Commit Graph

5 Commits

Author SHA1 Message Date
Khushbu Agarwal
3c39d279ab supporting prefill shapes for preshuffle block scale gemm (#2975)
* debugging

* debugging for prefill shapes

* comment unused code

* fix for prefill shapes

* clearing up the code

* add int4 to universal gemm example

* clang formatted

* adding test for prefill shapes in block scale gemm

* lil improv on the block pipeline

* Address Review Comment

---------

Co-authored-by: ThomasNing <thomas.ning@amd.com>
2025-10-10 15:36:24 -07:00
Cong Ma
1d4db30af9 [CK TILE GEMM] Refactor the code of transposeC and quantpreshuffle of AQuant Gemm (#2965)
Refactor the code of transposeC and quantpreshuffle of AQuant Gemm to make it easier to maintain.

Co-authored-by: Thomas Ning <Thomas.Ning@amd.com>
2025-10-08 00:05:38 -07:00
Cong Ma
6fc28ab493 [CK TILE GEMM] Support Aquant GEMM with transposeC and preshuffle (#2897)
* [CK TILE GEMM] Support Aquant GEMM with transposeC and preshuffle

When TransposeC and QuantPreshuffle are both true, Aquant generates
correct result.

* [CK TILE GEMM] Support Aquant GEMM with transposeC and preshuffle

- Add unit tests

* Fix bug in is_quantpreshuffle_enabled

* clang format

---------

Co-authored-by: ThomasNing <thomas.ning@amd.com>
2025-10-02 11:13:51 -07:00
Khushbu Agarwal
81458a6681 Weight Preshuffle Block Scale gemm support (#2877)
* initial commit

* remove extra files

* fixing errors

* updated ReadMe file for mapping of diff quants with diff configs

* addressing review comments

* addressing review comments

* Resolved merge conflicts

* [CK TILE GEMM] Replace get_preshuffle_or with is_quantpreshuffle_enabled

The get_preshuffle_or was not working as expected, which led to incorrect behavior
in the quantization preshuffle process. This change replaces it with the more reliable
is_quantpreshuffle_enabled function to properly determine when preshuffle should be applied.

* initial commit

* debugging

* working fp8 for init constant

* fp8 working with all inits

* updated block level code with comments

* changing the loop iter

* debugging

* debugging

* debugging

* code fix

* code clean up

* clang formatted

* Add comment

* code cleanup

* clang formatted

* merge conflicts fixes

* applying the latest int4 changes to the piepline

* fixing test code for updated traits

* Adding gtest

* review comments addressed

* addressing review comments

* remove c++20 code

* added flush cache changes

---------

Co-authored-by: Cong Ma <congma13@amd.com>
Co-authored-by: root <root@banff-cyxtera-s73-2.ctr.dcgpu>
2025-09-29 12:46:37 -07:00
Sami Remes
4363a82bd6 [CK_TILE] Tensor-wise scaled quant gemm kernel (#2846)
* rename gemm_group_quant to gemm_quant

* Add TensorWise quant mode

* Cshuffle epilogue tests with tensor scaling

* Add tensor quant to example

* Don't use readfirstlane for reading scales - doesn't work for some reason

* Add to changelog

* revert include - from a merge problem?

* revert common.hpp include

* revert host.hpp include

* remove unused utility function

* rename quant pipeline problem

* refactor quant tests

* remove aquant utils

* use TEST_F

* fix all tests by changing gemm config

* Use typed tests

* fix copyright
2025-09-19 16:52:35 -07:00