Skip to content

GitLab

"docs/git@developer.sourcefind.cn:gaoqiong/pybind11.git" did not exist on "3b265787f253b84598a0453fc2ff3fa25f3d5acf"
Unverified Commit aaf4defa authored by Illia Silin's avatar Illia Silin Committed by GitHub
Browse files

Merge pull request #9 from ROCmSoftwarePlatform/mergedown-from-public 

Merge down from public repo
parents e0c2a70f e51806cc
Hide whitespace changes
Inline Side-by-side
Showing with 252 additions and 317 deletions
docs/source/Disclaimer.rst deleted 100644 → 0
View file @ e0c2a70f
************
Disclaimer
************
-------------------------------
AMD's standard legal Disclaimer
-------------------------------
The information presented in this document is for informational purposes only and may contain technical inaccuracies, omissions, and typographical errors. The information contained herein is subject to change and may be rendered inaccurate for many reasons, including but not limited to product and roadmap changes, component and motherboard version changes, new model and/or product releases, product differences between differing manufacturers, software changes, BIOS flashes, firmware upgrades, or the like. Any computer system has risks of security vulnerabilities that cannot be completely prevented or mitigated. AMD assumes no obligation to update or otherwise correct or revise this information. However, AMD reserves the right to revise this information and to make changes from time to time to the content hereof without obligation of AMD to notify any person of such revisions or changes. THIS INFORMATION IS PROVIDED 'AS IS." AMD MAKES NO REPRESENTATIONS OR WARRANTIES WITH RESPECT TO THE CONTENTS HEREOF AND ASSUMES NO RESPONSIBILITY FOR ANY INACCURACIES, ERRORS, OR OMISSIONS THAT MAY APPEAR IN THIS INFORMATION. AMD SPECIFICALLY DISCLAIMS ANY IMPLIED WARRANTIES OF NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR ANY PARTICULAR PURPOSE. IN NO EVENT WILL AMD BE LIABLE TO ANY PERSON FOR ANY RELIANCE, DIRECT, INDIRECT, SPECIAL, OR OTHER CONSEQUENTIAL DAMAGES ARISING FROM THE USE OF ANY INFORMATION CONTAINED HEREIN, EVEN IF AMD IS EXPRESSLY ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. AMD, the AMD Arrow logo, Radeon, Ryzen, Epyc, and combinations thereof are trademarks of Advanced Micro Devices, Inc. Other product names used in this publication are for identification purposes only and may be trademarks of their respective companies. Google(R) is a registered trademark of Google LLC. PCIe(R) is a registered trademark of PCI-SIG Corporation. Linux(R) is the registered trademark of Linus Torvalds in the U.S. and other countries. Ubuntu(R) and the Ubuntu logo are registered trademarks of Canonical Ltd. Other product names used in this publication are for identification purposes only and may be trademarks of their respective companies. (C)2023 Advanced Micro Devices, Inc. All rights reserved.
----------------------
Third Party Disclaimer
----------------------
Third-party content is licensed to you directly by the third party that owns the content and is not licensed to you by AMD. ALL LINKED THIRD-PARTY CONTENT IS PROVIDED "AS IS" WITHOUT A WARRANTY OF ANY KIND. USE OF SUCH THIRD-PARTY CONTENT IS DONE AT YOUR SOLE DISCRETION AND UNDER NO CIRCUMSTANCES WILL AMD BE LIABLE TO YOU FOR ANY THIRD-PARTY CONTENT. YOU ASSUME ALL RISK AND ARE SOLELY RESPONSIBLE FOR ANY DAMAGES THAT MAY ARISE FROM YOUR USE OF THIRD-PARTY CONTENT.
docs/source/Linux_Install_Guide.rst deleted 100644 → 0
View file @ e0c2a70f
=====================
Getting Started Guide
=====================
------------
Introduction
------------
This document contains instructions for installing, using, and contributing to Composable Kernel (CK).
Documentation Roadmap
^^^^^^^^^^^^^^^^^^^^^
The following is a list of CK documents in the suggested reading order:
[TODO]
\ No newline at end of file
docs/source/Makefile deleted 100644 → 0
View file @ e0c2a70f
# Minimal makefile for Sphinx documentation
#
# You can set these variables from the command line.
SPHINXOPTS =
SPHINXBUILD = sphinx-build
SPHINXPROJ = CK
SOURCEDIR = .
