ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-20 12:59:49 +00:00
Code Issues Packages Projects Releases Wiki Activity

[CK Tile] Int8 Support on CK Tile GEMM (#2267)

Browse Source 
* updates to support int8 in 03_gemm example

* added comments, using aliases, helper functions

* test(gemm_universal): add test cases for int8 gemm pipeline

* fix(test_gemm): fix for failing test unit test for int8

* test(ck_tile): add int8 unit test for gemm universal

* refactor(gemm_universal): GPU reference verification for GEMM code improved

* style(gemm_universal): removed extra comments and did clang format

* merging recent changes to universal gemm to tile_engine

* ck tile engine integration work

* feat(tile_engine): add int8 support to tile engine ops/gemm

* feat(tile_engine): added 32 32 16 mfma instances to tile engine for int8

* style: Format code with clang-format-12

* refactor(tile_engine): address review comments

* style: removed unhelpful comments & unused variables.

* build: tile engine uses default config

* feat: add int8 support for CK_TILE GEMM

* style: added trailing commas to codegen_utils.py

* refactor: tile engine

* refactor: formatting and code review

* refactor: code formatting for python files

* fix: suppress build warning

* add support for gfx950

* refactor:KWarpTile size in gemms util

* Fix the branch and wrap up the k warp tile

* Add bf8 integration

* refactor: clang format and rebase

---------

Co-authored-by: zjli2013 <leezhengjiang@gmail.com>
Co-authored-by: AviralGoelAMD <aviral.goel@amd.com>
Co-authored-by: Khushbu Agarwal <khuagarw@amd.com>

[ROCm/composable_kernel commit: e03293ebce]
This commit is contained in:
Thomas Ning
2025-06-25 08:20:35 -07:00
committed by GitHub
parent 44656b6230
commit 90add28587
24 changed files with 815 additions and 301 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
include/ck_tile/core/numeric/integer.hpp
View File

@@ -7,6 +7,7 @@
namespace ck_tile {
using index_t = int32_t;
using int32_t = int32_t;
using long_index_t = int64_t;
using int8_t = int8_t;

17
include/ck_tile/core/tensor/buffer_view.hpp
View File

@@ -1009,6 +1009,15 @@ struct buffer_view<address_space_enum::lds,
std::is_same_v<remove_cvref_t<X>, int8x8_t>) ||
(std::is_same_v<remove_cvref_t<T>, int8x16_t> &&
std::is_same_v<remove_cvref_t<X>, int8x16_t>) ||
// int8 on thread buffer
(std::is_same_v<remove_cvref_t<T>, int8_t> &&
std::is_same_v<remove_cvref_t<X>, thread_buffer<int8_t, 8>>) ||
(std::is_same_v<remove_cvref_t<T>, int8_t> &&
std::is_same_v<remove_cvref_t<X>, thread_buffer<int8_t, 4>>) ||
(std::is_same_v<remove_cvref_t<T>, int8_t> &&
std::is_same_v<remove_cvref_t<X>, thread_buffer<int8_t, 2>>) ||
(std::is_same_v<remove_cvref_t<T>, int8_t> &&
std::is_same_v<remove_cvref_t<X>, thread_buffer<int8_t, 1>>) ||
// ext_vector_type for pk_int4 must use int8_t as type
(std::is_same_v<remove_cvref_t<T>, pk_int4_t> &&
std::is_same_v<remove_cvref_t<X>, thread_buffer<pk_int4_t, 1>>) ||
@@ -1031,6 +1040,8 @@ struct buffer_view<address_space_enum::lds,
if constexpr((std::is_same_v<remove_cvref_t<T>, int8_t> &&
std::is_same_v<remove_cvref_t<X>, int8_t>) ||
(std::is_same_v<remove_cvref_t<T>, int8_t> &&
std::is_same_v<remove_cvref_t<X>, thread_buffer<int8_t, 1>>) ||
(std::is_same_v<remove_cvref_t<T>, pk_int4_t> &&
std::is_same_v<remove_cvref_t<X>, thread_buffer<pk_int4_t, 1>>))
{
@@ -1041,6 +1052,8 @@ struct buffer_view<address_space_enum::lds,
}
else if constexpr((std::is_same_v<remove_cvref_t<T>, int8_t> &&
std::is_same_v<remove_cvref_t<X>, int8x2_t>) ||
(std::is_same_v<remove_cvref_t<T>, int8_t> &&
std::is_same_v<remove_cvref_t<X>, thread_buffer<int8_t, 2>>) ||
(std::is_same_v<remove_cvref_t<T>, pk_int4_t> &&
std::is_same_v<remove_cvref_t<X>, thread_buffer<pk_int4_t, 2>>))
{
@@ -1051,6 +1064,8 @@ struct buffer_view<address_space_enum::lds,
}
else if constexpr((std::is_same_v<remove_cvref_t<T>, int8_t> &&
std::is_same_v<remove_cvref_t<X>, int8x4_t>) ||
(std::is_same_v<remove_cvref_t<T>, int8_t> &&
std::is_same_v<remove_cvref_t<X>, thread_buffer<int8_t, 4>>) ||
(std::is_same_v<remove_cvref_t<T>, pk_int4_t> &&
std::is_same_v<remove_cvref_t<X>, thread_buffer<pk_int4_t, 4>>))
{
@@ -1061,6 +1076,8 @@ struct buffer_view<address_space_enum::lds,
}
else if constexpr((std::is_same_v<remove_cvref_t<T>, int8_t> &&
std::is_same_v<remove_cvref_t<X>, int8x8_t>) ||
(std::is_same_v<remove_cvref_t<T>, int8_t> &&
std::is_same_v<remove_cvref_t<X>, thread_buffer<int8_t, 8>>) ||
(std::is_same_v<remove_cvref_t<T>, pk_int4_t> &&
std::is_same_v<remove_cvref_t<X>, thread_buffer<pk_int4_t, 8>>))
{