Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

clean up the workspace after every stage #909

Merged
merged 1 commit into from
Sep 12, 2023
Merged

clean up the workspace after every stage #909

merged 1 commit into from
Sep 12, 2023

Conversation

illsilin
Copy link
Collaborator

This will make sure that the workspace folders get deleted from the slave nodes after each stage.

I have verified that the branch with these changes leaves the /data/workspace/ or /var/jenkins/workspace/ folders clean after each stage is completed.

@illsilin illsilin merged commit 56c0279 into develop Sep 12, 2023
@illsilin illsilin deleted the lwpck-918 branch September 18, 2023 17:16
shaojiewang added a commit to shaojiewang/composable_kernel that referenced this pull request Oct 12, 2023
* Fixed fp8 gemm (ROCm#882)

* add generic instances; fixed initi with fp8

* fixed comment

---------

Co-authored-by: Jing Zhang <[email protected]>

* added padding of K into gemm_v2r3 (ROCm#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 <[email protected]>

* Redesign the DPP8 GEMM kernel to use warp-wise component (ROCm#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

* [Navi3x] Add fp16/int8 wmma conv forward instances (ROCm#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 <[email protected]>
Co-authored-by: Chao Liu <[email protected]>

* Enable DPP8 GEMM on Navi3 (ROCm#892)

* Add codeowners for documentation (ROCm#902)

Co-authored-by: samjwu <[email protected]>

* Add new instances and support for small cases in DPP8 GEMM (ROCm#896)

* clean up the workspace after every stage (ROCm#909)

* Refactor f8_t, add bf8_t (ROCm#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

* Add a switch to build DL kernels and build them with staging compiler. (ROCm#907)

* enable building DL kernels with the daily staging compiler

* move the DL_KERNELS flag to another function

* fixed fp8 issues (ROCm#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 <[email protected]>

* Add grouped conv bwd weight dl instances and new layout (ROCm#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

* [Cmake] Set cmake default build type Release and path to /opt/rocm (ROCm#914)

* change the cmake update method (ROCm#918)

* Add  fp16/fp8 support into Grouped gemm FixedNK (ROCm#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 <[email protected]>

* Stylistic improvements for grouped convolution code

Remove unnecessary ignoring

Update test/grouped_convnd_bwd_weight/test_grouped_convnd_bwd_weight.cpp

* Add native conversions fp8<->fp32 (ROCm#908)

* Add native conversions

* Add bf8 conversions

* Fix DL GEMM instances with too large vector size (ROCm#901)

* Fix vector lengths of DL GEMM instances with padding
* Add checks for correctness of vector lenghts in DL GEMM

* fix the ckprofiler package build in a loop (ROCm#926)

* update to rocm5.7 by default (ROCm#925)

* update to rocm5.7 by default

* fix jenkinsfile syntax

* fix the building of the amd-stg-open compiler (ROCm#927)

* Refactoring cmake files to build data types separately. (ROCm#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

* Update naming (ROCm#937)

* Add 3d grouped conv fwd wmma instances (ROCm#935)

* Add 3d grouped conv fwd wmma instances

* Refactor fwd conv tests

* Split wmma instances for each specialization

* Minor stylistic fixes

* Resolve some data type issues and cmake policy. (ROCm#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

* split ckProfiler gfx9 package into gfx90 and gfx94 (ROCm#946)

* Add fp8 gemm instances (ROCm#920)

* Add fp8 gemm instances

* Update instance naming

* Fixed Gemmv2r3 kpad (ROCm#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 <[email protected]>

* Use lower case for ckprofiler package. (ROCm#948)

* split ckProfiler gfx9 package into gfx90 and gfx94

* use lower case for package names

* Add multiple A/B support (ROCm#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 <[email protected]>

* Add column to image kernel (ROCm#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 728b869.

* Instances reduction

* Add comments in device_column_to_image_impl

* Update changelog and Copyrights

* Improve changelog

* Handle type conversions to a const datatype (ROCm#944)

* Handle type conversions to a const datatype

* Review: Handle X being const data type as well

* Review: Remove typo

* Fix gemm_splitk test, add hip_check_error after kernel calls in kernel_launch. (ROCm#951)

* Added error check after kernel launch (ROCm#919)

Co-authored-by: Xiaodong Wang <[email protected]>
Co-authored-by: Xiaodong Wang <[email protected]>

* remove M=0 test cases for test_gemm_splitk

---------

Co-authored-by: Xiaodong Wang <[email protected]>
Co-authored-by: Xiaodong Wang <[email protected]>

* Add grouped convolution changes to changelog (ROCm#952)

