Skip to content

GitLab

Unverified Commit a8b539da authored by Chao Liu's avatar Chao Liu Committed by GitHub
Browse files

Update README.md

parent 4361cffd
Hide whitespace changes
Inline Side-by-side
Showing with 18 additions and 18 deletions
example/4_conv_xdl_bias_relu_add/README.md
View file @ a8b539da
# Instructions for ```gemm_xdl_bias_add``` Example # Instructions for ```conv_xdl_bias_relu_add``` Example
## Docker script ## Docker script
```bash ```bash
...@@ -13,7 +13,7 @@ rocm/tensorflow:rocm4.3.1-tf2.6-dev \ ...@@ -13,7 +13,7 @@ rocm/tensorflow:rocm4.3.1-tf2.6-dev \
/bin/bash /bin/bash
``` ```
## Build ```gemm_xdl_bias_add``` ## Build ```conv_xdl_bias_relu_add```
```bash ```bash
mkdir build && cd build mkdir build && cd build
``` ```
...@@ -30,32 +30,32 @@ cmake \ ...@@ -30,32 +30,32 @@ cmake \
``` ```
```bash ```bash
make -j gemm_xdl_bias_add make -j conv_xdl_bias_relu_add
``` ```
## Run ```gemm_xdl_bias_add``` ## Run ```conv_xdl_bias_relu_add```
```bash ```bash
#arg1: verification (0=no, 1=yes) #arg1: verification (0=no, 1=yes)
#arg2: initialization (0=no init, 1=integer value, 2=decimal value) #arg2: initialization (0=no init, 1=integer value, 2=decimal value)
#arg3: run kernel # of times (>1) #arg3: run kernel # of times (>1)
#arg4 to 9: M (256x), N(128x), K(32x), StrideA, StrideB, StrideC #arg4 to 18: N, K, C, Y, X, Hi, Wi, Sy, Sx, Dy, Dx, LeftPy, LeftPx, RightPy, RightPx
./example/gemm_xdl_bias_add 0 1 5 3840 4096 4096 4096 4096 4096 ./example/conv_xdl_bias_relu_add 0 1 5
``` ```
Result (MI100 @ 1087Mhz, 133.5TFlops peak FP16) Result (MI100 @ 1087Mhz, 133.5TFlops peak FP16)
``` ```
a_m_k: dim 2, lengths {3840, 4096}, strides {4096, 1} in_n_c_hi_wi: dim 4, lengths {128, 192, 71, 71}, strides {967872, 1, 13632, 192}
b_k_n: dim 2, lengths {4096, 4096}, strides {1, 4096} wei_k_c_y_x: dim 4, lengths {256, 192, 3, 3}, strides {1728, 1, 576, 192}
c_m_n: dim 2, lengths {3840, 4096}, strides {4096, 1} out_n_k_ho_wo: dim 4, lengths {128, 256, 36, 36}, strides {331776, 1, 9216, 256}
c0_m_n: dim 2, lengths {3840, 4096}, strides {4096, 1} bias_k: dim 1, lengths {256}, strides {1}
c1_m_n: dim 2, lengths {3840, 4096}, strides {1, 0} resi_n_k_ho_wo: dim 4, lengths {128, 256, 36, 36}, strides {331776, 1, 9216, 256}
arg.a_grid_desc_k0_m_k1_{512, 3840, 8} arg.a_grid_desc_k0_m_k1_{216, 165888, 8}
arg.b_grid_desc_k0_n_k1_{512, 4096, 8} arg.b_grid_desc_k0_n_k1_{216, 256, 8}
arg.c_grid_desc_m_n_{ 3840, 4096} arg.c_grid_desc_m_n_{ 165888, 256}
arg.c0_grid_desc_m_n_{ 3840, 4096} arg.c0_grid_desc_m_n_{ 165888, 256}
arg.c1_grid_desc_m_n_{ 3840, 4096} arg.c1_grid_desc_m_n_{ 165888, 256}
launch_and_time_kernel: grid_dim {480, 1, 1}, block_dim {256, 1, 1} launch_and_time_kernel: grid_dim {1296, 1, 1}, block_dim {256, 1, 1}
Warm up Warm up
Start running 5 times... Start running 5 times...
Perf: 1.27583 ms, 100.992 TFlops, 73.9688 GB/s Perf: 1.71779 ms, 85.4396 TFlops, 194.2 GB/s
``` ```
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment