Small refactoring + doc

include/ck_tile/core/tensor/tile_window.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
 
@@ -18,8 +18,17 @@
 
 namespace ck_tile {
 
-// Note: this tile window do not support single issue
-// you need to use tile_window_linear structure for this purpose
+/**
+ * @brief This class provides tile (windowed) view and access to the device memory.
+ *
+ * @note this tile window do not support single issue you need to use tile_window_linear
+ *       structure for this purpose
+ *
+ * @tparam BottomTensorView_        Class describing & holding device tensor memory.
+ * @tparam WindowLengths_           Spatial sizes of windowed view on tensor.
+ * @tparam StaticTileDistribution_  Thread distribution (mapping) into Tile dimensions
+ * @tparam NumCoord                 TBD
+ */
 template <typename BottomTensorView_,
           typename WindowLengths_,
           typename StaticTileDistribution_,
@@ -1009,6 +1018,14 @@ CK_TILE_DEVICE void move_tile_window(
     window.move(step);
 }
 
+/**
+ * @brief This class provides description of tile windowed view on the device memory.
+ *
+ * @note This class does not provide any functions to read or modify device memory.
+ *
+ * @tparam BottomTensorView_    Class describing & holding device tensor memory.
+ * @tparam WindowLengths_       Spatial sizes of windowed view on tensor.
+ */
 template <typename BottomTensorView_, typename WindowLengths_>
 struct tile_window_with_static_lengths
 {

include/ck_tile/ops/gemm/block/block_universal_gemm_as_bs_cr.hpp

@@ -80,7 +80,7 @@ struct BlockUniversalGemmAsBsCr
         static constexpr index_t InterWaveSchedulingMacClusters = 1;
 
         static constexpr index_t KPack      = WarpGemm::kKPerThread;
-        static constexpr index_t KPerThread = KPerBlock / WarpGemm::kK * KPack;
+        static constexpr index_t KPerThread = KIterPerWarp * KPack;
         static constexpr index_t KRepeat    = KPerThread / KPack;
     };
 

include/ck_tile/ops/gemm/pipeline/gemm_pipeline_agmem_bgmem_creg_v1.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
 
@@ -29,25 +29,14 @@ struct GemmPipelineAGmemBGmemCRegV1
     static constexpr index_t kNPerBlock = BlockGemmShape::kN;
     static constexpr index_t kKPerBlock = BlockGemmShape::kK;
 
-    static constexpr index_t VectorSizeA = Problem::VectorSizeA;
-    static constexpr index_t VectorSizeB = Problem::VectorSizeB;
-    static constexpr index_t VectorSizeC = Problem::VectorSizeC;
+    static constexpr index_t VectorSizeA = Policy::template GetVectorSizeA<Problem>();
+    static constexpr index_t VectorSizeB = Policy::template GetVectorSizeB<Problem>();
+    static constexpr index_t VectorSizeC = Policy::template GetVectorSizeC<Problem>();
 
     static constexpr bool kPadM = Problem::kPadM;
     static constexpr bool kPadN = Problem::kPadN;
     static constexpr bool kPadK = Problem::kPadK;
 
-    CK_TILE_HOST_DEVICE static constexpr index_t GetStaticLdsSize()
-    {
-        return integer_divide_ceil(
-                   sizeof(ADataType) *
-                       Policy::template MakeALdsBlockDescriptor<Problem>().get_element_space_size(),
-                   16) *
-                   16 +
-               sizeof(BDataType) *
-                   Policy::template MakeBLdsBlockDescriptor<Problem>().get_element_space_size();
-    }
-
     CK_TILE_HOST_DEVICE static constexpr index_t GetSmemSize()
     {
         return Policy::template GetSmemSize<Problem>();