* Add grouped convolution changes to changelog

* Fix 0.2.0 ck release rocm version

* Suggested CHANGELOG.md edits

* Update CHANGELOG.md

* Update CHANGELOG.md

* Update CHANGELOG.md

* Update CHANGELOG.md

* Update CHANGELOG.md

* Update CHANGELOG.md

---------

Co-authored-by: Lisa <[email protected]>

* Add grouped conv bwd data wmma (ROCm#950)

* Add grouped conv bwd data wmma

* Fix copyrights

* Add instances with smaller NPerBlock

* Update interface test

* Minor stylistic fixes

* Minor stylistic fixes

* Add support for mixed precision in contraction scale and bilinear (ROCm#936)

* Extract common functionality to separate files

* Reference contraction: Remove incorrect consts from type_converts

* Reference contraction: Add missing type_convert for dst value

* Reference contraction: Fix incorrect order of B matrix dimensions

* Add support for mixed precision in contraction scale and bilinear

* Move using statements from instances to a common file

* Move using statements from examples to a common file

* Fix the order of B matrix dimensions across examples and profiler

* Fix the computation of error threshold

* Make ComputeDataType an optional argument

* Include possible DataType -> ComputeDataType casting error in the threshold

* Remove commented code

* add gfx942 target to the daily ckprofiler package (ROCm#955)

* Contraction multi abd (ROCm#957)

* 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

* init commit for contraction_multi_ABD

* add examples

* add examples of multiA and broadcast

* update example

* fixed comments

* Update cmake-ck-dev.sh

* Update cmake-ck-dev.sh

* Add comments into the example

---------

Co-authored-by: Jing Zhang <[email protected]>

* get rid of gfx900/906, set rocm5.7 as default (ROCm#958)

* Add fp8 @ bf8 gemm support and example (ROCm#933)

* Add f8 bf8 gemm example

* Add element-wise ops

* Add intrinsics

* Update reference calculation

* Add an additional type option for xdlops gemm

* Fix build process

* Add bf8 to buffer addressing

* Update blockwise op, split typeA and typeB

* Update for compatibility

* Uppdate naming to f8->fp8

* Update naming

* Format

* add generic instances (ROCm#947)

Co-authored-by: Jing Zhang <[email protected]>

* Fixed contraction issues (ROCm#960)

* add missing ComputeType

* fixed

* Update cmake-ck-dev.sh

---------

Co-authored-by: Jing Zhang <[email protected]>

* changed test for grouped_gemm to be random (ROCm#959)

Co-authored-by: Jing Zhang <[email protected]>

* 3d grouped conv fwd with input/output fp16 and comp fp8 (ROCm#931)

* add f8 comp instance

* fixed

* fixed comments

* rename

* fixed dtype

* format

* fixed CI

* fixed ci

* add missing ComputeType

* fixed cit

* fixed

* Update cmake-ck-dev.sh

---------

Co-authored-by: Jing Zhang <[email protected]>

* Add conv bwd weight fp16 comp bf8 fp8 op, instances and example (ROCm#945)

* Add f8 bf8 gemm example

* Add element-wise ops

* Add intrinsics

* Update reference calculation

* Add an additional type option for xdlops gemm

* Fix build process

* Add bf8 to buffer addressing

* Update blockwise op, split typeA and typeB

* Update for compatibility

* Uppdate naming to f8->fp8

* Update naming

* Format

* Update naming (ROCm#937)

* Add a client example

* Add computetypes to device and gridwise ops

* Add instances, update instance factory

* Format

* Fix a flag

* Add ckProfiler mode

* Fix typos

* Add an example

* Add bf8 generator

* add bf8 mfma; fixed type_convert for bf8

* move verfication ahead of timing

* Update reference calculation

* Fix reference

* Narrow down float init range

* Fix bf8 bf8 mfma

* Add bf8 @ fp8 mfma

* Update example

* Update instances

* Update profiler api

* Update for compatibility

* Format

* Remove extra example

* Clean up

* workaround convert

---------

Co-authored-by: Jing Zhang <[email protected]>

* Grouped conv bwd data with fp16 input and bf8fp8 comp (ROCm#962)

* Add f8 bf8 gemm example

* Add element-wise ops

* Add intrinsics

* Update reference calculation

* Add an additional type option for xdlops gemm

* Fix build process

* Add bf8 to buffer addressing

* Update blockwise op, split typeA and typeB

* Update for compatibility

* Uppdate naming to f8->fp8

* Update naming

* Format

* Update naming (ROCm#937)

