ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-14 10:09:41 +00:00
Code Issues Packages Projects Releases Wiki Activity
1,015 Commits 742 Branches 38 Tags
9bc92adde37fe09118fd6b17f32a30acae2fa7cd
Commit Graph

1015 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Bartłomiej Kocot
9bc92adde3 Add column to image kernel (#930) 
* Add column to image kernel

* Minor fixes for dtypes and client examples

* Disable tests for disabled dtypes

* Disable add instances functions for disabled data types

* Minor stylistic fixes

* Revert "Disable add instances functions for disabled data types"

This reverts commit 728b869563.

* Instances reduction

* Add comments in device_column_to_image_impl

* Update changelog and Copyrights

* Improve changelog

[ROCm/composable_kernel commit: e2243a4d1e]
 2023-09-27 17:19:06 +02:00
zjing14
e21f21022c Add multiple A/B support (#906) 
* add gridwise_multi_abd

* move element_op into RunRead

* merge element_wise op with data read

* add multiABD example

* allow packed elementwise_op

* changed example

* clean

* clean

* add is_detected

* fix

* minor fix

* add scaleAdd_vec4 example

---------

Co-authored-by: Jing Zhang <jizha@amd.com>

[ROCm/composable_kernel commit: 11676c7e49]
 2023-09-26 21:16:23 -05:00
Illia Silin
832e7a6941 Use lower case for ckprofiler package. (#948) 
* split ckProfiler gfx9 package into gfx90 and gfx94

* use lower case for package names

[ROCm/composable_kernel commit: 420b5a0382]
 2023-09-26 17:43:09 -07:00
zjing14
6469878b49 Fixed Gemmv2r3 kpad (#938) 
* added kpad support into v2r3

* add generic instances

* fixed comments

* fixed mnk padding

* Update device_batched_gemm_xdl.hpp

* fixed kpad

---------

Co-authored-by: Jing Zhang <jizha@amd.com>

[ROCm/composable_kernel commit: 48ba6e8a69]
 2023-09-26 18:40:00 -05:00
Rostyslav Geyyer
9f2256f867 Add fp8 gemm instances (#920) 
* Add fp8 gemm instances

* Update instance naming

[ROCm/composable_kernel commit: 94bfa50256]
 2023-09-26 14:59:33 -05:00
Illia Silin
f8ae1bc9b5 split ckProfiler gfx9 package into gfx90 and gfx94 (#946) 
[ROCm/composable_kernel commit: 0b296a2722]
 2023-09-26 11:22:31 -07:00
Illia Silin
8bcd830114 Resolve some data type issues and cmake policy. (#940) 
* split the types in gemm_bilinear instances, add condition to cmake policy

* fix syntax

* split the data types in batchnorm examples

* fix the batchnorm_bwd test

* fix types in the batchnorm_bwd test

[ROCm/composable_kernel commit: 2ea75bd6d7]
 2023-09-26 08:39:11 -07:00
Bartłomiej Kocot
a85bb57471 Add 3d grouped conv fwd wmma instances (#935) 
* Add 3d grouped conv fwd wmma instances

* Refactor fwd conv tests

* Split wmma instances for each specialization

* Minor stylistic fixes

[ROCm/composable_kernel commit: c95538325b]
 2023-09-23 18:56:31 +02:00
Rostyslav Geyyer
c5a26b77ea Update naming (#937) 
[ROCm/composable_kernel commit: ede64ae9db]
 2023-09-22 10:08:45 -05:00
Illia Silin
43098d2a23 Refactoring cmake files to build data types separately. (#932) 
* refactor cmake files for the tests

