GitLab

* Refine the device batchnorm-backward base API templates and data type assignments

* Remove duplicated kernel file

* Add batchnorm backward instances and external API

* Add batchnorm-backward profiler and tests

* Add client example which uses batchnorm backward external API

* Merge test/batchnorm_fwd and test/batchnorm_bwd into one directory

* Loose the threshold for batchnorm-backward check_err()

* properly return error flag; reveals bug in split-k gemm

* fix bug in split k

* update split-k test case
Co-authored-by: Chao Liu <chao.liu2@amd.com>

* Implemented batchnorm-backward Blockwise and Multiblock kernels

* Add batchnorm-backward device op

* Add batchnorm-backward host-reference op

* Add batchnorm-backward example

* Parameters renaming in batchnorm backward kernels and device op

* Change in the example to loose the threshold for ScaleDiff checking

* Add comments to explain the implementation of batchnorm-backward

* Parameters renaming again in batchnorm backward kernels

* Improve the expression calculation for performance

* Add batchnorm backward to README

* Add comments to explain inv-variance in batchnorm forward and backward

* Renaming the batchnorm forward training and inferring examples

* Add/update the comments for batchnorm-backward kernels

* Renaming again

* Add block_sync_lds between two consecutive blockwise reductions

* Move common expression 1/N out of the static_for loops

* Add dy_elementwise_op

* Renaming in backward example again

* Add checking for reduceDims in reference_batchnorm_backward

* Update to comments and codes format

* Rename in the comments

* Remove common expression out of the loop in reference_batchnorm_backward_nhwc_c

* Add block_sync_lds() between blockwise reduction again

* Fix comments again

* Remove int8 from batchnorm-forward instances since it is not needed for forward training and could fail test

Remove int8 from batchnorm-forward instances since it is not needed for forward training and could fail test (#516)

* Update to device_batchnorm_forward base class to include all template parameters for problem description

* Add batchnorm forward instances and external api

* Add batchnorm forward profiler module which uses the external api

* Add some comments in batchnorm_forward example to explain the dimensions in lengths[]

* Replace the reference_batchnorm_forward_nhwc_c by generic reference_batchnorm_forward

* Improvement to the batchnorm infer base API

* Add batchnorm forward client example which shows using the batchnorm forward external API

* Add test for batchnorm forward

* Tuning the batchnorm profiler initialized values and error threshold

* Add support for bhalf_t in instances/external api/tests

* Add support for int8_t in instances/external api/tests

* Add support for double in instances/external api/tests

* Let ScaleDataType and BiasDataType be same as XDataType and YDataType when creating instances

* Checking before running best instance in batchnorm_fwd_nhwc client example

* Add checking for YElementwiseOp in batchnorm_forward external API

* Add more types in batchnorm forward profiler

* Add more test lengths
Co-authored-by: rocking5566 <ChunYu.Lai@amd.com>

* FastGelu support for more data types.

* AddFastGelu & FastGelu instances.

* Client example.

* clang-format

* Remove unused stride variable.

* Add new line at EOF.
Co-authored-by: Adam Osewski <aosewski@amd.com>

* workaround bf16 atten fwd issue on gfx908

* typo

* fixed bug in softmax reference & add bf16 examples for batched_gemm_scale_softmax_gemm

* added bf16 tests for batched_gemm_softmax_gemm_permute

* changed format of device_batched_gemm_softmax_gemm_permute_xdl_cshuffle_bf16_bf16_bf16_bf16_gmk_gnk_gno_gmo_instance.cpp

* changed format device_batched_gemm_softmax_gemm_permute_xdl_cshuffle_bf16_bf16_bf16_bf16_gmk_gnk_gno_gmo_instance.cpp

* aligned annotations

* modified CMakeLists for examples

* add common example code of fp16/bf16 version for batched_gemm_scale_softmax_gemm_xdl

* use macro to control the instances

* added macro control into instances

* clang-format some files

* changed error tolerance for bf16

* changed index for 10_elementwise_normalization

* fixed xdlops code bug in amd_xdlops.hpp
Co-authored-by: Po Yen Chen <PoYen.Chen@amd.com>

* start add example

* add device dl

* change launch kernel

* change init data method

* change example config

* add config valid check

* add instance for dl bwd

* add instance to ckProfiler

* reserver to profiler and cmakelist

* add instance to ckProfiler2

* change instance f32 config

* fix example return value
Co-authored-by: letaoqin <letaoqin@amd.com>
Co-authored-by: Po Yen Chen <PoYen.Chen@amd.com>

We can use this template to eliminate duplicated iterator computing
logics. By providing return type to ck::accumulate_n(), we can avoid
type conversion operations.

* Rangify STL algorithms

This commit adapts rangified std::copy(), std::fill() & std::transform()

* Re-write more std::copy() calls

* Re-write std::copy() calls in profiler

* Rangify check_err()

By rangifying check_err(), we can not only compare values between
std::vector<>s, but also compare any ranges which have same value
type.

* Re-format example code

* Add missing ignore expression

* Add missing include directive

* Rangify STL algorithms

This commit adapts rangified std::copy(), std::fill() & std::transform()

* Rangify check_err()

By rangifying check_err(), we can not only compare values between
std::vector<>s, but also compare any ranges which have same value
type.

* Allow constructing Tensor<> like a HostTensorDescriptor

* Simplify Tensor<> object construction logics

* Remove more unnecessary 'HostTensorDescriptor' objects

* Re-format example code

* Re-write more HostTensorDescriptor ctor call

* Add packages for example and profiler

* correct TEST_NAME -> EXAMPLE_NAME

Allow passing forward range to its call operator

* add client example for elementwise_normalization

* clang format elementwise_layernorm2d.cpp

* changed some naming to make it more understandable

* changed naming of input into ab_input

* fixed bug for threadwise_x_store

* add elementwise operation to reference

* Rename example folder for GroupedConvFwdMultipleD

* Unify example codes

* Change target names

* Add fp16 example for multiple d instance

* Re-format common.hpp

* Add interface 'DeviceGroupedConvFwd'

* Use simpler interface

* Move common conv params out

* Rename conv fwd client example folder

* Add missing include directive

* Update grouped conv instance implementations

* Simplify ckProfiler (grouped conv forward)

* Use GroupedConvFwd to implement client example

* Use greater groupe count in example

* Add custom target to group examples

* Add extra tag param to instance factory function

* Use tag to differentiate factory functions

* Add missing tag argument for factory function

* Remove inheritance relationship

* Remove no-longer used include directive

* Add license in front of file

* Remove redundant CMake setting

* Extract common code from files

* Rename folder 'convnd' to 'conv'

* Use std::array<> to accept compile-time kwnown # of arguments

* Fix compilation error of tuning parameter

* In example, use same setting as unit-test

* Remove no-longer used include directive

* Add interface for grouped conv bwd weight

* Add group support for conv bwd weight

* Add grouped conv bwd weight example

* Use group parameter in example

* Rename example folder

* Remove non-grouped version example source files

* Rename device op template

* Add group support to convolution backward weight

* Remove debug messages

* Use smaller group size in example

* Use named variable as loop terminate condition

* Prettify example output message

* Enlarge used grid size

* Allow real grid size exceeds expected grid size

* Rename interface file

* Add client example for group...

* Remove interface 'DeviceGroupedConvBwdData'

* Remove no-longer needed include directive

* Rename client example folder

* add fused addition lyernorm

* add fused addition lyernorm

* changed CMakelist

* removed annotates

* modified descriptor of C

* fixed bug in gridwise add layernorm

* format the files

* modified name from add&layernorm into elementwise&layernorm

* created fused elementwise layernorm branch

* change input into tuple type

* add sweep once to reduce load & read of C from global memory

* modified Argument api

* modified way to malloc c in global memory

* changed gamma and beta to m_k_desc

* fixed bug when sweep once and move CDataType when define device level struct

* add src dim for gamma and beta

* implement optimization for coalesced

* delete a annotation line

* fixed some bug to meet the requirements of ck

* add bandwidth computing in example, and fixed the time unit

* move device_elementwise_layernorm_impl.hpp into device/impl

* fixed bug in device_elementwise_layernorm_impl.hpp

* changed name from layernorm into normalization

* clang-format the changed files

* changed the names

* moved immidiate results into lds, it become faster in non-sweeponce cases

* changed naming of C into X to make the defination more clear

* changed naming in example

* add tests for elementwise normalization

* move example_elementwise_layernorm_blockwise into folder 44_elementwise_normalization

* move test_elementwise_layernorm_fp16 into new folder

* move elementwise_normalization_instances into a new folder

* add more tests in test_elementwise_layernorm_fp16.cpp

* added some corner cases in test

* fixed method to compute lds size for matrix X

* changed name of 44_elementwise_normalization into 45_elementwise_normalization

* modified some comments

* modified some other confused comments

* reduce redundant tests in test_elementwise_layernorm_fp16.cpp

* disable gtest discovery to run tests per-program not per-case

* register cmake target to ctest

* Sync the naming

* Sync the test of layernorm with groupnorm

* Sync the naming

* Minor change for comment and log

* [What] Add saveMean and SaveInvVariance in the interface.
[Why] These can optimize the backward

* Improve example reusability

* Remove no-longer used file

* Rename folder of grouped_conv_bwd_data example

* Add normal grouped conv bwd example

* Add interface 'DeviceGroupedConvBwdData'

* Prettify comment of device op type arguments

* Add grouped conv2d/conv3d backward data fp16 instances

* Fix wrong template argument

* Add grouped_conv2d_bwd_data client example

* Use simpler expression to calculate memory size

* Fix formating

* Remove grouped_conv3d_bw_data instances

Underlying device operator is not ready to handle 3D input

* Remove no-longer necessary include directive

* Add missing include directive

* Use more realistic conv param in example

* Add gridwise gemm pipeline v1/v2 selector

* Pipeline selector working, test-wise add pipeline options to one instance

* Add gemm instances

* Add debug info to DeviceGemmXdl

* Add debug info to DeviceGemmXdl_CShuffle

* Add debug info to DeviceGemmXdl_CShuffle and instances to gemm_add_add_fastgelu

* Minor fix

* Add debug info to DeviceBatchedGemmXdl and instances to batched_gemm

* set up inter-wave configuration

* use defualt loop scheduling for supported gemm ops

for blanket-applying interwave scheduling for all supported gemm ops, define macro CK_EXPERIMENTAL_DEFAULT_TO_INTER_WAVE_SCHEDULING=1. this should be discouraged though as it is not covered by CI

* Add enum PipelineVersion

* Update instances

* Format

* Fix the merge conflict

* Add flags to disable added instances

* Test disable flag check

* Disable flag check

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

* Add reduction across all dims cases.

* host softmax: handle all reduce

* Test cases when reduced dim is not innermost axis.

* Fix syntax.

* Test non innermost dim for fp32 and int8

* Group test suites wrt NumReduceDim.

* Additionally test failing cases.

* Throw error when Rank or NumReduceDims doesn't match arguments.

* Check reducedDims has correct values

* Move don't reuse DeviceReduceMultiblock IsSupportedArgument method.
Instead implement own. (in fact just get rid of one check to enable
reduction across inner dimensions).

* Reorganize unit tests to better cover use scenarios.

* Test input validation
* Test reduction of inner dimensions with custom op instances.

* Refactor fp32 and int8 unit tests.

* Fix FP32 instance template parameters.

* Add more instances.

* Instances with InSrcVectorDim=0.

* Do not initialize and copy data when arg not supported.

* ckProfiler Softmax use instance factory.

* Refactor device softmax IsSupported.

* Additionally add non-polymorphic api functions

* Split softmax instances into multiple files.

* Fix profiler.

* Reorganize tests to reuse profiler and cover edge cases.

* Clang-format

* I8 Softmax instances along with UT.

* Reuse type alias definitions from instance factory header.

* Clean included headers

* Fix variable names.

* Add missing checks in Argument constructor.
Co-authored-by: Adam Osewski <aosewski@amd.com>
Co-authored-by: Anthony Chang <ac.chang@outlook.com>

* Add conv2d requant example

* Fix bash error

* Rename example

* 1. Rename gemm quantization
2. shares the requantization lambda function with conv

* Refine declare type

* Add conv bias relu quantization exmaple

* clang format

* Fix compile error due to merge develop

* Fix CI error

* Extract quantization post operation into another file

* Support quantization for non piecewise linear function

* Add instance for conv quantization

* Add convolution quantization factory

* Add convolution quantization client example

* Add more instances with different template parameters

* clang format

* Sync the naming with the develop

* add device of dl

* fix k1 of GridwiseGemmDl_km_kn_mn_v1r3

* init version for dl conv

* add example(init)

* result right

* disable elementwise operation

* check parameters

* add fp32,int8 example and change check code

* change deive file and class name

* add check vector access of C

* add instance

* add to ckProfiler

* add Filter1x1Pad0 instances

* fix ignore error

* fix for CI
Co-authored-by: letaoqin <letaoqin@amd.com>

* Update to the batchnorm-forward API and base class

* Fix leeked header including in gridwise_set_buffer_value.hpp

* Add kernels and device file for batchnorm-forward welford supporting both blockwise and multi-block reduction

* Update to the batchnorm-forward example to use the new batchnorm-forward device interface

* Change the batchnorm-forward reference to use sequential welford method

* Change to assign the workspace into four buffers in the host layer

* Use GetReduceCountPerThread functor to replace the initial count for Blockwise and Multiblock welford

* Tiny correction and remove un-used file under example/34_batchnorm

* Renaming in the kernel arguments

* Explicitly use ck::math::sqrt in batchnorm-forward kernels

* Add some comments to some kernels

* Tiny fix

* Generalize the data types in reference_batchnorm_forward_nhwc_c

* Use ck::ignore to mark un-used parameters

* Move GetReduceCountPerThread functor codes from kernel to device

* Remove some un-used codes in device_batchnorm_forward_impl.hpp

* Tiny fix in batchnorm_forward example

* Move GetReduceCountPerThread() to welford_helper.hpp

* Use seperate data type for Scale and Bias

* Renaming in device Op

* Tiny fix in forward example

* Updata to batchnorm-infer (type spliting, renaming)

* Add time and bandwidth measurement to the batchnorm-forward example

* Add support of elementwise operation for batchnorm forward output

* Reduce object copying by passing object as reference type

* Tiny change for performance

* Updates for performance again

* Some Renamings

* Add GetActualVariance template parameter for ThreadwiseWelfordMerge

* Tiny update in reference batchnorm forward nhwc/c

* Move batchnorm multiblock kernel files to grid/batchnorm_multiblock sub-directory

* Fuse mean and bias in the normalization calculation
Co-authored-by: root <root@dc-smc-18.amd.com>
Co-authored-by: rocking5566 <ChunYu.Lai@amd.com>

* reduce the number of default targets

* re-write the setting of target flags

* move all options to one place

* add new custom target instances for installing CK

* reopen masking att instance due to CI is upgraded

* re-enable instances previously failed on 9110

* enable ksize-kpadding pair validity test

* add non-masked attention+permute test; expose masking boolean to attention kernel handles

* disable bench

* fix test

* move files

* bulk rename batched_gemm_masking_scale_softmax_gemm_permute to batched_gemm_softmax_gemm_permute

* format

* amend rename

* disable bench in test

* add mask/no-mask test for non-permute attention kernels

* disable broken kernel instance

* example working

add non-permuted problem statement

evaluating whether overhead comes from permutation or the extra kernel arg

* interface for bias addition without implementing it

* test and profiler running

* tidy

* mask type determined by enum class

* unify example code

* move masking specialization to its own header

* align formats

* extract helper functions

* experiment merging dims for attn w/ permute; shows perf parity with attn wo/ permute

* add tensor specialization to template args

since tensor spec packed shows perf parity when permutation isn't needed

remove redundant template args

comment on 'packed' tensor specialization

* grouped attention with input/output permute example

* format

* clean up

* refactor acc0 tile visitor

* fused attention client example

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

* reopen masking att instance due to CI is upgraded

* re-enable instances previously failed on 9110

* enable ksize-kpadding pair validity test

* add non-masked attention+permute test; expose masking boolean to attention kernel handles

* disable bench

* fix test

* move files

* bulk rename batched_gemm_masking_scale_softmax_gemm_permute to batched_gemm_softmax_gemm_permute

* format

* amend rename

* disable bench in test

* add mask/no-mask test for non-permute attention kernels

* disable broken kernel instance

* example working

add non-permuted problem statement

evaluating whether overhead comes from permutation or the extra kernel arg

* interface for bias addition without implementing it

* test and profiler running

* tidy

* mask type determined by enum class

* unify example code

* move masking specialization to its own header

* align formats

* extract helper functions

* experiment merging dims for attn w/ permute; shows perf parity with attn wo/ permute

* add tensor specialization to template args

since tensor spec packed shows perf parity when permutation isn't needed

remove redundant template args

comment on 'packed' tensor specialization

* grouped attention with input/output permute example

* format

* clean up

* refactor acc0 tile visitor
Co-authored-by: shaojiewang <wsjmessi@163.com>
Co-authored-by: Chao Liu <chao.liu2@amd.com>

* Fix for lwpck-425, update BlockTransferSrcVectorDim

* Revert "Fix for lwpck-425, update BlockTransferSrcVectorDim"

This reverts commit fd24e280e28ff238b452cfdde58a988affd46461.

* Add Batched Gemm int8 test, expect it to fail

* Format

* Re-add the fix