tidy

CMakeLists.txt

@@ -40,7 +40,7 @@ message(STATUS "Build with HIP ${hip_VERSION}")
 
 ## half
 #find_path(HALF_INCLUDE_DIR half.hpp)
-#message("HALF_INCLUDE_DIR: ${HALF_INCLUDE_DIR}")
+message("HALF_INCLUDE_DIR: ${HALF_INCLUDE_DIR}")
 
 ## tidy
 include(EnableCompilerWarnings)

composable_kernel/include/tensor_operation/gridwise_dynamic_gemm_xdlops_v2r3.hpp

@@ -411,9 +411,6 @@ struct GridwiseDynamicGemm_k0mk1_k0nk1_mn_xdlops_v2r3
         constexpr auto a_block_space_size =
             math::integer_least_multiple(a_k0_m_k1_block_desc.GetElementSpaceSize(), max_lds_align);
 
-        constexpr auto b_block_space_size =
-            math::integer_least_multiple(b_k0_n_k1_block_desc.GetElementSpaceSize(), max_lds_align);
-
         FloatAB* p_a_block = p_shared_block;
         FloatAB* p_b_block = p_shared_block + a_block_space_size;
 
@@ -574,8 +571,6 @@ struct GridwiseDynamicGemm_k0mk1_k0nk1_mn_xdlops_v2r3
                 make_dynamic_naive_tensor_descriptor_packed_v2(make_tuple(
                     I1, I1, I1, I1, Number<M0>{}, Number<1>{}, Number<M2>{}, Number<1>{}));
 
-            StaticBuffer<AddressSpaceEnum_t::Vgpr, FloatC, BlkSize> c_blk_buf_;
-
             // calculate origin of thread output tensor on global memory
             //     blockwise GEMM c matrix starting index
             const auto c_thread_mtx_on_block =

host/driver_offline/CMakeLists.txt

@@ -9,7 +9,6 @@ include_directories(BEFORE
     ${PROJECT_SOURCE_DIR}/composable_kernel/include/problem_transform
     ${PROJECT_SOURCE_DIR}/composable_kernel/include/driver
     ${PROJECT_SOURCE_DIR}/external/rocm/include
-    ${PROJECT_SOURCE_DIR}/external/half/include
 )
 
 set(CONV_FWD_DRIVER_OFFLINE_SOURCE src/conv_fwd_driver_offline.cpp)

host/driver_offline/include/device_dynamic_convolution_forward_implicit_gemm_v4r4r4_xdlops_nhwc_kyxc_nhwk.hpp

@@ -338,9 +338,6 @@ void device_dynamic_convolution_forward_implicit_gemm_v4r4r4_xdlops_nhwc_kyxc_nh
             const auto K = out_n_ho_wo_k_lengths[I3];
             const auto C = wei_k_y_x_c_lengths[I3];
 
-            const auto Hi = in_n_hi_wi_c_lengths[I1];
-            const auto Wi = in_n_hi_wi_c_lengths[I2];
-
             const auto Ho = out_n_ho_wo_k_lengths[I1];
             const auto Wo = out_n_ho_wo_k_lengths[I2];
 

host/driver_offline/src/conv_fwd_driver_offline.cpp

@@ -20,11 +20,11 @@
 #include "device_dynamic_convolution_forward_implicit_gemm_v4r4r4_xdlops_nhwc_kyxc_nhwk.hpp"
 
 #define USE_DYNAMIC_MODE 1
-#define USE_CONV_FWD_V4R4_NCHW 0
-#define USE_CONV_FWD_V4R4R2_NHWC 0
-#define USE_CONV_FWD_V6R1_NCHW 0
+#define USE_CONV_FWD_V4R4_NCHW 1
+#define USE_CONV_FWD_V4R4R2_NHWC 1
+#define USE_CONV_FWD_V6R1_NCHW 1
 #define USE_CONV_FWD_V5R1_NCHW 0
-#define USE_CONV_FWD_V4R4R2_XDL_NCHW 0
+#define USE_CONV_FWD_V4R4R2_XDL_NCHW 1
 #define USE_CONV_FWD_V4R4R4_XDL_NHWC 1
 
 enum ConvForwardAlgo

host/host_tensor/CMakeLists.txt

@@ -10,7 +10,7 @@ set(HOST_TENSOR_SOURCE
 ## the library target
 add_library(host_tensor SHARED ${HOST_TENSOR_SOURCE}) 
 
-#target_include_directories(host_tensor SYSTEM PUBLIC $<BUILD_INTERFACE:${HALF_INCLUDE_DIR}>)
+target_include_directories(host_tensor SYSTEM PUBLIC $<BUILD_INTERFACE:${HALF_INCLUDE_DIR}>)
 
 target_link_libraries(host_tensor PRIVATE hip::device)
 target_link_libraries(host_tensor INTERFACE hip::host)