Merge branch 'develop' into universal_streamk

.azuredevops/rocm-ci.yml

@@ -23,20 +23,7 @@ trigger:
     - Jenkinsfile
     - LICENSE
 
-pr:
-  autoCancel: true
-  branches:
-    include:
-    - develop
-  paths:
-    exclude:
-    - .github
-    - docs
-    - '.*.y*ml'
-    - '*.md'
-    - Jenkinsfile
-    - LICENSE
-  drafts: false
+pr: none
 
 jobs:
   - template: ${{ variables.CI_COMPONENT_PATH }}/composable_kernel.yml@pipelines_repo

CMakeLists.txt

@@ -117,7 +117,7 @@ else()
     add_definitions(-DPROFILER_ONLY)
     set(GPU_TARGETS "" CACHE STRING "" FORCE)
     if(GPU_TARGETS)
-        message(FATAL_ERROR "For PROFILE_ONLY build, please do not set GPU_TARGETS, use GPU_ARCH = gfx90, gfx94, gfx10, or gfx11")
+        message(FATAL_ERROR "For PROFILE_ONLY build, please do not set GPU_TARGETS, use GPU_ARCH = gfx90, gfx94, gfx10, gfx11 or gfx12")
     endif()
     if(GPU_ARCH MATCHES "gfx90")
         rocm_check_target_ids(DEFAULT_GPU_TARGETS TARGETS "gfx908;gfx90a")
@@ -127,8 +127,10 @@ else()
         rocm_check_target_ids(DEFAULT_GPU_TARGETS TARGETS "gfx1030")
     elseif(GPU_ARCH MATCHES "gfx11")
         rocm_check_target_ids(DEFAULT_GPU_TARGETS TARGETS "gfx1100;gfx1101;gfx1102")
+    elseif(GPU_ARCH MATCHES "gfx12")
+        rocm_check_target_ids(DEFAULT_GPU_TARGETS TARGETS "gfx1200;gfx1201")
     else()
-        message(FATAL_ERROR "For PROFILE_ONLY build, please specify GPU_ARCH as gfx90, gfx94, gfx10, or gfx11")
+        message(FATAL_ERROR "For PROFILE_ONLY build, please specify GPU_ARCH as gfx90, gfx94, gfx10, gfx11 or gfx12")
     endif()
     set(GPU_TARGETS "${DEFAULT_GPU_TARGETS}" CACHE STRING " " FORCE)
 endif()

Jenkinsfile

@@ -493,6 +493,7 @@ def Build_CK(Map conf=[:]){
 
         def variant = env.STAGE_NAME
         def retimage
+
         gitStatusWrapper(credentialsId: "${env.status_wrapper_creds}", gitHubContext: "Jenkins - ${variant}", account: 'ROCm', repo: 'composable_kernel') {
             try {
                 (retimage, image) = getDockerImage(conf)
@@ -660,9 +661,6 @@ CRON_SETTINGS = BRANCH_NAME == "develop" ? '''0 23 * * * % RUN_FULL_QA=true;ROCM
 
 pipeline {
     agent none
-    triggers {
-        parameterizedCron(CRON_SETTINGS)
-    }
     options {
         parallelsAlwaysFailFast()
     }

cmake/EnableCompilerWarnings.cmake

@@ -66,7 +66,7 @@ else()
             -Wunreachable-code
             -Wunused
             -Wno-reserved-identifier
-            -Werror
+	    -Werror
             -Wno-option-ignored
             -Wsign-compare
             -Wno-extra-semi-stmt

example/01_gemm/gemm_wmma_fp16.cpp

@@ -23,45 +23,45 @@ static constexpr auto GemmDefault = ck::tensor_operation::device::GemmSpecializa
 
