revise comment in gemm_softmax_gemm example

dab6bb8d · Anthony Chang · 18ead6e6 · dab6bb8d
Commit dab6bb8d authored Aug 22, 2022 by Anthony Chang
Hide whitespace changes
Inline Side-by-side

Showing with 5 additions and 5 deletions

example/32_batched_gemm_scale_softmax_gemm/batched_gemm_scale_softmax_gemm_xdl_fp16.cpp ...softmax_gemm/batched_gemm_scale_softmax_gemm_xdl_fp16.cpp +5 -5

No files found.
--- a/example/32_batched_gemm_scale_softmax_gemm/batched_gemm_scale_softmax_gemm_xdl_fp16.cpp
+++ b/example/32_batched_gemm_scale_softmax_gemm/batched_gemm_scale_softmax_gemm_xdl_fp16.cpp
@@ -2,11 +2,11 @@
 // Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.

 /*
-Gemm + Gemm fused operation. Computes C_m_o = A_m_k * B0_k_n * B1_n_o
-                                              |------------|
-                                                   Gemm0
-                                              |---------------------|
-                                                       Gemm1
+Gemm + Softmax + Gemm fused operation. Computes C_g_m_o = Softmax(A_g_m_k * B0_g_k_n) * B1_g_n_o
+                                                                  |-----------------|
+                                                                          Gemm0
+                                                          |-------------------------------------|
+                                                                          Gemm1
 */

 #include <iostream>