* refactor cmake files for examples

* fix cmake for gemm example

* fix the cmake file for all examples

* add splitting by data types in gemm_splitk instance header

* rename test to reflect only dl instances are used

* clean up CI workspace, update cmake for instances

* change the jenkinsfile syntax

* build all instances except DL on gfx11

* move workspace cleanup after stages

* clean up workspace after every stage

* isolate data types in grouped_conv_fwd header

* isolate dl instances for grouped_conv2d_fwd

* fix syntax

* fix cmake and batchnorm instances

* fix typo

* fix reduction instances

* fix grouped_conv headers

* fix syntax

* replace parsing logic for instances, replace bfp16 with bf16

* fix the client examples build

* clean up DTYPES from instances cmake files

* update the parsing logic in cmake files

* make an exception for reduction kernels

* update few remaining cmake files to handle DTYPES

* fix syntax

* fix cmake conflicts

* replace f8 with fp8 test name

* resolve conflicts for dpp instances

[ROCm/composable_kernel commit: bba085d2b5]
 2023-09-20 22:15:56 -07:00
Illia Silin
2f7d38e8bf fix the building of the amd-stg-open compiler (#927) 
[ROCm/composable_kernel commit: 58817bf967]
 2023-09-19 18:50:58 -07:00
Illia Silin
828f8e5554 update to rocm5.7 by default (#925) 
* update to rocm5.7 by default

* fix jenkinsfile syntax

[ROCm/composable_kernel commit: 718065ebd2]
 2023-09-19 09:35:45 -07:00
Illia Silin
c469df8124 fix the ckprofiler package build in a loop (#926) 
[ROCm/composable_kernel commit: 5a4416c8a7]
 2023-09-19 09:17:39 -07:00
Bartlomiej Wroblewski
f26463b810 Fix DL GEMM instances with too large vector size (#901) 
* Fix vector lengths of DL GEMM instances with padding
* Add checks for correctness of vector lenghts in DL GEMM

[ROCm/composable_kernel commit: 63cd459248]
 2023-09-18 14:08:23 +02:00
Rostyslav Geyyer
ecc69920c3 Add native conversions fp8<->fp32 (#908) 
* Add native conversions

* Add bf8 conversions

[ROCm/composable_kernel commit: f17af2e9ed]
 2023-09-17 20:56:27 -05:00
Bartlomiej Kocot
775ae67af3 Stylistic improvements for grouped convolution code 
Remove unnecessary ignoring

Update test/grouped_convnd_bwd_weight/test_grouped_convnd_bwd_weight.cpp


[ROCm/composable_kernel commit: bc2d0583d3]
 2023-09-15 20:03:47 +02:00
zjing14
a49f4ff995 Add fp16/fp8 support into Grouped gemm FixedNK (#874) 
* move all arguments into device

* add b2c_tile_map

* add examples

* add SetDeviceKernelArgs

* dedicated fixed_nk solution

* init client api

* add grouped_gemm_bias example

* add a instance

* add instances

* formatting

* fixed cmake

* Update EnableCompilerWarnings.cmake

* Update cmake-ck-dev.sh

* clean; fixed comments

* fixed comment

* add instances for fp32 output

* add instances for fp32 output

* add fp32 out client example

* fixed CI

* init commit for kbatch

* add splitk gridwise

* format

* fixed

* clean deviceop

* clean code

* finish splitk

* fixed instances

* change m_loops to tile_loops

* add setkbatch

* clean code

* add splitK+bias

* add instances

* opt mk_nk instances

* clean examples

* fixed CI

* remove zero

* finished non-zero

* clean

* clean code

* optimized global_barrier

* fixed ci

* fixed CI

* instance and client

* removed AddBias

* format

* fixed CI

* fixed CI

* move 20_grouped_gemm to 21_grouped_gemm

* clean

* formatting

* clean

* clean

* fixed computeType

---------

Co-authored-by: Jing Zhang <jizha@amd.com>

[ROCm/composable_kernel commit: f9d0eddb90]
 2023-09-14 21:04:10 -05:00
Illia Silin
5ee23dd30d change the cmake update method (#918) 
[ROCm/composable_kernel commit: 0d8efaa13d]
 2023-09-14 09:36:26 -07:00
Jun Liu
4b789f8934 [Cmake] Set cmake default build type Release and path to /opt/rocm (#914) 
[ROCm/composable_kernel commit: 5fe687fa27]
 2023-09-13 14:38:12 -07:00
Bartłomiej Kocot
a13a1cb6d3 Add grouped conv bwd weight dl instances and new layout (#897) 
* Add grouped conv bwd weight dl instances and new layout