 // clang-format off
 using DeviceGemmInstance = ck::tensor_operation::device::DeviceGemmWmma_CShuffle
-         < ALayout,             
-           BLayout,             
-           CLayout,             
-           ADataType, 
+         < ALayout,
+           BLayout,
+           CLayout,
+           ADataType,
            BDataType,
-           CDataType, 
-           AccDataType, 
-           CShuffleDataType,  
-           AElementOp,  
-           BElementOp,  
-           CElementOp,    
-           GemmDefault, 
+           CDataType,
+           AccDataType,
+           CShuffleDataType,
+           AElementOp,
+           BElementOp,
+           CElementOp,
+           GemmDefault,
            1,           // Prefetch stage
            128,         // BlockSize
            64,          // MPerBlock
            128,         // NPerBlock
            64,          // KPerBlock
-           8,           // K1
+           2,           // K1
            16,          // MPerWmma
            16,          // NPerWmma
            2,           // M-Repeat // M-PerWmma / M-Repeat = M-Wave
            4,           // N-Repeat // N-PerWmma / N-Repeat = N-Wave
-           S<4, 32, 1>,     
-           S<1, 0, 2>,     
-           S<1, 0, 2>,              
-           2,              
-           8,              
-           8,      
-           true,     
-           S<4, 32, 1>,     
-           S<1, 0, 2>,     
-           S<1, 0, 2>,             
-           2,              
-           8,              
-           8,      
-           true,           
+           S<4, 32, 1>,
+           S<1, 0, 2>,
+           S<1, 0, 2>,
+           2,
+           2,
+           2,
+           true,
+           S<4, 32, 1>,
+           S<1, 0, 2>,
+           S<1, 0, 2>,
+           2,
+           2,
+           2,
+           true,
            1,           // C shuffle (M Repeat) Per store
            1,           // C shuffle (N Repeat) Per store
-           S<1, 32, 1,  4>,               
+           S<1, 32, 1,  4>,
            8>;
 // clang-format on
 

example/01_gemm/run_gemm_example.inc

@@ -159,7 +159,7 @@ bool run_gemm(const ProblemType& problem_size, const ExecutionConfig& config)
         ck::utils::FillUniformDistributionIntegerValue<BDataType>{-5.f, 5.f}(b_k_n);
         break;
     case 4:
-        ck::utils::FillUniformDistributionIntegerValue<ADataType>{1.f, 1.f}(a_m_k);
+        ck::utils::FillUniformDistributionIntegerValue<ADataType>{-5.f, 5.f}(a_m_k);
         ck::utils::FillUniformDistributionIntegerValue<BDataType>{1.f, 1.f}(b_k_n);
         break;
     case 5:

example/04_gemm_add_add_fastgelu/CMakeLists.txt

@@ -24,4 +24,4 @@ foreach(gpu IN LISTS GPU_TARGETS)
         add_example_dependencies(example_gemm_add_add_fastgelu_xdl example_gemm_add_add_fastgelu_xdl_lds_direct_load_fp32)
         set(target 1)
     endif()
-endforeach()
+endforeach()
\ No newline at end of file

example/29_batched_gemm_bias_e_permute/batched_gemm_bias_e_permute_wmma_fp16.cpp

@@ -83,14 +83,14 @@ using DeviceOpInstanceKKNN =
                                                                                   2,
                                                                                   4,
                                                                                   4,
-                                                                                  true,
+                                                                                  false,
                                                                                   S<4, 32, 1>,
                                                                                   S<1, 0, 2>,
                                                                                   S<1, 0, 2>,
                                                                                   2,
                                                                                   4,
                                                                                   4,
-                                                                                  true,
+                                                                                  false,
                                                                                   1,
                                                                                   1,
                                                                                   S<1, 64, 1, 2>,

example/32_batched_gemm_scale_softmax_gemm/cross_attention_forward_wmma_fp16.cpp

@@ -71,7 +71,7 @@ static constexpr auto TensorSpecC  = ck::tensor_operation::device::TensorSpecial
 #define CK_MHA_USE_WAVE_1
 #define CK_MHA_USE_WAVE_2
 #define CK_MHA_USE_WAVE_4
-#define CK_MHA_USE_WAVE_8
+//#define CK_MHA_USE_WAVE_8
 using DeviceMHAFactory = 
     std::tuple<
 #ifdef CK_MHA_USE_WAVE_1
@@ -277,10 +277,10 @@ using DeviceMHAFactory =
             S<2, 8, 8>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 1, false,
             // CShuffleBlockTransfer MN
             1, 1, S<1, 64, 1, 2>, 8,             
-            MaskingSpec>,
+            MaskingSpec>
 #endif
 #ifdef CK_MHA_USE_WAVE_8
-        ck::tensor_operation::device::DeviceBatchedGemmSoftmaxGemmPermute_Wmma_CShuffle<
+        ,ck::tensor_operation::device::DeviceBatchedGemmSoftmaxGemmPermute_Wmma_CShuffle<
             NumDimG, NumDimM, NumDimN, NumDimK, NumDimO,
             ADataType, B0DataType, B1DataType, CDataType, Acc0BiasDataType, Acc0DataType, Acc1BiasDataType, Acc1DataType, CShuffleDataType,
             AElementOp, B0ElementOp, Acc0ElementOp, B1ElementOp, CElementOp,

example/32_batched_gemm_scale_softmax_gemm/self_attention_forward_wmma_fp16.cpp

@@ -71,7 +71,7 @@ static constexpr auto TensorSpecC  = ck::tensor_operation::device::TensorSpecial
 #define CK_MHA_USE_WAVE_1
 #define CK_MHA_USE_WAVE_2
 #define CK_MHA_USE_WAVE_4
-#define CK_MHA_USE_WAVE_8
+//#define CK_MHA_USE_WAVE_8
 using DeviceMHAFactory = 
     std::tuple<
 #ifdef CK_MHA_USE_WAVE_1
@@ -277,10 +277,10 @@ using DeviceMHAFactory =
             S<2, 8, 8>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 1, false,
             // CShuffleBlockTransfer MN
             1, 1, S<1, 64, 1, 2>, 8,             
-            MaskingSpec>,
+            MaskingSpec>
 #endif
 #ifdef CK_MHA_USE_WAVE_8
-        ck::tensor_operation::device::DeviceBatchedGemmSoftmaxGemmPermute_Wmma_CShuffle<
+        ,ck::tensor_operation::device::DeviceBatchedGemmSoftmaxGemmPermute_Wmma_CShuffle<
             NumDimG, NumDimM, NumDimN, NumDimK, NumDimO,
             ADataType, B0DataType, B1DataType, CDataType, Acc0BiasDataType, Acc0DataType, Acc1BiasDataType, Acc1DataType, CShuffleDataType,
             AElementOp, B0ElementOp, Acc0ElementOp, B1ElementOp, CElementOp,

example/CMakeLists.txt

@@ -67,7 +67,7 @@ function(add_example_executable EXAMPLE_NAME FILE_NAME)
     endforeach()
     #Do not build any WMMA examples if gfx11 targets are not on the list
     foreach(source IN LISTS FILE_NAME)
-        if(NOT EX_TARGETS MATCHES "gfx11" AND source MATCHES "_wmma")
+	if(NOT GPU_TARGETS MATCHES "gfx11" AND NOT GPU_TARGETS MATCHES "gfx12" AND source MATCHES "_wmma")
             message("removing wmma example ${source} ")
             list(REMOVE_ITEM FILE_NAME "${source}")
         endif()
@@ -154,7 +154,7 @@ function(add_example_executable_no_testing EXAMPLE_NAME FILE_NAME)
     endforeach()
     #Do not build any WMMA examples if gfx11 targets are not on the list
     foreach(source IN LISTS FILE_NAME)
-        if(NOT EX_TARGETS MATCHES "gfx11" AND source MATCHES "_wmma")
+	if(NOT GPU_TARGETS MATCHES "gfx11" AND NOT GPU_TARGETS MATCHES "gfx12" AND source MATCHES "_wmma")
             message("removing wmma example ${source} ")
             list(REMOVE_ITEM FILE_NAME "${source}")
         endif()

example/ck_tile/01_fmha/CMakeLists.txt

 # generate a list of kernels, but not actually emit files at config stage
 execute_process(
   COMMAND ${Python3_EXECUTABLE} ${CMAKE_CURRENT_LIST_DIR}/generate.py
-  --direction fwd --list_blobs ${CMAKE_CURRENT_BINARY_DIR}/fwd_blob_list.txt
+  --api fwd,fwd_splitkv --list_blobs ${CMAKE_CURRENT_BINARY_DIR}/fwd_blob_list.txt
 )
 
 execute_process(
   COMMAND ${Python3_EXECUTABLE} ${CMAKE_CURRENT_LIST_DIR}/generate.py
-  --direction bwd --list_blobs ${CMAKE_CURRENT_BINARY_DIR}/bwd_blob_list.txt
+  --api bwd --list_blobs ${CMAKE_CURRENT_BINARY_DIR}/bwd_blob_list.txt
 )
 
 # NOTE: for cmake, the FMHA_FWD_GEN_BLOBS/FMHA_BWD_GEN_BLOBS files must be in the same directory
@@ -17,13 +17,13 @@ file(STRINGS ${CMAKE_CURRENT_BINARY_DIR}/bwd_blob_list.txt FMHA_BWD_GEN_BLOBS)
 add_custom_command(
   OUTPUT ${FMHA_FWD_GEN_BLOBS}
   COMMAND ${Python3_EXECUTABLE} ${CMAKE_CURRENT_LIST_DIR}/generate.py
-  --direction fwd --output_dir ${CMAKE_CURRENT_BINARY_DIR}
+  --api fwd,fwd_splitkv --output_dir ${CMAKE_CURRENT_BINARY_DIR}
 )
 
 add_custom_command(
   OUTPUT ${FMHA_BWD_GEN_BLOBS}
   COMMAND ${Python3_EXECUTABLE} ${CMAKE_CURRENT_LIST_DIR}/generate.py
-  --direction bwd --output_dir ${CMAKE_CURRENT_BINARY_DIR}
+  --api bwd --output_dir ${CMAKE_CURRENT_BINARY_DIR}
 )
 
 set(EXAMPLE_FMHA_FWD "tile_example_fmha_fwd")

example/ck_tile/01_fmha/codegen/cmake_config.py

+# SPDX-License-Identifier: MIT
+# Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
+# generate kernel instances to speed up compilation
+
+GEN_DIR = ""    # in Cmake, have to generate files in same folder
\ No newline at end of file

