K1|          |  PerShuffle|  PerShuffle| _NBlock_NXdlPerWave_NWaveNPerXdl|   _NWaveNPerXdl|         Scheduler|                     Verision|
+        //################################|        |        |                 |        |     |      |                |      |        |         |            |            |                |              |      |      |      |      |    |    |    |     |     |     |                |               |               |               |               |               |          |                |               |               |              |               |               |          |            |            |                                 |                |                  |                             |
+#if defined(__gfx94__) || defined(CK_USE_GFX94) || defined(CK_USE_FP8_ON_UNSUPPORTED_ARCH)
+        // Latency friendly 
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    32,   16,    128, 16,  16,  16,   16,    1,    1,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,      S<2, 2, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v1, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,    64,    16,   16,    128, 16,  16,  16,   16,    1,    1,     S<8,  8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8,  8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 4>,      S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v1, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    16,   32,    128, 16,  16,  16,   16,    1,    1,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,      S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v1, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,    64,   16,    256, 16,  16,  16,   16,    1,    1,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 64, 1, 4>,      S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v1, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    32,   16,    256, 16,  16,  16,   16,    1,    1,     S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 4>,      S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v1, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,    64,    16,   16,    256, 16,  16,  16,   16,    1,    1,     S<16, 4, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 4, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 4>,      S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v1, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    16,   32,    256, 16,  16,  16,   16,    1,    1,     S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,      S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v1, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,    16,   64,    256, 16,  16,  16,   16,    1,    1,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 16>,     S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v1, F8>,
+        // Memory friendly
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,   128,   32,    128, 16,  16,  32,   32,    1,    1,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,     S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,   128,   16,    128, 16,   8,  16,   16,    2,    1,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,              8,              8,          0,          1,           1,                   S<1, 64, 1, 4>,     S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    64,   32,    128, 16,  16,  32,   32,    1,    1,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,     S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    64,   16,    128, 16,  16,  16,   16,    2,    1,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,     S<2, 2, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    32,   16,    128, 16,  16,  16,   16,    1,    1,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,     S<2, 2, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,    64,    16,   16,    128, 16,  16,  16,   16,    1,    1,     S<8,  8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8,  8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 4>,     S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    16,   32,    128, 16,  16,  16,   16,    1,    1,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,     S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,    16,   64,    128,  8,  16,  16,   16,    1,    1,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,              8,              8,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 16>,    S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    32,   64,    128, 16,  16,  32,   32,    1,    1,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,     S<8, 8, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,    16,  128,    128,  8,  16,  16,   16,    1,    2,     S<16, 16, 1>,    S<1, 0, 2>,    S<1, 0, 2>,               2,              8,              8,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           2,                   S<1, 16, 1, 16>,    S<8, 8, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,    32,  128,    128, 16,  16,  32,   32,    1,    1,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 16>,    S<8, 8, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,    64,   16,    512, 16,  16,  16,   16,    1,    1,     S<32, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<32, 8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 64, 1, 4>,     S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    32,   16,    512, 16,  16,  16,   16,    1,    1,     S<32, 4, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<32, 4, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 4>,     S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,    64,    16,   16,    512, 16,  16,  16,   16,    1,    1,     S<32, 2, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<32, 2, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 4>,     S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    16,   32,    512, 16,  16,  16,   16,    1,    1,     S<32, 4, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<32, 4, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,     S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,    16,  256,    128, 16,  16,  16,   16,    1,    4,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 16>,    S<4, 4, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,    32,  256,    128, 16,  16,  32,   32,    1,    2,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 16>,    S<8, 8, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v2, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,    32,  128,    128, 16,  16,  32,   32,    1,    1,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,      S<8, 8, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v1, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,    32,  128,    256, 16,  16,  32,   32,    1,    1,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,      S<8, 8, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v1, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    32,  128,    128, 16,  16,  32,   32,    1,    2,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,      S<8, 8, 1>,  BlkGemmPipeSched, BlockGemmPipelineVersion::v1, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   128,    32,  128,    128, 16,  16,  32,   32,    1,    2,     S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 16, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 16, 1, 8>,      S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,    32,  128,    128, 16,  16,  32,   32,    1,    1,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,      S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>,
+        DeviceGemmMultiD_Xdl_CShuffle_V3<  Row,     Col,     Tuple<Row, Col>, Row,     F8,     F8,    Tuple<F32, F32>, F16,   F32,     F32,  PassThrough, PassThrough, MultiplyMultiply,     GemmSpec,   256,    32,  128,    256, 16,  16,  32,   32,    1,    1,     S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,    S<8, 32, 1>,     S<1, 0, 2>,    S<1, 0, 2>,               2,             16,             16,          0,          1,           1,                   S<1, 32, 1, 8>,      S<8, 8, 1>,  BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v3, F8>
+#endif
+    // clang-format on
+    >;
+} // namespace instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck
--- a/library/src/tensor_operation_instance/gpu/gemm_multiply_multiply/device_gemm_multiply_multiply_xdl_f8_f8_f16/device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn_comp_default_instance.cpp
+++ b/library/src/tensor_operation_instance/gpu/gemm_multiply_multiply/device_gemm_multiply_multiply_xdl_f8_f8_f16/device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn_comp_default_instance.cpp
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2024, Advanced Micro Devices, Inc. All rights reserved.