* Add M and N padding

* Remove todo comment

* Enable grouped conv fwd dl k,c=1 generic instance

* Comment fixes

[ROCm/composable_kernel commit: 475188ca2e]
 2023-09-13 10:14:31 -05:00
zjing14
8af8296c29 fixed fp8 issues (#894) 
* fixed fp8 init; and reference gemm

* Update host_tensor_generator.hpp

* fixed convert

* fixed reference gemm

* fixed comments

* fixed comments

* fixed ci

* fixed computeType

---------

Co-authored-by: Jing Zhang <jizha@amd.com>

[ROCm/composable_kernel commit: a66d14edf2]
 2023-09-12 22:17:56 -05:00
Illia Silin
5b9d3268d7 Add a switch to build DL kernels and build them with staging compiler. (#907) 
* enable building DL kernels with the daily staging compiler

* move the DL_KERNELS flag to another function

[ROCm/composable_kernel commit: 74d32f0719]
 2023-09-12 20:14:33 -05:00
Rostyslav Geyyer
2e227b8581 Refactor f8_t, add bf8_t (#792) 
* Refactor f8_t to add bf8_t

* Add check_err impl for f8_t

* Update fp8 test

* Format

* Revert the fix

* Update vector_type implementation

* Add bf8 test

* Add bf8, use BitInt types

* Add bf8 conversion methods

* Update type_convert for fp8/bf8

* Add check_err fp8/bf8 support

* Add subnorm fp8 tests

* Add subnorm bf8 tests

* Fix conversion

* Add bf8 cmake bindings

* Add macros to enable build with disabled fp8/bf8

* Remove is_native method

* Update flag combination for mixed precision instances

* Add more flag checks

* Add another flag to a client example

* Add type traits, decouple f8/bf8 casting

* Clean up

* Decouple fp8 and bf8 flags

* Remove more redundant flags

* Remove leftover comments

[ROCm/composable_kernel commit: 62d4af7449]
 2023-09-12 17:04:27 -05:00
Illia Silin
e885110f62 clean up the workspace after every stage (#909) 
[ROCm/composable_kernel commit: 56c0279bbd]
 2023-09-12 08:57:12 -07:00
Bartlomiej Wroblewski
df01b7c45a Add new instances and support for small cases in DPP8 GEMM (#896) 
[ROCm/composable_kernel commit: 547dbcfbc2]
 2023-09-12 10:05:23 -05:00
Sam Wu
a41a9fa3cd Add codeowners for documentation (#902) 
Co-authored-by: samjwu <samjwu@users.noreply.github.com>

[ROCm/composable_kernel commit: 85e2e1e2e2]
 2023-09-11 11:01:36 -06:00
Bartlomiej Wroblewski
20d295b14b Enable DPP8 GEMM on Navi3 (#892) 
[ROCm/composable_kernel commit: 8f84a01237]
 2023-09-08 11:14:57 -05:00
Haocong WANG
68d7430ec5 [Navi3x] Add fp16/int8 wmma conv forward instances (#746) 
* fix wmma gemm int8; add grouped conv int8 example

* Add int8 gemm-bilinear instances

* compile sanity check unknown

* Sanity pass + clang-format

* add int8 conv profiler instances

* solve merge conflict

---------

Co-authored-by: zjing14 <zhangjing14@gmail.com>
Co-authored-by: Chao Liu <chao.liu2@amd.com>

[ROCm/composable_kernel commit: 562b4cec48]
 2023-09-07 21:59:26 -05:00
Bartlomiej Wroblewski
88bb9d5fac Redesign the DPP8 GEMM kernel to use warp-wise component (#863) 
* Redesign the DPP8 GEMM kernel to use warp-wise component

* Review: Improve error messages

* Review: Remove unnecessary empty lines

* Review: Fix M, N per thread names

* Review: Rename mfma_input_type to dpp_input_type

* Review: Fix tensor adaptor; remove unnecessary element

* Review: Remove calls to dpp_gemm's MakeCDescriptor

* Review: Add blockwise doc, change function names to include dimension names

* Review: Remove duplicated code; Move Block2CtileMap alias to the top of the file