example/ck_tile/01_fmha/codegen/cpp_symbol_map.py

+# SPDX-License-Identifier: MIT
+# Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
+# generate kernel instances to speed up compilation
+
+DTYPE_MAP = {
+    "fp16": "ck_tile::fp16_t",
+    "bf16": "ck_tile::bf16_t",
+    "fp8" : "ck_tile::fp8_t"
+}
+
+MASK_IMPL = {
+    "generic" : "ck_tile::GenericAttentionMask",
+    "simplified"  : "ck_tile::SimplifiedGenericAttentionMask"
+}
+
+_MASK_SIMPLIFIED_MAP = {
+    "s_no" : "ck_tile::SimplifiedGenericAttentionMask<false>",
+    "s_mask" : "ck_tile::SimplifiedGenericAttentionMask<true>",
+}
+
+_MASK_MAP = {
+    "no" : "FmhaMasks::NoMask",
+    "causal" : "FmhaMasks::CausalMask",
+    "generic" : "FmhaMasks::GenericMask"
+}
+
+def get_mask_map(mask : str):
+    if mask == "generic":
+        return _MASK_MAP
+    elif mask == "simplified":
+        return _MASK_SIMPLIFIED_MAP
+    else:
+        assert False
+        return None
+
+_MASK_CHECK_MAP = {
+    "no" : "t.mask_type == mask_enum::no_mask",
+    "causal" : "t.mask_type == mask_enum::mask_top_left || t.mask_type == mask_enum::mask_bottom_right",
+    "generic" : "t.mask_type == mask_enum::window_generic",
+}
+
+_MASK_SIMPLIFIED_CHECK_MAP = {
+    "s_no" : "t.mask_type == mask_enum::no_mask",
+    "s_mask" : "t.mask_type != mask_enum::no_mask",
+}
+
+def get_mask_check_map(mask : str):
+    if mask == "generic":
+        return _MASK_CHECK_MAP
+    elif mask == "simplified":
+        return _MASK_SIMPLIFIED_CHECK_MAP
+    else:
+        assert False
+        return None
+
+BIAS_MAP = {
+    "no" : "ck_tile::BlockAttentionBiasEnum::NO_BIAS",
+    "bias"  : "ck_tile::BlockAttentionBiasEnum::ELEMENTWISE_BIAS",
+    "alibi" : "ck_tile::BlockAttentionBiasEnum::ALIBI"
+}
+
+# TODO: this is ugly
+BIAS_CHECK_MAP = {
+    "no" : "bias_enum::no_bias",
+    "bias"  : "bias_enum::elementwise_bias",
+    "alibi" : "bias_enum::alibi"
+}
+
+MODE_MAP = {
+    "batch" : "false",
+    "group" : "true"
+}
+
+LAYOUT_MAP = {
+    "row" : "true",
+    "col" : "false"
+}
+
+PIPELINE_MAP = {
+    "qr" : "ck_tile::BlockFmhaPipelineQRKSVS",
+    "qr_async" : "ck_tile::BlockFmhaPipelineQRKSVSAsync",
+}
+
+PIPELINE_ENUM_MAP = {
+    "qr" : "ck_tile::BlockFmhaPipelineEnum::QRKSVS",
+    "qr_async" : "ck_tile::BlockFmhaPipelineEnum::QRKSVS_ASYNC",
+}
+
+BOOL_MAP = {
+    "t" : "true",
+    "f" : "false"
+}
\ No newline at end of file

example/ck_tile/01_fmha/fmha_fwd.cpp

@@ -114,6 +114,9 @@ auto create_args(int argc, char* argv[])
         .insert("drop_seed", "1", "seed for random number generator")
         .insert("drop_offset", "0", "offset for random number generator")
         .insert("timer", "gpu", "gpu:gpu timer, cpu:cpu timer")
+        .insert("num_splits",
+                "1",
+                "# of splits for key/value. 0 to determine actual number by heuristic")
         .insert("warmup", "5", "number of iterations before benchmark the kernel")
         .insert("repeat", "20", "number of iterations to benchmark the kernel");
 
@@ -155,6 +158,106 @@ auto get_elimit<ck_tile::fp8_t>(std::string init_method)
     }
 }
 