BUILDDIR = _build
# Put it first so that "make" without argument is like "make help".
help:
@$(SPHINXBUILD) -M help "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O)
.PHONY: help Makefile
# Catch-all target: route all unknown targets to Sphinx using the new
# "make mode" option. $(O) is meant as a shortcut for $(SPHINXOPTS).
%: Makefile
@$(SPHINXBUILD) -M $@ "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O)
docs/source/conf.py deleted 100644 → 0
View file @ e0c2a70f
"""Copyright (C) 2018-2023 Advanced Micro Devices, Inc. All rights reserved.
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell cop-
ies of the Software, and to permit persons to whom the Software is furnished
to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IM-
PLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS
FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR
COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER
IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNE-
CTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
"""
# -*- coding: utf-8 -*-
#
# Composable Kernel (CK) docuumentation build configuration file, based on
# rocBLAS documentation build configuration file, created by
# sphinx-quickstart on Mon Jan 8 16:34:42 2018.
#
# This file is execfile()d with the current directory set to its
# containing dir.
#
# Note that not all possible configuration values are present in this
# autogenerated file.
#
# All configuration values have a default; values that are commented out
# serve to show the default.
# If extensions (or modules to document with autodoc) are in another directory,
# add these directories to sys.path here. If the directory is relative to the
# documentation root, use os.path.abspath to make it absolute, like shown here.
#
# import os
# import sys
# sys.path.insert(0, os.path.abspath('.'))
import os
import sys
import subprocess
read_the_docs_build = os.environ.get('READTHEDOCS', None) == 'True'
if read_the_docs_build:
subprocess.call('../run_doxygen.sh')
# -- General configuration ------------------------------------------------
# If your documentation needs a minimal Sphinx version, state it here.
#
# needs_sphinx = '1.0'
# Add any Sphinx extension module names here, as strings. They can be
# extensions coming with Sphinx (named 'sphinx.ext.*') or your custom
# ones.
extensions = ['sphinx.ext.mathjax', 'breathe']
breathe_projects = { "CK": "../docBin/xml" }
breathe_default_project = "CK"
# Add any paths that contain templates here, relative to this directory.
templates_path = ['_templates']
# The suffix(es) of source filenames.
# You can specify multiple suffix as a list of string:
#
# source_suffix = ['.rst', '.md']
source_suffix = '.rst'
# The master toctree document.
master_doc = 'index'
# General information about the project.
project = u'Composable Kernel (CK)'
copyright = u'2018-2023, Advanced Micro Devices'
author = u'Advanced Micro Devices'
# The version info for the project you're documenting, acts as replacement for
# |version| and |release|, also used in various other places throughout the
# built documents.
#
# The short X.Y version.
#version = u'0.8'
# The full version, including alpha/beta/rc tags.
#release = u'0.8'
# The language for content autogenerated by Sphinx. Refer to documentation
# for a list of supported languages.
#
# This is also used if you do content translation via gettext catalogs.
# Usually you set "language" from the command line for these cases.
language = 'en'
# List of patterns, relative to source directory, that match files and
# directories to ignore when looking for source files.
# This patterns also effect to html_static_path and html_extra_path
exclude_patterns = ['_build', 'Thumbs.db', '.DS_Store']
# The name of the Pygments (syntax highlighting) style to use.
pygments_style = 'sphinx'
# If true, `todo` and `todoList` produce output, else they produce nothing.
todo_include_todos = False
# -- Options for HTML output ----------------------------------------------
# The theme to use for HTML and HTML Help pages. See the documentation for
# a list of builtin themes.
#
# html_theme = 'alabaster'
#if read_the_docs_build:
# html_theme = 'default'
#else:
import sphinx_rtd_theme
html_theme = "sphinx_rtd_theme"
html_theme_path = [sphinx_rtd_theme.get_html_theme_path()]
html_logo = "rocm_logo.png"
# Theme options are theme-specific and customize the look and feel of a theme
# further. For a list of options available for each theme, see the
# documentation.
html_theme_options = {
'logo_only': True,
'display_version': True
}
# Add any paths that contain custom static files (such as style sheets) here,
# relative to this directory. They are copied after the builtin static files,
# so a file named "default.css" will overwrite the builtin "default.css".
#html_static_path = ['_static']
