ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-07-11 01:27:34 +00:00
Code Issues Packages Projects Releases Wiki Activity

improved multiply_add profiler layout help

Browse Source
This commit is contained in:
Zoltan Lakatos
2025-05-30 13:16:42 +00:00
parent 8727762e74
commit 63513c372b
1 changed files with 4 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
profiler/src/profile_gemm_add_multiply.cpp
View File

@@ -35,10 +35,10 @@ int profile_gemm_add_multiply(int argc, char* argv[])
// clang-format off
printf("arg1: tensor operation (" OP_NAME ": " OP_DESC ")\n");
printf("arg2: data type (0: fp32; 1: fp16; 2: bf16; 3: int8)\n");
printf("arg3: matrix layout (0: E[m, n] = AddMultiply((A[m, k] * B[k, n] + D0[m, n]) x D1[m, n]);\n");
printf(" 1: E[m, n] = AddMultiply((A[m, k] * B[n, k] + D0[m, n]) x D1[m, n]);\n");
printf(" 2: E[m, n] = AddMultiply((A[k, m] * B[k, n] + D0[m, n]) x D1[m, n]);\n");
printf(" 3: E[m, n] = AddMultiply((A[k, m] * B[n, k] + D0[m, n]) x D1[m, n]))\n");
printf("arg3: matrix layout (0: E[m, n] = (A[m, k] * B[k, n] + D0[m, n]) x D1[m, n];\n");
printf(" 1: E[m, n] = (A[m, k] * B[n, k] + D0[m, n]) x D1[m, n];\n");
printf(" 2: E[m, n] = (A[k, m] * B[k, n] + D0[m, n]) x D1[m, n];\n");
printf(" 3: E[m, n] = (A[k, m] * B[n, k] + D0[m, n]) x D1[m, n])\n");
printf("arg4: verification (0: no; 1: yes)\n");
printf("arg5: initialization (0: no init; 1: integer value; 2: decimal value)\n");
printf("arg6: print tensor value (0: no; 1: yes)\n");