+int num_splits_heuristic(int batch_nhead_mblocks, int num_SMs, int num_n_blocks, int max_splits)
+{
+    // If we have enough to almost fill the SMs, then just use 1 split
+    if(batch_nhead_mblocks >= 0.8f * num_SMs)
+    {
+        return 1;
+    }
+    max_splits           = std::min({max_splits, num_SMs, num_n_blocks});
+    float max_efficiency = 0.f;
+    std::vector<float> efficiency;
+    efficiency.reserve(max_splits);
+    auto ceildiv = [](int a, int b) { return (a + b - 1) / b; };
+    // Some splits are not eligible. For example, if we have 64 blocks and choose 11 splits,
+    // we'll have 6 * 10 + 4 blocks. If we choose 12 splits, we'll have 6 * 11 + (-2) blocks
+    // (i.e. it's 11 splits anyway).
+    // So we check if the number of blocks per split is the same as the previous num_splits.
+    auto is_split_eligible = [&ceildiv, &num_n_blocks](int num_splits) {
+        return num_splits == 1 ||
+               ceildiv(num_n_blocks, num_splits) != ceildiv(num_n_blocks, num_splits - 1);
+    };
+    for(int num_splits = 1; num_splits <= max_splits; num_splits++)
+    {
+        if(!is_split_eligible(num_splits))
+        {
+            efficiency.push_back(0.f);
+        }
+        else
+        {
+            float n_waves = float(batch_nhead_mblocks * num_splits) / num_SMs;
+            float eff     = n_waves / ceil(n_waves);
+            // printf("num_splits = %d, eff = %f\n", num_splits, eff);
+            if(eff > max_efficiency)
+            {
+                max_efficiency = eff;
+            }
+            efficiency.push_back(eff);
+        }
+    }
+    for(int num_splits = 1; num_splits <= max_splits; num_splits++)
+    {
+        if(!is_split_eligible(num_splits))
+        {
+            continue;
+        }
+        if(efficiency[num_splits - 1] >= 0.85 * max_efficiency)
+        {
+            // printf("num_splits chosen = %d\n", num_splits);
+            return num_splits;
+        }
+    }
+    return 1;
+}
+
+int override_num_splits_if_necessary(
+    int batch, int nhead, int max_seqlen_q, int hdim_v, float p_drop, int num_splits)
+{
+    int device;
+    auto status = hipGetDevice(&device);
+    if(status != hipSuccess)
+    {
+        return num_splits;
+    }
+
+    hipDeviceProp_t props{};
+    status = hipGetDeviceProperties(&props, device);
+    if(status != hipSuccess)
+    {
+        return num_splits;
+    }
+
+    // tile size should match the generate.py
+    const int kM0 = 64;
+    const int kN1 = hdim_v;
+
+    const int num_m_blocks = ck_tile::integer_divide_ceil(max_seqlen_q, kM0);
+    const int num_n_blocks = ck_tile::integer_divide_ceil(hdim_v, kN1);
+
+    if(num_splits < 1 && p_drop == 0.0f)
+    {
+        return num_splits_heuristic(
+            batch * nhead * num_m_blocks, props.multiProcessorCount * 2, num_n_blocks, 128);
+    }
+
+    return num_splits;
+}
+
+float fmha_fwd_dispatch(fmha_fwd_traits traits,
+                        fmha_fwd_args args,
+                        const ck_tile::stream_config& config)
+{
+    if(1 < args.num_splits)
+    {
+        return fmha_fwd_splitkv(traits, args, config);
+    }
+    else
+    {
+        return fmha_fwd(traits, args, config);
+    }
+}
+
 template <typename DataType>
 bool run(const ck_tile::ArgParser& arg_parser)
 {
@@ -260,6 +363,8 @@ bool run(const ck_tile::ArgParser& arg_parser)
         seed.reset();
     }
 
+    int num_splits = arg_parser.get_int("num_splits");
+
     int stream_warmup = arg_parser.get_int("warmup");
     int stream_repeat = arg_parser.get_int("repeat");
     bool kname        = arg_parser.get_bool("kname");
@@ -320,6 +425,18 @@ bool run(const ck_tile::ArgParser& arg_parser)
         }
     }
 
+    // legalize num_splits according to other options
+    if(num_splits < 1)
+    {
+        num_splits = override_num_splits_if_necessary(
+            batch, nhead, max_seqlen_q, hdim_v, p_drop, num_splits);
+    }
+    if(128 < num_splits)
+    {
+        std::cerr << "num_splits greater than 128 is not supported" << std::endl;
+        return false;
+    }
+
     auto get_lengths = [&](bool permute,
                            ck_tile::index_t b /*batch*/,
                            ck_tile::index_t h /*nhead*/,
@@ -361,7 +478,15 @@ bool run(const ck_tile::ArgParser& arg_parser)
                                    : std::array<ck_tile::index_t, 2>{batch, nhead})
             : std::array<ck_tile::index_t, 2>{1, 1});
 
