Merge branch 'ck_tile/layernorm2d_fwd_optimize' into ck_tile/ln_add_cache_clear

0475a327 · dummycoderfe · c9b961ab · 27ff3dec · 0475a327 · 0475a327
Commit 0475a327 authored Nov 04, 2024 by dummycoderfe
20 changed files
--- a/example/ck_tile/11_add_rmsnorm2d_rdquant/instances/add_rmsnorm2d_rdquant_fwd_fp16_n4096_instance.cpp
+++ b/example/ck_tile/11_add_rmsnorm2d_rdquant/instances/add_rmsnorm2d_rdquant_fwd_fp16_n4096_instance.cpp
--- a/example/ck_tile/11_add_rmsnorm2d_rdquant/instances/add_rmsnorm2d_rdquant_fwd_fp16_n4096_tp_instance.cpp
+++ b/example/ck_tile/11_add_rmsnorm2d_rdquant/instances/add_rmsnorm2d_rdquant_fwd_fp16_n4096_tp_instance.cpp
--- a/example/ck_tile/11_add_rmsnorm2d_rdquant/instances/add_rmsnorm2d_rdquant_fwd_fp16_n512_instance.cpp
+++ b/example/ck_tile/11_add_rmsnorm2d_rdquant/instances/add_rmsnorm2d_rdquant_fwd_fp16_n512_instance.cpp
--- a/example/ck_tile/11_add_rmsnorm2d_rdquant/instances/add_rmsnorm2d_rdquant_fwd_fp16_n64_n128_instance.cpp
+++ b/example/ck_tile/11_add_rmsnorm2d_rdquant/instances/add_rmsnorm2d_rdquant_fwd_fp16_n64_n128_instance.cpp
--- a/example/ck_tile/11_add_rmsnorm2d_rdquant/instances/add_rmsnorm2d_rdquant_fwd_fp16_n768_instance.cpp
+++ b/example/ck_tile/11_add_rmsnorm2d_rdquant/instances/add_rmsnorm2d_rdquant_fwd_fp16_n768_instance.cpp
--- a/example/ck_tile/11_add_rmsnorm2d_rdquant/instances/add_rmsnorm2d_rdquant_fwd_instance_common.hpp
+++ b/example/ck_tile/11_add_rmsnorm2d_rdquant/instances/add_rmsnorm2d_rdquant_fwd_instance_common.hpp
--- a/example/ck_tile/11_add_rmsnorm2d_rdquant/script/perf_test.sh
+++ b/example/ck_tile/11_add_rmsnorm2d_rdquant/script/perf_test.sh
--- a/example/ck_tile/11_add_rmsnorm2d_rdquant/script/smoke_test.sh
+++ b/example/ck_tile/11_add_rmsnorm2d_rdquant/script/smoke_test.sh
--- a/example/ck_tile/12_smoothquant/CMakeLists.txt
+++ b/example/ck_tile/12_smoothquant/CMakeLists.txt
--- a/example/ck_tile/12_smoothquant/README.md
+++ b/example/ck_tile/12_smoothquant/README.md
--- a/example/ck_tile/12_smoothquant/example_smoothquant.cpp
+++ b/example/ck_tile/12_smoothquant/example_smoothquant.cpp
--- a/example/ck_tile/12_smoothquant/instances/smoothquant_bf16_n1024_instance.cpp
+++ b/example/ck_tile/12_smoothquant/instances/smoothquant_bf16_n1024_instance.cpp
--- a/example/ck_tile/12_smoothquant/instances/smoothquant_bf16_n1536_instance.cpp
+++ b/example/ck_tile/12_smoothquant/instances/smoothquant_bf16_n1536_instance.cpp
--- a/example/ck_tile/12_smoothquant/instances/smoothquant_bf16_n2048_instance.cpp
+++ b/example/ck_tile/12_smoothquant/instances/smoothquant_bf16_n2048_instance.cpp
--- a/example/ck_tile/12_smoothquant/instances/smoothquant_bf16_n256_instance.cpp
+++ b/example/ck_tile/12_smoothquant/instances/smoothquant_bf16_n256_instance.cpp
--- a/example/ck_tile/12_smoothquant/instances/smoothquant_bf16_n3072_instance.cpp
+++ b/example/ck_tile/12_smoothquant/instances/smoothquant_bf16_n3072_instance.cpp
--- a/example/ck_tile/12_smoothquant/instances/smoothquant_bf16_n4096_instance.cpp
+++ b/example/ck_tile/12_smoothquant/instances/smoothquant_bf16_n4096_instance.cpp
--- a/example/ck_tile/12_smoothquant/instances/smoothquant_bf16_n4096_tp_instance.cpp
+++ b/example/ck_tile/12_smoothquant/instances/smoothquant_bf16_n4096_tp_instance.cpp
--- a/example/ck_tile/12_smoothquant/instances/smoothquant_bf16_n512_instance.cpp
+++ b/example/ck_tile/12_smoothquant/instances/smoothquant_bf16_n512_instance.cpp
--- a/example/ck_tile/12_smoothquant/instances/smoothquant_bf16_n64_n128_instance.cpp
+++ b/example/ck_tile/12_smoothquant/instances/smoothquant_bf16_n64_n128_instance.cpp