GitLab

* ad gelu and fast_gelu

* added GeLU and fast GeLU

* clean up

* add gemm+fastgelu example

* add gemm+gelu instances

* update profiler

* clean up

* clean up

* adding gemm+bias+activation

* clean

* adding bias

* clean

* adding gemm multiple d

* debugging

* add gemm bias add fastgelu

* rename, clean

* refactoring; add readme

* refactor

* refactor

* refactor

* refactor

* refactor

* refactor

* fix

* fix

* update example

* update example

* rename

* update example

* add ckProfiler

* clean

* clean

* clean

* clean

* add comment

* use type_convert

* clean

* clean element wise op

* use pre-built docker instead of building a new one

* try docker.image.pull

* change syntax in docker.image()

* add 30 min timeout

* increase timeout to 3 hours

* move performance tests to first stage for testing

* set image variable to the new container name

* update image name

* check available images

* check available images in both places

* try different image name

* use image ID to refer to image

* run performance on gfx90a

* fix the gpu_arch labeling, add parameter

* move env vars out of stages

* add stand-alone performance script, MI200 tests, CU numbers

* dos2unix for run_perf_tests.sh

* try the new git credentials

* use env var for git credentials

* don't look up /sys/module/amdgpu/version
Co-authored-by: Chao Liu <chao.liu2@amd.com>

* Remove template from Reducton operation classes and add template to their operator() and GetIdentityValue() interfaces

* Change to unary elementwise operators and the reduce_unary_operator (class for mapping) and dependent variations in all host layers

* Remove the data type template parameter from reduce_binary_operator (class for mapping) and dependent variations in host layers

* Add InMemoryDataOperatonSupportedOnDataType to check the matching between data type and InMemoryDataOperation

* Use struct-scope operator template instantiation for binary and unary element-wise operations

* Change a few more elementwise operations to use template for operator()

* Tiny correction in Normalize operator

* Add static_assert to check the data type appliability for some reduction accumulator and element-wise operatons

* Correction in some examples with regard to using ReduceAccDataType

* Use static_assert for UnaryDivide

* Update to merged codes to use Element-wise operations and Reduction Accumulator operations correctly

* Tiny fix with regard to SetWorkSpacePointer()

* Copy "gemm reduce" to "gemm bias add reduce"

* Implement gemm bias add reduction

* Fix compiler error due to merge from develop

* Add tensor operation for gemm + bias + add + reduce

* Add gemm_bais_add_reduce to ckProfiler

* Add c1 functor

* Refine type

* Use reduceAccDataType instead of explicitly float

* Change to use check_err()

* Do relu in float32 instead of bhalf_t. Because bhalf_t is unsigned

* Refactor relu. using type_trait instead of overloading

* Rename DxsReduceAccElementwiseOperation to DxsReduceAccElementwiseOperation

* Fix denominator

* Refine nameing

* Fix denominator  in host

* Remove useless include header

* Use AccDataType

* Fix static_cast order

* Refine type

* [What] Remove tuple type in the base class
[Why] External api depend on base class. if base class has relationship with type, we will need many class for different type

* add GetWorkSpaceSize to base arg and make an example on convnd_bwd_weight

* add bwd weight for bf16: init

* remove redundant compute

* use datatype and split k to check whether a workspace is used

* remove unused computation for work space size

* add some code for bfp16

* add device/grid unary op

* add unary type convert to bwd-weight example

* support bf16 splitk kernel for convnd bwd weight

* 1. remove comments. 2. add checkvalidity. 3. add gridsize computation

* add workspace size check

* fix format

* change function name

* use pre-built docker instead of building a new one

* try docker.image.pull

* change syntax in docker.image()

* add 30 min timeout

* increase timeout to 3 hours

* move performance tests to first stage for testing

* set image variable to the new container name

* update image name

* check available images

* check available images in both places

* try different image name

* use image ID to refer to image

* run performance on gfx90a

* fix the gpu_arch labeling, add parameter

* move env vars out of stages

* add stand-alone performance script, MI200 tests, CU numbers

* dos2unix for run_perf_tests.sh

* try the new git credentials

* use env var for git credentials

* use pre-built docker instead of building a new one

* try docker.image.pull

* change syntax in docker.image()