-    // self define lse data layout as [shape_batch, nhead, shape_seqlen_q]
+    ck_tile::HostTensor<LSEDataType> lse_acc_host(
+        1 < num_splits ? std::array<ck_tile::index_t, 4>{num_splits, batch, nhead, max_seqlen_q}
+                       : std::array<ck_tile::index_t, 4>{1, 1, 1, 1});
+    ck_tile::HostTensor<OaccDataType> o_acc_host(
+        1 < num_splits
+            ? std::array<ck_tile::index_t, 5>{num_splits, batch, nhead, max_seqlen_q, hdim_v}
+            : std::array<ck_tile::index_t, 5>{1, 1, 1, 1, 1});
+
+    // self define lse data layout as [batch, nhead, max_seqlen_q]
     ck_tile::HostTensor<LSEDataType> lse_host(
         lse ? std::array<ck_tile::index_t, 3>{batch, nhead, max_seqlen_q}
             : std::array<ck_tile::index_t, 3>{1, 1, 1} /* dummy shape for simplifying code */);
@@ -443,6 +568,8 @@ bool run(const ck_tile::ArgParser& arg_parser)
     ck_tile::DeviceMem k_buf(k_host.get_element_space_size_in_bytes());
     ck_tile::DeviceMem v_buf(v_host.get_element_space_size_in_bytes());
     ck_tile::DeviceMem bias_buf(bias_host.get_element_space_size_in_bytes());
+    ck_tile::DeviceMem lse_acc_buf(lse_acc_host.get_element_space_size_in_bytes());
+    ck_tile::DeviceMem o_acc_buf(o_acc_host.get_element_space_size_in_bytes());
     ck_tile::DeviceMem lse_buf(lse_host.get_element_space_size_in_bytes());
     ck_tile::DeviceMem o_buf(o_host.get_element_space_size_in_bytes());
     ck_tile::DeviceMem seqstart_q(seqstart_q_host.size() * sizeof(int32_t));
@@ -479,7 +606,12 @@ bool run(const ck_tile::ArgParser& arg_parser)
                                       : (std::string("(") + std::to_string(seqlen_kpads[0]) + ")"))
               << ", d:" << hdim_q << "/" << hdim_v << ", scale_s:" << scale_s << ", bias:" << bias
               << ", p_drop:" << p_drop << ", lse:" << lse << ", squant:" << squant
-              << ", mask:" << mask << ", v:" << vlayout << std::flush;
+              << ", mask:" << mask << ", v:" << vlayout;
+    if(1 < num_splits)
+    {
+        std::cout << ", num_splits:" << num_splits;
+    }
+    std::cout << std::flush;
 
     auto fmha_traits = fmha_fwd_traits{hdim_q,
                                        hdim_v,
@@ -523,6 +655,7 @@ bool run(const ck_tile::ArgParser& arg_parser)
         }();
         const ck_tile::index_t stride_bias    = (i_perm ? shape_seqlen_k : 1 * shape_seqlen_k);
         const ck_tile::index_t stride_randval = (max_seqlen_k);
+        const ck_tile::index_t stride_o_acc   = hdim_v;
         const ck_tile::index_t stride_o       = (o_perm ? hdim_v : nhead * hdim_v);
         // setup nhead_stride_* arguments
         const ck_tile::index_t nhead_stride_q = (i_perm ? shape_seqlen_q * hdim_q : hdim_q);
@@ -537,6 +670,8 @@ bool run(const ck_tile::ArgParser& arg_parser)
             (i_perm ? 0 * shape_seqlen_q * shape_seqlen_k : 0 * shape_seqlen_k);
         const ck_tile::index_t nhead_stride_randval = (shape_seqlen_q * max_seqlen_k);
         const ck_tile::index_t nhead_stride_lse     = max_seqlen_q;
+        const ck_tile::index_t nhead_stride_lse_acc = max_seqlen_q;
+        const ck_tile::index_t nhead_stride_o_acc   = (max_seqlen_q * hdim_v);
         const ck_tile::index_t nhead_stride_o       = (o_perm ? shape_seqlen_q * hdim_v : hdim_v);
         // setup batch_stride_* arguments
         const ck_tile::index_t batch_stride_q       = (nhead * shape_seqlen_q * hdim_q);
@@ -545,7 +680,12 @@ bool run(const ck_tile::ArgParser& arg_parser)
         const ck_tile::index_t batch_stride_bias    = (0 * nhead * shape_seqlen_q * shape_seqlen_k);
         const ck_tile::index_t batch_stride_randval = (nhead * shape_seqlen_q * max_seqlen_k);
         const ck_tile::index_t batch_stride_lse     = (nhead * max_seqlen_q);
+        const ck_tile::index_t batch_stride_lse_acc = (nhead * max_seqlen_q);
+        const ck_tile::index_t batch_stride_o_acc   = (nhead * max_seqlen_q * hdim_v);
         const ck_tile::index_t batch_stride_o       = (nhead * shape_seqlen_q * hdim_v);
+        // setup split_stride_* arguments (only used in split-kv kernel)
+        const ck_tile::index_t split_stride_lse_acc = (batch * nhead * max_seqlen_q);
+        const ck_tile::index_t split_stride_o_acc   = (batch * nhead * max_seqlen_q * hdim_v);
 
         return fmha_fwd_args{q_buf.GetDeviceBuffer(),
                              k_buf.GetDeviceBuffer(),
@@ -553,6 +693,8 @@ bool run(const ck_tile::ArgParser& arg_parser)
                              bias.type == bias_enum::alibi ? alibi_slope_buf.GetDeviceBuffer()
                                                            : bias_buf.GetDeviceBuffer(),
                              randval_buf.GetDeviceBuffer(),
+                             lse_acc_buf.GetDeviceBuffer(),
+                             o_acc_buf.GetDeviceBuffer(),
                              lse_buf.GetDeviceBuffer(),
                              o_buf.GetDeviceBuffer(),
                              seqstart_q.GetDeviceBuffer(),
@@ -566,6 +708,7 @@ bool run(const ck_tile::ArgParser& arg_parser)
                              hdim_v,
                              nhead,
                              nhead_k,
+                             num_splits,
                              scale_s,
                              scale_p,
                              scale_o,
@@ -575,6 +718,7 @@ bool run(const ck_tile::ArgParser& arg_parser)
                              bias.type == bias_enum::alibi ? (bias.rank_info == 0 ? 0 : nhead)
                                                            : stride_bias,
                              stride_randval,
+                             stride_o_acc,
                              stride_o,
                              nhead_stride_q,
                              nhead_stride_k,
@@ -582,6 +726,8 @@ bool run(const ck_tile::ArgParser& arg_parser)
                              nhead_stride_bias,
                              nhead_stride_randval,
                              nhead_stride_lse,
+                             nhead_stride_lse_acc,
+                             nhead_stride_o_acc,
                              nhead_stride_o,
                              batch_stride_q,
                              batch_stride_k,
@@ -589,7 +735,11 @@ bool run(const ck_tile::ArgParser& arg_parser)
                              batch_stride_bias,
                              batch_stride_randval,
                              batch_stride_lse,
+                             batch_stride_lse_acc,
+                             batch_stride_o_acc,
                              batch_stride_o,
+                             split_stride_lse_acc,
+                             split_stride_o_acc,
                              mask.left,
                              mask.right,
                              static_cast<ck_tile::index_t>(mask.type),
@@ -598,7 +748,7 @@ bool run(const ck_tile::ArgParser& arg_parser)
                              {drop_seed, drop_offset}};
     }();
 
-    float ave_time = fmha_fwd(fmha_traits, fmha_args, stream_config);
+    float ave_time = fmha_fwd_dispatch(fmha_traits, fmha_args, stream_config);
 
     if(ave_time < 0)
     {
@@ -849,14 +999,14 @@ bool run(const ck_tile::ArgParser& arg_parser)
             lse_host_result.ForEach(
                 [&](auto& self, auto idx) { self(idx) = lse_host(wb, idx[0], idx[1]); });
 
-            bool lse_pass = ck_tile::check_err(lse_host_result,
-                                               lse_host_ref,
-                                               "LSE Error: Incorrect results!",
-                                               rtol,
-                                               atol,
-                                               /* allow_infinity_ref = */ true);
+            cur_pass = ck_tile::check_err(lse_host_result,
+                                          lse_host_ref,
+                                          "LSE Error: Incorrect results!",
+                                          rtol,
+                                          atol,
+                                          /* allow_infinity_ref = */ true);
 
-            pass &= lse_pass;
+            pass &= cur_pass;
             if(!cur_pass)
             {
                 std::cerr << "LSE mismatch found at batch: " << wb << std::endl