* Review: Add __restrict__ keywords

* Review: Use MatrixPadder for padding A, B, C matrices

* Review: Remove hardcoded datatypes

* Review: Change names from FloatX to XDataType

* Review: Introduce AK0 and BK0 instead of a single K0

* Review: Remove construction of dpp_datatypes object

* Review: Rename DppInstrRunner to DppLanegroupGemm

[ROCm/composable_kernel commit: 37a8c1f756]
 2023-09-06 11:44:09 -05:00
zjing14
3446ff1e7d added padding of K into gemm_v2r3 (#887) 
* added kpad support into v2r3

* add generic instances

* fixed comments

* fixed mnk padding

* Update device_batched_gemm_xdl.hpp

---------

Co-authored-by: Jing Zhang <jizha@amd.com>

[ROCm/composable_kernel commit: 3786bfe1cc]
 2023-09-06 10:15:52 -05:00
zjing14
88af65157c Fixed fp8 gemm (#882) 
* add generic instances; fixed initi with fp8

* fixed comment

---------

Co-authored-by: Jing Zhang <jizha@amd.com>

[ROCm/composable_kernel commit: a61b8b785e]
 2023-09-06 09:59:20 -05:00
Illia Silin
a9d5a61e90 set warnings as errors in doxygen (#864) 
[ROCm/composable_kernel commit: aae4df5596]
 2023-09-05 14:29:37 -07:00
Bartlomiej Wroblewski
423aab15f6 Add contribution guidelines to the documentation (#843) 
Add contribution guidelines to the documentation

[ROCm/composable_kernel commit: 1e1f82d9b0]
 2023-09-05 21:25:28 +02:00
Illia Silin
3b3f0f225e fix syntax (#890) 
[ROCm/composable_kernel commit: 7dcb14d9d4]
 2023-09-05 11:29:44 -07:00
Bartłomiej Kocot
2ec7a9084a Add image to column kernel (#867) 
* Add image to column kernel

* Add instances, tests, profiler, example

* Add client example

* Several fixes of image to column

* Fix variable name in device_image_to_column_impl

* Several fixes of image to column profiler

* Fix num_btype calculation

* Make new mesaurements for correct bytes calculation

[ROCm/composable_kernel commit: 0077eeb3be]
 2023-09-05 10:11:40 -05:00
Bartłomiej Kocot
f64298eef4 Add nhwgc dl generic instances for grouped conv fwd (#879) 
[ROCm/composable_kernel commit: 0c9a1d25b3]
 2023-09-05 10:07:56 -05:00
Bartłomiej Kocot
fcafba0fd4 Fix K padding calculation for grouped conv data (#876) 
* Fix K padding calculation for grouped conv data

* Restore previous padd for 1x1 specialization

[ROCm/composable_kernel commit: c981f6d033]
 2023-09-05 10:07:41 -05:00
Lauren Wrubleski
47958ebd07 Fix config header installation (#880) 
[ROCm/composable_kernel commit: bd8024b84a]
 2023-09-04 09:49:40 -07:00
zjing14
bc88b1a50b Grouped Gemm with Fixed K and N with SplitK (#818) 
* move all arguments into device

* add b2c_tile_map

* add examples

* add SetDeviceKernelArgs

* dedicated fixed_nk solution

* init client api

* add grouped_gemm_bias example

* add a instance

* add instances

* formatting

* fixed cmake

* Update EnableCompilerWarnings.cmake

* Update cmake-ck-dev.sh

* clean; fixed comments

* fixed comment

* add instances for fp32 output

* add instances for fp32 output

* add fp32 out client example

* fixed CI

* init commit for kbatch

* add splitk gridwise

* format

* fixed

* clean deviceop

* clean code

* finish splitk

* fixed instances

* change m_loops to tile_loops

* add setkbatch

* clean code

* add splitK+bias

* add instances

* opt mk_nk instances

* clean examples

* fixed CI

* remove zero

* finished non-zero

* clean

* clean code

* optimized global_barrier

* fixed ci

* fixed CI

* removed AddBias

* format

* fixed CI

* fixed CI

* move 20_grouped_gemm to 21_grouped_gemm

---------

Co-authored-by: Jing Zhang <jizha@amd.com>

[ROCm/composable_kernel commit: f5ec04f091]
 2023-08-31 09:22:12 -05:00
rocking
e422d088a3 MaxPool & AvgPool bwd instances, test, ckProfiler, client example (#861) 
* Add maxpool instances

* Rename index pool to max pool.

* Add maxpool bwd bf16 instances

* Add avg pool bwd instances

* Rename avgpool and maxpool to avg_pool3d and max_pool

* Add bf16 pool fwd instances

* Add max pool bwd to ckProfiler

* Add avg pool3d bwd to ckProfiler

* Add avg pool bwd test

* Fix bug of reference pool fwd (dilation)

* Fix bug of max pool bwd  (dilation and initZero)

* Support bf16 compute data type

* Force compute type be f32. Because atomicAdd only support f32

* Add max pool bwd test

* Rename folder

* Rename pool

* Add max pool bwd client example

* Add avg pool bwd client example

* Add missing workspace

* clang format

* Rename macro

* remove useless header

* remove useless layout

[ROCm/composable_kernel commit: 866377de18]
 2023-08-31 21:01:50 +08:00
Illia Silin
620a9319de fix gemm_streamk example on mi300 (#875) 
[ROCm/composable_kernel commit: bf1912ed3d]
 2023-08-30 20:18:38 -07:00
Bartłomiej Kocot
bdfbeb5a26 Add number of error when fail (#868) 
[ROCm/composable_kernel commit: 9e86ebd62d]
 2023-08-30 10:33:11 -05:00
zjing14
d6005c1e0d add an example of customized type convert - bfp16_rtn (#869) 
* add an example of customized bfp16_rtn

* fixed threadwise_copy

---------

Co-authored-by: Jing Zhang <jizha@amd.com>

[ROCm/composable_kernel commit: 38ada109ea]
 2023-08-29 12:31:24 -05:00
zjing14
17d03c86d2 Fp16/fp8 mixed-precision Gemm with multiply+add fusion (#865) 
* add compute_type

* add multiply_add ckProfiler

* add f8_fp16 support

* clean

* clean

* fixed lds size calc

* format

---------

Co-authored-by: Jing Zhang <jizha@amd.com>

[ROCm/composable_kernel commit: 31ea132aa2]
 2023-08-28 16:27:32 -05:00
Jun Liu
a2392f1887 [HotFix] add config and version files to pass on build info (#856) 
* experiment with config file

* experiment with version.h config

* add more info to version.h

* minor updates

* minor updates

* fix case where DTYPE is not used

* large amount of files but minor changes

* remove white space

* minor changes to add more MACROs

* fix cmakedefine01

* fix issue with CK internal conflict

* fix define and define value

* fix clang-format

* fix formatting issue

* experiment with cmake

* clang format v12 to be consistent with miopen

* avoid clang-format for config file

[ROCm/composable_kernel commit: c8a8385fdd]
 2023-08-23 11:36:17 -07:00
Qianfeng
b0bd6e3895 Add workspace setting up for batchnorm bwd/fwd client examples (#860) 
[ROCm/composable_kernel commit: 350d64f351]
 2023-08-24 01:13:07 +08:00
Illia Silin
d81ec56b52 use correct data types in cmake conditions for splitk gemm example (#862) 
[ROCm/composable_kernel commit: 7c71dc7e70]
 2023-08-23 09:52:11 -07:00
zjing14
d91fc00427 add generic instances (#858) 
Co-authored-by: Jing Zhang <jizha@amd.com>

[ROCm/composable_kernel commit: 8ebea3a56e]
 2023-08-23 09:18:10 -05:00
zjing14
3d88622dda Ck profiler splitk (#857) 
* updated regular gemm

* update ckProfiler

* fixed gtests

---------

Co-authored-by: Jing Zhang <jizha@amd.com>

[ROCm/composable_kernel commit: ca3115e7e8]
 2023-08-22 16:54:34 -07:00
Bartłomiej Kocot
80659b5bc1 Fix transform and instances for grouped conv bwd data (#848) 
* Fix transform and instances for grouped conv bwd data

* Add instances for small K and small C

* Remove workaround after fix

* Fix interface tests

[ROCm/composable_kernel commit: 595d23be14]
 2023-08-22 11:25:41 -05:00