NHWC conv 2d: bwd fp32/fp16/bfp16/int8, Device level tuning and host API (#92)

* start conv2d bwd api * kernel running * add bwd reference * change to no shuffle * fix bwd reference * pass verification * add Filter1x1Stride1Pad0 and start testing * change some tuning parameter * fix test error * add fp16 tuning parameter * add bf16 tuning parameter * add int8 tuning parameters * change fp32 tuning parameter * add bwd to profiler * fix bug for bwd profiler * fix ckProfiler bug * change conv2d_bwd_xdl to fp16 * fix bug in comments * fix precompile id * fix enum conv name * chage _bwd_ to _bwd_data_ * change conv2d_bwd example id * bwd to bwd data * fix prehead * fix MakeDefaultBlock2CTileMap ,import form merge develop * format bwd instance * bwd to bwd data * change name bwd to bwd data * change name bwd to bwd data in example * formate code * change conv2d bwd data id in example * rewrite readme for example * fix CalculateMagicNumbers about div zero * add workaround CK_WORKAROUND_SWDEV_325164 * change test_conf2d_bwd_data show info * format * fix bug for workaround:CK_WORKAROUND_SWDEV_325164 * formate tuning parameters * formate tuning parameters again * formate tuning parameters 3 * formate tuning parameters 4 * remove add function template * format * update comment Co-authored-by: ltqin <letaoqin@amd.com> Co-authored-by: Chao Liu <chao.liu2@amd.com>
2026-05-01 20:21:23 +00:00 · 2022-03-04 14:08:26 +08:00
parent 992f71e371
commit c254e5abd2
22 changed files with 2652 additions and 94 deletions
--- a/example/12_conv2d_bwd_data_xdl/README.md
+++ b/example/12_conv2d_bwd_data_xdl/README.md
@@ -0,0 +1,79 @@
+# Instructions for ```conv2d_bwd_data_xdl``` Example
+
+## Docker script
+```bash
+docker run                                                                   \
+-it                                                                          \
+--rm                                                                         \
+--privileged                                                                 \
+--group-add sudo                                                             \
+-w /root/workspace                                                           \
+-v ${PATH_TO_LOCAL_WORKSPACE}:/root/workspace                                \
+rocm/tensorflow:rocm4.3.1-tf2.6-dev                                          \
+/bin/bash
+```
+
+## Build ```conv2d_bwd_data_xdl```
+```bash
+mkdir build && cd build
+```
+
+```bash
+# Need to specify target ID, example below is gfx908
+cmake                                                                  \
+-D BUILD_DEV=OFF                                                       \
+-D CMAKE_BUILD_TYPE=Release                                            \
+-D CMAKE_CXX_FLAGS="-DCK_AMD_GPU_GFX908 --amdgpu-target=gfx908 -O3 "   \
+-D CMAKE_CXX_COMPILER=/opt/rocm/bin/hipcc                              \
+-D CMAKE_PREFIX_PATH=/opt/rocm                                         \
+..
+```
+
+```bash
+ make -j conv2d_bwd_data_xdl
+```
+
+## Run ```conv2d_bwd_data_xdl```
+```bash
+#arg1: verification (0=no, 1=yes)
+#arg2: initialization (0=no init, 1=integer value, 2=decimal value)
+#arg3: run kernel # of times (>1)
+#arg4 to 18: N, K, C, Y, X, Hi, Wi, Sy, Sx, Dy, Dx, LeftPy, LeftPx, RightPy, RightPx
+./bin/conv2d_bwd_data_xdl 0 1 5
+```
+
+Result
+```
+in_n_c_hi_wi: dim 4, lengths {128, 256, 71, 71}, strides {1290496, 1, 18176, 256}
+wei_k_c_y_x: dim 4, lengths {256, 256, 3, 3}, strides {2304, 1, 768, 256}
+out_n_k_ho_wo: dim 4, lengths {128, 256, 36, 36}, strides {331776, 1, 9216, 256}
+arg.a_grid_desc_k0_m_k1_container_{128, 175232, 8}
+arg.b_grid_desc_k0_n_k1_container_{128, 256, 8}
+arg.c_grid_desc_m_n_container_{ 175232, 256}
+arg.c_grid_desc_m0_n0_m1_n1_m2_m3_m4_n2_container_( 2738, 4, 2, 2, 4, 2 ) 
+launch_and_time_kernel: grid_dim {2738, 1, 1}, block_dim {256, 1, 1} 
+Warm up
+Start running 1 times...
+arg.a_grid_desc_k0_m_k1_container_{64, 175232, 8}
+arg.b_grid_desc_k0_n_k1_container_{64, 256, 8}
+arg.c_grid_desc_m_n_container_{ 175232, 256}
+arg.c_grid_desc_m0_n0_m1_n1_m2_m3_m4_n2_container_( 2738, 4, 2, 2, 4, 2 ) 
+launch_and_time_kernel: grid_dim {2738, 1, 1}, block_dim {256, 1, 1} 
+Warm up
+Start running 1 times...
+arg.a_grid_desc_k0_m_k1_container_{64, 175232, 8}
+arg.b_grid_desc_k0_n_k1_container_{64, 256, 8}
+arg.c_grid_desc_m_n_container_{ 175232, 256}
+arg.c_grid_desc_m0_n0_m1_n1_m2_m3_m4_n2_container_( 2738, 4, 2, 2, 4, 2 ) 
+launch_and_time_kernel: grid_dim {2738, 1, 1}, block_dim {256, 1, 1} 
+Warm up
+Start running 1 times...
+arg.a_grid_desc_k0_m_k1_container_{32, 175232, 8}
+arg.b_grid_desc_k0_n_k1_container_{32, 256, 8}
+arg.c_grid_desc_m_n_container_{ 175232, 256}
+arg.c_grid_desc_m0_n0_m1_n1_m2_m3_m4_n2_container_( 2738, 4, 2, 2, 4, 2 ) 
+launch_and_time_kernel: grid_dim {2738, 1, 1}, block_dim {256, 1, 1} 
+Warm up
+Start running 1 times...
+Perf: 2.45966 ms, 79.5597 TFlops, 169.325 GB/s
+```