* add 30 min timeout

* increase timeout to 3 hours

* move performance tests to first stage for testing

* set image variable to the new container name

* update image name

* check available images

* check available images in both places

* try different image name

* use image ID to refer to image

* run performance on gfx90a

* fix the gpu_arch labeling, add parameter

* move env vars out of stages

* add stand-alone performance script, MI200 tests, CU numbers

* add resnet50 test to performance tests

* add blanks before gpu_arch in log files

* add resnet50 test with N=4 and process its results

* add ROCM and HIP versions to test tables

* uncomment the sql queries

* fix script syntax in jenkinsfile

* Use the unified naming for math functions on host and HIP kernel

* Corresponding change/simplification in reduction host/profiler/examples due to unified math functions renaming

* Renaming GetReductionZeroVal() to GetIdentityValue()

* Tiny renaming in profile_reduce_impl.hpp

* More renaming in profile_reduce_impl.hpp

* Replace zeroVal by identiyVal

* Remove ck_ prefix in the naming of ck::math provided functions

* moved gemm_descs_args into const buff

* use CK_CONSTANT_ADDRESS_SPACE instead of global constant

* clean

* moved hipMemAlloc outside of deviceOp

* add SetWorkSpacePointer

* fix ignore

* Reference CGEMM + test stub

* Format.

* Incomplete simple implementation

* Library instances

* Sketch of tests

* Test fixes.

* Example added

* Cosmetics

* Add elementwise operation kernel and example

* Add comment

* Add template argument of dim . Prepare to support multiple dimension

* Rename example

* Support 1 dimension

* Add static assert

* Add comment

* Second auxiliary buffer added

* Extract pad

* Remove redundant argument

* Support any dimension for elementwise operation

* Remove line

* Let it be the multiple number of CU

* Move thread per block to the parameter of constructor

* Consuming binary ops to do A+B / A-B

* Fix + cosmetics + bf16 test commented out temporarily

* Format

* Enabling bf16 test

* Revert "Enabling bf16 test"

This reverts commit f497e2ba.

* Fix + test reenabled

* fix build

* Revert "fix build"

This reverts commit d7310238

.

* post PR #235 merge fix

* amend

* Single workspace for cgemm + helper

* Perf calc fix

* Review remarks: static_cast

* Review remarks: binary ops templated

* Cleaning

* Removal of instances and their tests

* Review remarks from aosew addressed

* Review remark: unnecessary attribute

* Post-merge fixes

* Restrict 4gemm to PassThrough + bug fix

* Review remarks

* update licence

* change cgemm example to fp16
Co-authored-by: rocking <chunylai@amd.com>
Co-authored-by: Chao Liu <chao.liu2@amd.com>
Co-authored-by: Anthony Chang <ac.chang@outlook.com>

* fix example

* update IsSupportedArgument

* fix

* disable fp64 conv example as test

* Implement reduction meand and reduction square mean

* Refine file name

* Add reduce mean and square mean

* Fix parameter name

* Add normalize device op (not implement invoker::run())

* Remove epislon

* Refine deviceop

* Add 5ary elementwise for normalization

* Add layernorm example

* layerNorm verication

* Fix compiler error due to merge from develop

* Fix typo

* Fix compile error

* Refine naming

* [What] Suport non pointer for invoker and argument
[Why] Snyc coding style with gemm

* Refine folder name

* Refine class name

* Evaluate perf of the kernel

* Fix compile error

* [What] Refine perf evaluation in example of gemm + reduction
[Why] evaluation of gemm + reduction may cause verification fail. Because evaluation will not initial global memory

* clang-format

* debugging conv

* fix oversight where ctile map is constructed before initializing c desc

* example program should returns error code

* clean up

* changed Block2CTileMap in conv2d and convnd

* clean up

* clean up

* cleanup
Co-authored-by: Anthony Chang <ac.chang@outlook.com>

* add intrin_mfma_f64_16x16x4f64

* add example

* gemm reference add double data type

* chang init data

* fix M N PerXdlops

* fix ifdef

* add comparsion config

* add conv fwd example

* format log out

* change rc matrix egister layout

* reorganize example

* reorganize example 2

* format,because merge develop

* fix call impl adding acc data type

* lost ;

* add compiler warning

* change example tunning parameters

* add test for fp64

* add instance

* add test/gemm/gemm_fp64.cpp

* fix get name issue

* remove some tunning parameter

* fix conflict

* format

* use integer value for GEMM test

* add acc data type

* remove typeid because fp16

* fix streamconfig etc bug from merging develop

* format

* remove test_gemm_xdl_fp64

* add AccDataType

* AccDataType problem
Co-authored-by: qinletao <letaoqin@amd.com>
Co-authored-by: Chao Liu <chao.liu2@amd.com>

* Add example for computing LayerNorm mean and meansquare

* Refactor the pool2d_fwd example and add example for float type testing

* Revert "Add example for computing LayerNorm mean and meansquare"

This reverts commit df52e6f9d897b00c981baa48f291450bcd60925d.

* Tiny fix in pool2d_fwd_common.hpp

* Support different length of ScalarPerVector

* Add example of broadcast on fastest axis

* Typo

* Refine fastest example

* Add dimension check

* Modify fastest broadcast example to 3d

* Enforce users give scalarPerVector explicitely

* 1. Add CscalarPerVedctor
2. Not only broadcast on fastest need to set scalarPerVector to 1

* Rename var

* Move IsScalarPerVectorValid() inside IsSupportedArgument()

* Separate GridDesc_M0 into A, B and C

* rename var

* Rename var of length
Co-authored-by: rocking <chunylai@amd.com>

* fix build

* Revert "fix build"

This reverts commit d7310238

.

* post PR #235 merge fix

* amend

* adds tensile-stype c-tile map

* make it dynamic version

* add k-split flavor tile map

* apply tensile-style tile map to all xdl gridwise gemms

* remove dead code
Co-authored-by: Chao Liu <chao.liu2@amd.com>

* minor fix

* clean

* start adding navi21 GEMM

* navi_gemm_km_kn_mn_fp32 compiles and passes one test.

* rename variables and functions in gridwise_gemm_dlops_v1r3

* add other 3 layouts; format instance

* adding more tuning parameters

add tuning parameters for other 3 layouts

* add gemm_dlops_f16

* tmp

* add dependence of DeviceGemm::IsSupportedArg() on arch

* minor changes

* minor changes

* minor changes

* minor changes

* minor changes

* minor changes

* minor changes

* push gemm_dlops into profiler

* minor changes

* if using xdl or dlops is moved into profiler_gemm_impl

* minor changes

* minor changes

* remove is_xdl from profile_gemm_impl

* make IsSupportedArg dependent on arch for other device_gemm

* minor changes

* minor changes

* fix a bug in f_generate_tensor_value

* add 64x64x64 for gemm_dlops_int8

* add 64x64x64 for gemm_dlops_int8

* comment out 3 layouts in gemm_dlops_int8; add 32x32x32 for gemm_dlops_int8; init A values to 1

* fix

* start fixing tuning parameters

* monir

* minor changes

* minor changes

* minor changes

* fixing

* adding example

* adding example

* adding example

* add gemm fp32 example

* clean up

* use 128x128x16 as MNK tile in navi21 gemm example

* bug fix

* fix test

* use new block c tile

* clean

* fix build
Co-authored-by: Chao Liu <chao.liu2@amd.com>
Co-authored-by: shaojiewang <wsjmessi@163.com>

* Tiny fix in dynamic_buffer.hpp to support vectorized AtomicAdd for double type

* Update to host layer and host reduction

* Merge and remove reduction kernels

* Merge and remove reduction device interfaces and update pooling device interface

* Merge and remove useless reduction device instances

* Update to reduction profiler and reduction ctests

* Update to reduction and pooling examples and add one reduction example

* Change to reduction examples to let them testable by ctest

* Add explicit pass checking for reduction and pooling examples

* Explicit assignment of tensor shapes in example reduce_blockwise_two_call

* Use atomic_add to repace atomicAdd and add atomic_add for double type

* Add reduce ctest support for double data type

* Replace to_int_vector() by using c++ std::vector::assign()

* Keep DeviceReduceThreadWise separated from DeviceReduceBlockWise

* Merge DeviceReduceBlockWise and DeviceReduceMultiBlockAtomicAdd into DeviceReduceMultiBlock

* Add GetAtomicOperationZeroValue() support for AtomicMax

* Tiny change to reduce example README.md

* Fix some tiny issues due to branch merging

* Revoke previous change in dynamic_buffer.hpp and add atomic_add for double2_t

* Add reduce multiblock_atomic_add instances for fp64 to verify vectorized atomic_add on fp64

* Renaming

* Clean the header includings in device_reduce instances header files

* modify ckProfiler_gemm output

* fix syntax

* change ckProfiler output and return 0

* fix syntax

* output datatype

* fix syntax

* output datatype in another way

* fix syntax

* fix syntax

* test return values of ckProfiler

* add layout info and tests, make sure ckprofiler returns 0

* fix syntax

* change layout output

* fix syntax

* fix syntax again

* update script to process perf results

* rearrange jenkins stages

* fix typo

* add python packages to Docker file

* adding setuptools-rust package

* modify parsing for new test parameters

* test db credentials on jenkins

* fix syntax

* update python script to handle incomplete lines

* ungrade python to 3.8 and write the gemm_params table

* add sqlalchemy package to docker

* move perf data processing to master node

* move the master node inside a steps region

* add new stage for result processing

* move results processing to separate stage

* reduce number of tests to speedup debugging

* pass config to processPerfResults stage

* run script on master in a docker container

* replace show_node_info

* try loading docker on master node again

* use ansible node instead of master

* get rid of pymysql package

* try ssh connection using paramiko

* put back pymysql

* put the perf data processing back on the gpu node

* put back artifact definition

* archive the perf_log before parsing

* clean up jenkinsfile, fix parsing

* fix typo

* enable all perf tests

* put all stages in original order, finalize script

* fix gpu_arch version

* update parsing script

* remove obsolete file causing merge conflict

* add GetWorkSpaceSize to base arg and make an example on convnd_bwd_weight

* remove redundant compute

* use datatype and split k to check whether a workspace is used

* remove unused computation for work space size

* fix build

* Revert "fix build"

This reverts commit d7310238

.

* post PR #235 merge fix

* amend
Co-authored-by: Anthony Chang <ac.chang@outlook.com>

* enable example of conv 1d/3d for bwd weight

* make bf16 kernel do not use atomic add

* using new gridwise gemm for bwd weight on convnd bwd weight
Co-authored-by: Chao Liu <chao.liu2@amd.com>

* refactor block-to-ctile-map

* gridwise gemm block2ctile generic validity check

* format

* amend split-k gemm block2ctile map refactor

* add test

* format

* amend

* revert to calculating batch index in kernel instead of passing as block_id_z

* move file

* add valid ctile index check to gridwise v2r4

* add some instance to develop

* avoid bank conflicts for wrw for all instance

* add small K1 test

* delete some unused instance

* binding gemm k1 to conv n

* try using half_4 to do ds_read

* reset buffer load oob and ds memcpy to default option

* remove useless instances

* remove redandunt space

* remove printf code

* clang-format-10 change

* use fastest config

* fix clang format for the other files

* remove gemmk0 pad for output

* add gemmk padding macro

* add bank length computation

* add template to distinguish the instance that need lds padding for wrw

* use rocm5.1 as docker

* use integer value for GEMM test

* add Right padding macro

* add 2 test asm code

* using 256x256x32 tile size

* 1. move dedicated transform into gridwisegemm's head file. 2. make lds tensor params a struct templete. 3. remove useless code

* using small vec

* 256*128 kernel size for example

* remove asm files

* use a new gridwise gemm header for bwd-weight

* revert gridwise gemm v2r4r2

* change foramt

* reset gridwise gemm v2r4r2

* remove unused code

* revert instance file

* revert example instance

* format file

* remove macros

* resolve compile error

* rename wrw kernel invoker

* use gridwisegemm pipeline struct instead of implement run fucntion in the same header
Co-authored-by: Chao Liu <chao.liu2@amd.com>

[Perf][Bwd-weights]Lds re-layout to avoid ds read/write bank conflict and balance ds ops with address calculations (#190)

* add some instance to develop

* avoid bank conflicts for wrw for all instance

* add small K1 test

* delete some unused instance

* reset buffer load oob and ds memcpy to default option

* remove useless instances

* remove redandunt space

* remove printf code

* clang-format-10 change

* fix clang format for the other files

* add bank length computation

* add template to distinguish the instance that need lds padding for wrw

* use rocm5.1 as docker

* use integer value for GEMM test

* 1. move dedicated transform into gridwisegemm's head file. 2. make lds tensor params a struct templete. 3. remove useless code

* use a new gridwise gemm header for bwd-weight

* revert gridwise gemm v2r4r2

* change foramt

* rename kernel invoker
Co-authored-by: Chao Liu <chao.liu2@amd.com>

* Use vector constructor instead

* Fix typo

* Move blockSize to the MakeArgumentPointer

* Fix naming

* Fix clang format

* remove blockSize from DeviceBinaryElementwise::Argument()
Co-authored-by: rocking <chunylai@amd.com>
Co-authored-by: Chao Liu <chao.liu2@amd.com>

* [What] Rename the example
[Why] Prepare to add unary reduction

* Add global oparation to the parameter

* Add atomicmax

* Fix compile error

* Support atomicMax (hip library)

* Rename the reduction example

* Fix target name

* use p_d1_grid as the indicator directly

* Prevent performance issue. Let passthrough handle it.

* Implement the function template the specialize the float2

* No need to separate into two lines

* Remove empty line

* add comment

* Fix compile error due to merge from develop

* make the implementation of atomic_max / atomic_add explicit for each datatype

* Refine typo

* For future CI test

* Fix compiler error in ckProfiler

* Merge commit 'de2769e3a6695b38a20529261273ddc5cdaab2fe'

* simply use remove_pointer

* Rename type and var

* Refine example

* Modify reducemax example

* Fix bug in reduction

* Change initialize range

* Implement F64 version of atomicMax

* Move reduction  code together

* Add buffer atomic_max

* Fix coding style by clang-format

* Integrate new api of DeviceGemmReduce_Xdl_CShuffle

* Integrate Batch gemm reduction

* Fix example

* fix example

* clean up

* Fix batch gemm tensor operation

* Fix coding style

* Fix template augument

* Fix clang format

* Keep flexible of different stride for each D tensor

* Fix compile error for ckProfiler

* Fix typo

* [What] Fix naming
[Why] Prepare to add out elementop

* Add DoutElementOp
Co-authored-by: Chao Liu <chao.liu2@amd.com>
Co-authored-by: rocking <chunylai@amd.com>

* Add elementwise operation kernel and example

* Add comment

* Add template argument of dim . Prepare to support multiple dimension

* Rename example

* Support 1 dimension

* Add static assert

* Add comment

* Extract pad

* Remove redundant argument

* Support any dimension for elementwise operation

* Remove line

* Let it be the multiple number of CU

* Move thread per block to the parameter of constructor

* rename threadPerBlock with blockSize

* Support double

* rename kernel function name

* remove redundant include header

* Refine type

* Need to the final dimension

* Refine variable name

* Refine type

* Use index_t instead of int in API
Co-authored-by: rocking <chunylai@amd.com>

* validate examples in ctest runs

* format

* fix usage of check_err

* amend

* add example codes to custom target 'check'
Co-authored-by: Chao Liu <chao.liu2@amd.com>

* Add host API

* manually rebase on develop

* clean

* manually rebase on develop

* exclude tests from all target

* address review comments

* update client app name

* fix missing lib name

* clang-format update

* refactor

* refactor

* refactor

* refactor

* refactor

* fix test issue

* refactor

* refactor

* refactor

* upate cmake and readme
Co-authored-by: Chao Liu <chao.liu2@amd.com>