+
+#include "device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace instance {
+
+void add_device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn_comp_default_instances(
+    std::vector<std::unique_ptr<DeviceGemmMultipleDSplitK<Row,
+                                                          Col,
+                                                          Tuple<Row, Col>,
+                                                          Row,
+                                                          F8,
+                                                          F8,
+                                                          Tuple<F32, F32>,
+                                                          F16,
+                                                          PassThrough,
+                                                          PassThrough,
+                                                          MultiplyMultiply>>>& instances)
+{
+    add_device_operation_instances(
+        instances,
+        device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn_comp_instances<GemmDefault>{});
+}
+
+} // namespace instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck
--- a/library/src/tensor_operation_instance/gpu/gemm_multiply_multiply/device_gemm_multiply_multiply_xdl_f8_f8_f16/device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn_comp_kpadding_instance.cpp
+++ b/library/src/tensor_operation_instance/gpu/gemm_multiply_multiply/device_gemm_multiply_multiply_xdl_f8_f8_f16/device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn_comp_kpadding_instance.cpp
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2024, Advanced Micro Devices, Inc. All rights reserved.
+
+#include "device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace instance {
+
+void add_device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn_comp_kpadding_instances(
+    std::vector<std::unique_ptr<DeviceGemmMultipleDSplitK<Row,
+                                                          Col,
+                                                          Tuple<Row, Col>,
+                                                          Row,
+                                                          F8,
+                                                          F8,
+                                                          Tuple<F32, F32>,
+                                                          F16,
+                                                          PassThrough,
+                                                          PassThrough,
+                                                          MultiplyMultiply>>>& instances)
+{
+    add_device_operation_instances(
+        instances,
+        device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn_comp_instances<GemmKPadding>{});
+}
+
+} // namespace instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck
--- a/library/src/tensor_operation_instance/gpu/gemm_multiply_multiply/device_gemm_multiply_multiply_xdl_f8_f8_f16/device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn_mem_v1_default_instance.cpp
+++ b/library/src/tensor_operation_instance/gpu/gemm_multiply_multiply/device_gemm_multiply_multiply_xdl_f8_f8_f16/device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn_mem_v1_default_instance.cpp
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2024, Advanced Micro Devices, Inc. All rights reserved.
+
+#include "device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace instance {
+
+void add_device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn_mem_v1_default_instances(
+    std::vector<std::unique_ptr<DeviceGemmMultipleDSplitK<Row,
+                                                          Col,
+                                                          Tuple<Row, Col>,
+                                                          Row,
+                                                          F8,
+                                                          F8,
+                                                          Tuple<F32, F32>,
+                                                          F16,
+                                                          PassThrough,
+                                                          PassThrough,
+                                                          MultiplyMultiply>>>& instances)
+{
+    add_device_operation_instances(
+        instances,
+        device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn_mem_instances<Intrawave,
+                                                                           GemmDefault>{});
+}
+
+} // namespace instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck
--- a/library/src/tensor_operation_instance/gpu/gemm_multiply_multiply/device_gemm_multiply_multiply_xdl_f8_f8_f16/device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn_mem_v1_kpadding_instance.cpp
+++ b/library/src/tensor_operation_instance/gpu/gemm_multiply_multiply/device_gemm_multiply_multiply_xdl_f8_f8_f16/device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn_mem_v1_kpadding_instance.cpp
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2024, Advanced Micro Devices, Inc. All rights reserved.
+
+#include "device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace instance {
+
+void add_device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn_mem_v1_kpadding_instances(
+    std::vector<std::unique_ptr<DeviceGemmMultipleDSplitK<Row,
+                                                          Col,
+                                                          Tuple<Row, Col>,
+                                                          Row,
+                                                          F8,
+                                                          F8,
+                                                          Tuple<F32, F32>,
+                                                          F16,
+                                                          PassThrough,
+                                                          PassThrough,
+                                                          MultiplyMultiply>>>& instances)
+{
+    add_device_operation_instances(
+        instances,
+        device_gemm_multiply_multiply_xdl_f8_f8_f16_mk_nk_mn_mem_instances<Intrawave,
+                                                                           GemmKPadding>{});
+}
+
+} // namespace instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck