composable_kernel

ROCm/composable_kernel

Fork 0

mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-18 20:09:25 +00:00

Commit Graph

Author SHA1 Message Date

Author	SHA1	Message	Date
linqunAMD	01b8e6a80f	[CK_TILE] Refine Generic2dBlockShape to fix ck_tile example 2,10,11,14 on rdna3 and 4 (#2795 ) BlockWarps, WarpTile in Generic2dBlockShape are wave size dependent, it causes mangled name mismatch between host and device side. Solution: Replace them with ThreadPerBlock and move BlockWarps, WarpTile calculation into Generic2dBlockShape [ROCm/composable_kernel commit: `c254f3d7b4`]	2025-09-10 08:29:20 +08:00
Emily Martins	70f25296e6	[CK_Tile] Refactor MOE Sorting and Smoothquant ctests to gtests (#2596 ) * refactor moe_sorting ctests to use gtest framework * Refactor ctests for smoothquant to gtests * fix clang format to use version 18 * Print local_eid in MOE sorting gtests * Remove extra space in smoothquant output [ROCm/composable_kernel commit: `70dce4e0c6`]	2025-08-14 10:54:57 -07:00
Emily Martins	356180d6a3	Tests for CK Tile Batched Transpose and Smoothquant (#2453 ) * Create tests for ck tile batched transpose using example * Create ck tile tests for smoothquant using examples * fix precision input strings and convert batched transpose to regression tests * Code cleanup and fix asserts * add missing licenses * update copyright and licensing in files * Update smoothquant tests to use example's smoothquant.cpp * Add custom target for batched transpose tests * Add missing new lines at end of files for CMakelists * fix typo in batched transpose CMakeList target_compile_options --------- Co-authored-by: root <root@ctr-ubbsmc16.amd.com> [ROCm/composable_kernel commit: `c08986b026`]	2025-07-17 09:53:34 -06:00

linqunAMD

01b8e6a80f

[CK_TILE] Refine Generic2dBlockShape to fix ck_tile example 2,10,11,14 on rdna3 and 4 (#2795 )

BlockWarps, WarpTile in Generic2dBlockShape are wave size dependent, it causes mangled name mismatch between host and device side.

Solution: Replace them with ThreadPerBlock and move BlockWarps, WarpTile calculation into Generic2dBlockShape

[ROCm/composable_kernel commit: c254f3d7b4]

2025-09-10 08:29:20 +08:00

Emily Martins

70f25296e6

[CK_Tile] Refactor MOE Sorting and Smoothquant ctests to gtests (#2596 )

* refactor moe_sorting ctests to use gtest framework

* Refactor ctests for smoothquant to gtests

* fix clang format to use version 18

* Print local_eid in MOE sorting gtests

* Remove extra space in smoothquant output

[ROCm/composable_kernel commit: 70dce4e0c6]

2025-08-14 10:54:57 -07:00

Emily Martins

356180d6a3

Tests for CK Tile Batched Transpose and Smoothquant (#2453 )

* Create tests for ck tile batched transpose using example

* Create ck tile tests for smoothquant using examples

* fix precision input strings and convert batched transpose to regression tests

* Code cleanup and fix asserts

* add missing licenses

* update copyright and licensing in files

* Update smoothquant tests to use example's smoothquant.cpp

* Add custom target for batched transpose tests

* Add missing new lines at end of files for CMakelists

* fix typo in batched transpose CMakeList target_compile_options

---------

Co-authored-by: root <root@ctr-ubbsmc16.amd.com>

[ROCm/composable_kernel commit: c08986b026]

2025-07-17 09:53:34 -06:00

3 Commits