* Add a client example

* Add computetypes to device and gridwise ops

* Add instances, update instance factory

* Format

* Fix a flag

* Add ckProfiler mode

* Fix typos

* Add an example

* Add bf8 generator

* add bf8 mfma; fixed type_convert for bf8

* move verfication ahead of timing

* Update reference calculation

* Fix reference

* Narrow down float init range

* Fix bf8 bf8 mfma

* Add bf8 @ fp8 mfma

* Update example

* Update instances

* Update profiler api

* Update for compatibility

* Format

* Remove extra example

* Clean up

* workaround convert

* added instance of f16_bf8f8, and client example

* fixed mfma selector

* format

---------

Co-authored-by: Rostyslav Geyyer <[email protected]>
Co-authored-by: Rostyslav Geyyer <[email protected]>
Co-authored-by: Jing Zhang <[email protected]>

* remove example 60 (ROCm#963)

Co-authored-by: Jing Zhang <[email protected]>

* Revert "Add support for mixed precision in contraction scale and bilinear" (ROCm#967)

* Revert "Add support for mixed precision in contraction scale and bilinear (ROCm#936)"

This reverts commit f074850.

* revert commits ROCm#957 and ROCm#960

* Replace CMake `return` from later CMake (ROCm#970)

* Fixed f8_gemm NaN (ROCm#975)

* workaround nan problem by changing output to fp16

* enable f8/bf8 gemm tests on MI200

* workaround f16 to f8 conversion

---------

Co-authored-by: Jing Zhang <[email protected]>

* Fix MNKPadding in gridwise_gemm_xdlops_v2r3 (ROCm#981)

* Grouped Gemm with looping over the tiles. (ROCm#788)

* Introduce LocalBlockToCTileMap.

* Change the signature of CalculateBottomIndex() function which now does
not accept any argument. The B2C map which is already passed as an
argument to the kernel Run function is calculating block's local id
already outside at kernel entry point __global__ function.
The LocalB2C map stores as members local block ID.

* Use LocalBlockToCTile map in device ops.

* First draft of tile loop work distribution.

* Fix typo.

* Simplify kernel arguments.

Calculate descriptors & B2C maps on the device.

* Use looping kernel.

* Fix B2C constructor.

* Fix Navi21 errors.

* Calculate tile start/end in device kernel.

* Change Run API to accept user provided workspace buffer.

* Add new line at EOF.

* Move Gemm KernelArguments to device op interface.

* Remove unused code.

* Update API.

* Launch grid size which is min of occupancy vs tile count

* Get back to use constant memory for gemm descriptors.

* Remove unused code.

* Add default virtual method implementation.

* Update comments to conform with doxygen style.

* Fix doc style and unused parameters.

* Add thread cluster lengths to kernel name.

* Remove old splitk impl and replace it with tile looping one.

* Modify instances.

* set KPerBlock to 64
* maximize wherever possible vector load size.

* Fix instances cluster lengths.

* Change comment style.

* Use 128b store where possible in instances.

* Update test cases, since KPerBlock has doubled.

* Update output stream operator for Sequence.

* Add pipeline version to GroupedGEMM device op type string.

* Fix pipeline version type logging.

* Fix input tensors type after merge.

* Fix compiler error.

* Fix output stream operator for Pipeline version.

* Store using 128b.

* Set of instances with kpb 32/64

* Limit number of instances

* Remove commented out instances.

* Fix function name.

* Limit the number of instances.

Add pipline version to the regular instances

* Change thr cluster layout for reading B tensor.

* disabled failed instances

---------

Co-authored-by: Adam Osewski <[email protected]>
Co-authored-by: zjing14 <[email protected]>
Co-authored-by: Jing Zhang <[email protected]>

* Revert "Grouped Gemm with looping over the tiles. (ROCm#788)" (ROCm#982)

This reverts commit a4f72a3.

* simplified buffer_load/store (ROCm#971)

* simplified buffer_load/store

* add bfp8/fp8

* fixed

* fixed buffer_load

* fixed buffer_store

---------

Co-authored-by: Jing Zhang <[email protected]>

---------

Co-authored-by: zjing14 <[email protected]>
Co-authored-by: Jing Zhang <[email protected]>
Co-authored-by: Bartlomiej Wroblewski <[email protected]>
Co-authored-by: Haocong WANG <[email protected]>
Co-authored-by: Chao Liu <[email protected]>
Co-authored-by: Sam Wu <[email protected]>
Co-authored-by: samjwu <[email protected]>
Co-authored-by: Illia Silin <[email protected]>
Co-authored-by: Rostyslav Geyyer <[email protected]>
Co-authored-by: Bartłomiej Kocot <[email protected]>
Co-authored-by: Jun Liu <[email protected]>
Co-authored-by: Xiaodong Wang <[email protected]>
Co-authored-by: Xiaodong Wang <[email protected]>
Co-authored-by: Lisa <[email protected]>
Co-authored-by: Rostyslav Geyyer <[email protected]>
Co-authored-by: Lauren Wrubleski <[email protected]>
Co-authored-by: Adam Osewski <[email protected]>
Co-authored-by: Adam Osewski <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants