GitLab

* Add license header.

* Reduce number of logged output. Add constant initialization.

* Add functional tests for grouped_gemm with different kbatch value.

* Add debug log informations + remove unused code.

* Don't pass kbatch to CalculateKPadded.

* Turn on logging in grouped gemm and gemm splitk profiler

* Debug: limit number of test cases to run;

* Log more information and initialize with constant value.

* Turn on DEBUG_LOG

* Add more debug log informations.

* Limit the number of instances to compile.

* Use GridwiseGemmPipeline

* Use KBatch to calculate K0

* Multiple DebugLog messages.

* Unit tests for multiple KBatch values.

* Refactoring

* Disable logging
* extract out of if statement KBatch update.

* Uncomment instances.

* Disable DebugLog.

* Use Kbatch when calculate KPadded.

* Fix CGridDesc padding.

* Use available helper functions.

* Uncomment code commented for debuggin.

* Remove unnecessary debug log messages.

* Uncomment previously commented code for debug purposes.

* Add KBatch info to profiler output summary log.

* Add gtests for gemm splitk using ckProfiler API.

* Add more test-cases for different data layout.

* Add more test cases for gemm splitk

* Remove old test.

* Unit tests for MKNK ggemm interface.

* Fix and add more unit-tests.

* Constepxr everything!

* Increase error threshold for fp16 and splitk.

Since we're using fp16 atomic add for splitk there's a
known precision loss.

---------
Co-authored-by: Adam Osewski <aosewski@amd.com>
Co-authored-by: zjing14 <zhangjing14@gmail.com>

* Add instances for fp16/int8 Gemm kernels (Navi21)

* Extend instances with smaller tiles

* Fix SrcVectorTensor for km_kn_mn int8

* fix headers for gpu instances

* remove unused headers

---------
Co-authored-by: zjing14 <zhangjing14@gmail.com>

* Expand the base class of pool2d, prepare to share base class with pool3d

* Add pool3d device op

* Add pool3d f16 example

* Refactor the base class. implement generic pooling in the future

* clang format

* get original index in max pooling

* Add outputindex to base class

* Fix dimension

* Add pooling instance

* Use indexType instead

* Remove useless header

* Extract IndexDataType to template

* Extract pooling reference code

* clang format

* clang format

* Fix typo

* Add tensor stride

* Add missing header

* Add index stride and output stride

* Refine naming

* Add type to base class

* Rename file

* Use proper size

* Fix typo

* Refine naming

* Modify the argument into vector.

* Add max pool profiler

* Refine naming

* Support f32 pool

* Fix typo

* Add avg pool2d fwd in profiler

* clang format

* Rename AccDatatype to ComputeDatatype

* Fix init

* test pool

* Extract variable

* Add client example

* Check the pooling dim

* clang format

* Connect argv and arg_parser

* Add found check

* Remove useless header

* Refine naming

* Adjust the order of device_pool_fwd

* Add contraction profiler and tests

* Build and style fixes

* Allow to use any elementwise operator for ref_contraction

* Introduce profile_contraction_scale and profile_contraction_bilinear

* Make ref_contraction generic and extend interface tests

* Stylistic minor fixes

* Extend test_contraction_interface

* Add TypeConvert class and start refactoring

* Refactor TypeConvert as a struct

* Get back to template functions type_convert

* Add a type_convert_bf16_rtn, set rtz as default

* Clean up

* Add UnaryConvertPrecision struct for high-precision workloads

* Format

* Update type_convert to UnaryConvert on threadwise level

* Update UnaryConvertPrecision

* Format

* Fix chmod

* Add a flag to pick converion method

* Format

* Remove the added flag

* Merge elementwise op with type conversion

* Move type_convert to elemwise op, update the op

* Update type_convert_precision -> bf16_convert_rtn

* Clean up

* Update comments

* Update the CK_WORKAROUND_DENORM_FIX flag handling

* Update the unneeded op to work but warn user

* Remove the message

* Use a PassThrough instead of ConvertBF16RTN to calcaulate reference

* Format

* Add missing include

* simplify karg in device/grid split-k op

* fix mk_kn_mn instances

* add more instances

* B2C with 3D grid for KSplit

* Remove unused code.

* Use default B2C (3D grid) in grid gemm v2r4r2.

* Device gemm splitk use B2C map.

* Device GroupedGemmXdlSplitKCShuffle

* Example for GroupedGemm Xdl SplitK

* Introduce Device GroupedGemmSplitK

* Fix updating kbatch size.

* Add instance mk-nk-mn

* Enable set kbatch in profiler.

* Add GGemmSplitK mk-kn-mn instances

* Add more instances & split into multiple files.

* minor fix

* tuning

* clean

* disabled failed instances

* use pipe v2

* Ignore arg on not supported arch.

* fix warning

---------
Co-authored-by: carlushuang <carlus.huang@amd.com>
Co-authored-by: Adam Osewski <aosewski@amd.com>
Co-authored-by: zjing14 <zhangjing14@gmail.com>
Co-authored-by: Jing Zhang <jizhan@amd.com>
Co-authored-by: root <root@ctr-ubbsmc15.amd.com>

* [What] Remove pure conv int8 instance
[Why] We will never use pure int8 conv in AI, use int8 quantization instead

* Change layout

* Share the kernel parameter

* Support more type of NHWGC for group conv

* Revise client example of conv 2d, use NHWGC layout

* Add instance to cmake

* Revise layout of group conv quantization instance

* Revise layout of external api of group conv quantization

* Revise layout of group conv quantization client example

* Fix clang format

* Add comment to describe meaning of each parameter

* simplify karg in device/grid split-k op

* fix mk_kn_mn instances

* add more instances

* use name from tensor layout

---------
Co-authored-by: carlushuang <carlus.huang@amd.com>

* Rename to proper naming

* Add example of groupnorm + swish

* Extract duplicate code in example

* Add groupnorm + swish instances

* Ractor instance generation, split into multiple cpp file

* Add external api and client example

* Refine profiler message

* Use ck math version of exp

* Refine problem size in example

* Add host version of exp

This reverts commit bb5530af.

Co-authored-by: root <root@ctr-ubbsmc15.amd.com>

* simplify karg in device/grid split-k op

* fix mk_kn_mn instances

* add more instances

* use name from tensor layout

* Rename file. Prepare to support another activation

* Add comment for quantization

* Extract out_elementop

* Add tanh example

* Add conv + bias + tanh quantization instance

* Add missing parameter

* Refine cmake

* Add external api and client example

* Extract variable in example

* Fix the comment

---------
Co-authored-by: zjing14 <zhangjing14@gmail.com>

* add workaround 637

* format

* change id

---------
Co-authored-by: zjing14 <zhangjing14@gmail.com>

* Add conv perlayer quantization

* Add gemm_dlops quantization

* Support int8 for innerproduct

* Refine gemm dlops int8 kernel parameter

* Support gfx908(MI100) and gfx90a(MI200)

* clang-format

* Rename example number

* Support different layout for d tensor

* Add conv dlops perchannel quantization example

* Move to example 40

* Extract the common code for different platform (dlops and xdlops)

* Move ot subfolder. Prepare to add other op of quantization

* Refine the quantization instance library

* Add conv dl instances and client example

* Remove unnecessary type

* Add gemm quantization instance

* Add external api and client example

* Refine num_bytes

* Separete different layout to different cpp

* Add more xdl instances

* Revert "Remove unnecessary type"

This reverts commit 82086918

.

* Remove CShuffleDataType in dlops
Let acc and CShuffleDataType be the same in xdlops

---------
Co-authored-by: zjing14 <zhangjing14@gmail.com>

* Grouped gemm + Gelu instances.

* Device Instance Factory for GroupedGemm+Gelu

* Client example

* Rangify fill helper functions.

* Fix name clash.

* Profiler for grouped_gemm+gelu

* No need to use full namespace name.

* Add check for MRaw divisible by vector load.

* Ugly fix for big errors.

* Add grouped_gemm+gelu to profiler CMakelists.

* Store in argument additional info.

* Information about Mraw, Nraw, Kraw values.

* Use FastGelu instead of Gelu.

* Change client ex to use FastGelu

* Remove relaxed error precision.

* Remove duplicate output elementwise-op

---------
Co-authored-by: Adam Osewski <aosewski@amd.com>
Co-authored-by: zjing14 <zhangjing14@gmail.com>

* Modify Doxygen config to pick up include directories recursively

* Add DeviceMem struct to API Reference guide

* Add classes that are used in Flash Attention kernel

* Add a reference and config for generating bibliography
Co-authored-by: Philip Maybank <Philip.Maybank@amd.com>

* Sync the order of type string with template parameter

* Add more instances

* Check the vector size and remove redundant var

* Extract var to static, prepare to separate sweep once kernel

* Separate sweeponce flow and optimize the flow

* 1. Rename AccDatatype in normalization to computeData
2. Rename AccElementwiseOperation to YElementwiseOperation in normalization

* Remove useless code

* Update naive variance kernel

* Refine string

* Fix typo

* Support naive variance for device_normalization

* Check the blocksize

* Share the VGPR of x and y

* Share the VGPR of gamma and beta

* Add more instances

* Support fp16 sqrt for experiment

* Add CHANGELOG

* Fix typo

* clang-format

* Conv3d bwd weight client example.

* Update year in license

* Convolution bwd data 3D fp16/fp32 client example.

* Client example for convnd fwd fp16 fp32

* clang-format

* Review remarks.

* Fix compiler err.

* Update data layout to standard one.

* Add conv 3d fwd NDHWGC instances

* clang-format

* Conv3d fwd NDHWGC instances.

---------
Co-authored-by: Adam Osewski <aosewski@amd.com>
Co-authored-by: zjing14 <zhangjing14@gmail.com>

* Adding more bigger tiles.

* Remove failing instance.

* Remove instances which that don't improve perf.

---------
Co-authored-by: Adam Osewski <aosewski@amd.com>
Co-authored-by: zjing14 <zhangjing14@gmail.com>

* Add gemm + layernorm instance

* Add ckProfiler

* Add test

* Add client example

* Detect if user forger to set the workrspace

* Use literal in the example

* [What] use builtin function for sqrt
[Why] compiler will not use v_sqrt_f64_e64 if we use ::sqrt()

* check gemm vaildity in IsSupportedArgument

* Add more testcases

* Merge duplicated folder in client example

* Print more infomation

* Use better kernel parameter for MS problem size

* clang format

* Add constexpr for if condition and remove redundant include

* Remove cstdlib and add constexpr

* added instances for large N

* add instance for elementwise normlization

* added supported restrict in device_elementwise_normalization_impl.hpp

* add instance for gemm bias softmax gemm

* add client example

* change CGridDesc_G_M_N to CGridDesc_G_M_O

* add gridwise

* change c grid name

* device add d0s data

* fix 08 client_example

* add example 47_fused_attention

* example output correct

* add d0 to example

* add d0 element op

* rechange instance code

* change Acc0ElementwiseOperation to C0DEElementwiseOperation

* change example name

* update instance for cdeelementwiseop

* add bhalf_t ScaleAdd

* add test

* not surport geem1 bias

* remove some ignore

* fix test bug

Co-authored-by: Adam Osewski <aosewski@amd.com>

* File renaming and class renaming for device element-wise operation

* Add batchnorm-infer instances, external API and client example

* Add batchnorm-infer profiler module and gtests

* Remove file device_elementwise_extension.hpp and move NormalizeInInfer operation to element_wise_operation.hpp

* Remove the using of class aliasing for DeviceElementwiseForBatchNormInfer

* Rename class and file due to conflict from device_elementwise_2d.hpp

* Fix namespace in batcnnorm_infer_nhwc client example

* Use double as alpha/beta values type in reduce device op api

* Use double as alpha/beta values type in softmax device op api

* Use double as alpha/beta values type in multiple-reduce device op api

* Use double as epsilon value type in normalization/elementwise-normalization device op api

* start add example

* fix config

* fix showinfo bug

* add an elementop

* change to padding

* add xdl example

* change elementwiseop

* add instance

* add instance to profiler

* change file name

* fix deive not support issue

* add client example

* fix client gemm_add_multiply name

* change AddMultiply elementwiseop

* fix elementwiseop

* fix client example

* fix addmultiply op

* fix comments and fun name
Co-authored-by: letaoqin <letaoqin@amd.com>

* add example

* fix example

* add instance for gemm permute

* add to client example

* change configs

* change instance file name

* formate

* change client example file name and remove example

* Change to the DeviceReduce base class template to include all problem description information

* Add external api for reduction

* Add client example to test the reduction external api

* Spelling correction

* Re-implement the host_reduction to follow the DeviceReduce base API format

* Change the reduce profiler to call the external API for collecting device instances

* Rename reduce client example directory from 08_reduce to 12_reduce

* Remove (void) before the functional call

* Tiny update in reduce client example

* Tiny update in profile_reduce_impl.hpp

* Rename the reduce client example directory
Co-authored-by: Po Yen Chen <PoYen.Chen@amd.com>

* Add device op of gemm layernorm

* [What] Rename F to H
[Why] F and G prepare for welford tensor

* Add gridwise gemm + welford

* Extract template parameter

* Rename kernel. Prepare to add second half kernel

* Extract var

* Add second kernel for gemm+layernorm

* Move to the gemm_layernorm folder

* Rename F and G to mean and var

* Do not use snakeCurved, it makes determination of padding  for welford difficult

* Rewrite the device interface and rename some var

* Add welford count

* Update interface

* Sync code, prepare to test on MI200

* Clean the code

* Implement layernorm

* Add comment to mension hipFree

* Wrtie out the e for debug.
This could be remove and use h for instead

* 1. Allocate mean, var and count into by SetWorkSpacePointer.
2. Add GetWorkSpaceSize to calculate the space size

* Add gemm layernorm host code

* use reference layernorm

* Fix bug of blockwise welford for first kernel

* Fix bug of mean var padding for layernorm

* Use sgpr for shuffleM_index

* padding for GemmMeanVarCountGridDescriptor_M_NBlock

* Add layout parameter

* Check argument for gemm

* calculate max count for tail block

* Share E and H memory in device op

* Hard code the vector dim

* Refine the MakeDescriptor

* 1. Remove E parameter, because E is inside of device op
2. Check vector size

* [What] Rename MakeMeanVarDescriptor_M_N
[Why] Prepare to add count version of make descriptor

* Use 1D global memory for count

* Prevent redundant IO

* Update parameter

* Add pipeline v1/v2 selector

* Rename the example name

* Add base class for gemm layernorm

* Refine naming to distinguish naive and welford

* Add comment to explan in detail

* We don't need to pad in N dimension in gemm for mean/var/count. Set NPerTile 1

* Rewrite the 2st kernel, use multiple block along N dimension in layernorm kernel

* Share the vector size

* Refine var name

* [What] Force LayernormThreadSliceSize_N = vector size.
[Why] Memory coalesce

* Add comment

* Extract divisor out of the loop in reference layernorm

* Pad different size for E and H in layernorm kernel according to different block tile

* Refine naming

* Refine naming

* Prevent implicit cast

* [What] use ck::math::sqrt instead of __builtin_amdgcn_sqrtf
[Why] __builtin_amdgcn_sqrtf is only support float, double will cause casting

* Cast only constant

* Change of post shuffle thread descriptor

* Add EMeanVarDataType parameter.

* Merge the mean and var threadwise copy

* Add missing index

* Fix Typo

* Sync the variable with previous if

* 1. Declare e inside the host_gemm_layernorm()
2. Prevent implicit cast in reference code
Co-authored-by: Po Yen Chen <PoYen.Chen@amd.com>

* add mnk padding, support m=0

* clean code

* clean code
Co-authored-by: Rostyslav Geyyer <46627076+geyyer@users.noreply.github.com>

Add padding device_gemm_add_add_fastgelu_xdl_c_shuffle instances to enable arbitrary problem size (#535)

* Add padding device_gemm_add_add_fastgelu_xdl_c_shuffle instances

* Add padding device_gemm_add_fastgelu_xdl_c_shuffle instances

* Add gemm_add_fastgelu profiler impl

* Add padding device_gemm_fastgelu_xdl_c_shuffle instances

* Add gemm_fastgelu profiler impl

Co-authored-by: Rosty Geyyer <rosty.geyyer@amd.com>
Co-authored-by: Chao Liu <chao.liu2@amd.com>

* start add example

* add multiple d fp16 example

* device transfer elementwiseop to gridwise

* gridwise add multiple d

* change example for multiple d

* fix spill registers

* fix for passthrough element op

* fix int8 overflow

* change example file name

* add instance for dl multiple d

* example add DsDataType

* remove grouped_convolution_forward_dl.hpp

* add head file(was deleted before)

* fix not support device issue

* format

* remove passthrough check
Co-authored-by: letaoqin <letaoqin@amd.com>

* Use gemm_multiple_D instead

* Add gemm bias relu quantization example

* Add pure gemm quantization example

* Add quantization of perchannel conv + bias + relu example

* Refine the code

* Rename multiplier to requant_scale

* Rename the folder

* Remove redundant comment

* Rename the file. Prepare to add perchannel

* Add conv perchannel instance

* Move to quantization folder

* Add conv perchannel client example

* Apply Rangify constructor of HostTensorDescriptor & Tensor<>

* Fix merge error

* 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()

* 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