# Custom sidebar templates, must be a dictionary that maps document names
# to template names.
#
# This is required for the alabaster theme
# refs: http://alabaster.readthedocs.io/en/latest/installation.html#sidebars
# html_sidebars = {
# '**': [
# 'relations.html', # needs 'show_related': True theme option to display
# 'searchbox.html',
# ]
# }
mathjax3_config = {
'tex': {
'macros': {
'diag': '\\operatorname{diag}',
}
}
}
# -- Options for HTMLHelp output ------------------------------------------
# Output file base name for HTML help builder.
htmlhelp_basename = 'CKdoc'
# -- Options for LaTeX output ---------------------------------------------
latex_elements = {
# The paper size ('letterpaper' or 'a4paper').
#
'papersize': 'letterpaper',
# The font size ('10pt', '11pt' or '12pt').
#
'pointsize': '10pt',
# Additional stuff for the LaTeX preamble.
#
'preamble': r'''
\setcounter{tocdepth}{5}
\newcommand{\diag}{\operatorname{diag}}
''',
# Latex figure (float) alignment
#
# 'figure_align': 'htbp',
}
# Grouping the document tree into LaTeX files. List of tuples
# (source start file, target name, title,
# author, documentclass [howto, manual, or own class]).
latex_documents = [
(master_doc, 'CK.tex', u'Composabl Kernel (CK) Documentation',
u'Advanced Micro Devices', 'manual'),
]
# -- Options for manual page output ---------------------------------------
# One entry per manual page. List of tuples
# (source start file, name, description, authors, manual section).
man_pages = [
(master_doc, 'ck', u'Composable Kernel (CK) Documentation',
[author], 1)
]
# -- Options for Texinfo output -------------------------------------------
# Grouping the document tree into Texinfo files. List of tuples
# (source start file, target name, title, author,
# dir menu entry, description, category)
texinfo_documents = [
(master_doc, 'CK', u'Composable Kernel (CK) Documentation',
author, 'CK', 'Composable Kernel for AMD ROCm',
'Miscellaneous'),
]
docs/source/index.rst deleted 100644 → 0
View file @ e0c2a70f
============================
Composable Kernel User Guide
============================
.. toctree::
:maxdepth: 5
:caption: Contents:
:numbered:
Linux_Install_Guide
tutorial_hello_world
dockerhub
Supported_Primitives_Guide
API_Reference_Guide
Contributors_Guide
Disclaimer
\ No newline at end of file
example/01_gemm/CMakeLists.txt
View file @ aaf4defa
...@@ -38,7 +38,7 @@ add_example_executable_no_testing(example_gemm_xdl_fp64 gemm_xdl_fp64.cpp) ...@@ -38,7 +38,7 @@ add_example_executable_no_testing(example_gemm_xdl_fp64 gemm_xdl_fp64.cpp)
add_dependencies(example_gemm_xdl example_gemm_xdl_skip_b_lds_fp16) add_dependencies(example_gemm_xdl example_gemm_xdl_skip_b_lds_fp16)
add_dependencies(example_gemm_xdl example_gemm_xdl_fp64) add_dependencies(example_gemm_xdl example_gemm_xdl_fp64)
if(GPU_TARGETS MATCHES "gfx1100") if(GPU_TARGETS MATCHES "gfx1100" OR GPU_TARGETS MATCHES "gfx1101" OR GPU_TARGETS MATCHES "gfx1102")
add_custom_target(example_gemm_wmma) add_custom_target(example_gemm_wmma)
add_example_executable(example_gemm_wmma_fp16 gemm_wmma_fp16.cpp) add_example_executable(example_gemm_wmma_fp16 gemm_wmma_fp16.cpp)
add_dependencies(example_gemm_wmma example_gemm_wmma_fp16) add_dependencies(example_gemm_wmma example_gemm_wmma_fp16)
......
example/01_gemm/gemm_wmma_fp16.cpp
View file @ aaf4defa
...@@ -19,15 +19,15 @@ using AElementOp = PassThrough; ...@@ -19,15 +19,15 @@ using AElementOp = PassThrough;
using BElementOp = PassThrough; using BElementOp = PassThrough;
using CElementOp = PassThrough; using CElementOp = PassThrough;
static constexpr auto GemmDefault = ck::tensor_operation::device::GemmSpecialization::Default; static constexpr auto GemmMNKPadding = ck::tensor_operation::device::GemmSpecialization::MNKPadding;
// clang-format off // clang-format off
using DeviceGemmInstance = ck::tensor_operation::device::DeviceGemmWmma_CShuffle using DeviceGemmInstance = ck::tensor_operation::device::DeviceGemmWmma_CShuffle
// ######| ALayout| BLayout| CLayout| AData| BData| CData| AccData| CShuffle| A| B| C| GEMM| Block| MPer| NPer| K0Per| K1| MPer| NPer|MRepeat|NRepeat| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CBlockTransferClusterLengths| CBlockTransfer| // ######| ALayout| BLayout| CLayout| AData| BData| CData| AccData| CShuffle| A| B| C| GEMM| Block| MPer| NPer| K0Per| K1| MPer| NPer|MRepeat|NRepeat| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CBlockTransferClusterLengths| CBlockTransfer|
// ######| | | | Type| Type| Type| Type| DataType| Elementwise| Elementwise| Elementwise| Spacialization| Size| Block| Block| Block| | WMMA| WMMA| | | ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN|MWmmaPerWave|NWmmaPerWave| _MBlock_MWaveMPerWmma| ScalarPerVector| // ######| | | | Type| Type| Type| Type| DataType| Elementwise| Elementwise| Elementwise| Spacialization| Size| Block| Block| Block| | WMMA| WMMA| | | ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN|MWmmaPerWave|NWmmaPerWave| _MBlock_MWaveMPerWmma| ScalarPerVector|
// ######| | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerWmma| _NWaveNPerWmma| // ######| | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerWmma| _NWaveNPerWmma|
// ######| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | // ######| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
< ALayout, BLayout, CLayout, ADataType, BDataType, CDataType, AccDataType, CShuffleDataType, AElementOp, BElementOp, CElementOp, GemmDefault, 256, 128, 256, 8, 8, 16, 16, 4, 4, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 32, 1, 8>, 8, 1>; < ALayout, BLayout, CLayout, ADataType, BDataType, CDataType, AccDataType, CShuffleDataType, AElementOp, BElementOp, CElementOp, GemmMNKPadding, 256, 128, 256, 8, 8, 16, 16, 4, 4, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 32, 1, 8>, 8, 1>;
// clang-format on // clang-format on
using ReferenceGemmInstance = ck::tensor_operation::host:: using ReferenceGemmInstance = ck::tensor_operation::host::
......
example/02_gemm_bilinear/CMakeLists.txt
View file @ aaf4defa
add_example_executable(example_gemm_bilinear_xdl_fp16 gemm_bilinear_xdl_fp16.cpp) add_example_executable(example_gemm_bilinear_xdl_fp16 gemm_bilinear_xdl_fp16.cpp)
if(GPU_TARGETS MATCHES "gfx1100") if(GPU_TARGETS MATCHES "gfx1100" OR GPU_TARGETS MATCHES "gfx1101" OR GPU_TARGETS MATCHES "gfx1102")
add_example_executable(example_gemm_bilinear_wmma_fp16 gemm_bilinear_wmma_fp16.cpp) add_example_executable(example_gemm_bilinear_wmma_fp16 gemm_bilinear_wmma_fp16.cpp)
endif() endif()
example/04_gemm_add_add_fastgelu/common.hpp
View file @ aaf4defa
...@@ -62,7 +62,7 @@ struct ExecutionConfig final ...@@ -62,7 +62,7 @@ struct ExecutionConfig final
}; };
inline bool inline bool
parse_cmd_args(int argc, char* argv[], ProblemSize& problem_size, ExecutionConfig config) parse_cmd_args(int argc, char* argv[], ProblemSize& problem_size, ExecutionConfig& config)
{ {
if(argc == 1) if(argc == 1)
{ {
......
example/04_gemm_add_add_fastgelu/gemm_add_add_fastgelu_xdl_bf16.cpp
View file @ aaf4defa
...@@ -7,10 +7,11 @@ using ADataType = BF16; ...@@ -7,10 +7,11 @@ using ADataType = BF16;
using BDataType = BF16; using BDataType = BF16;
using AccDataType = F32; using AccDataType = F32;
using CShuffleDataType = F32; using CShuffleDataType = F32;
using D0DataType = BF16; using CDataType = F32; // C matrix doesn't exsit in GPU memory, this is used for host verification
using D1DataType = BF16; using D0DataType = BF16;
using DsDataType = ck::Tuple<D0DataType, D1DataType>; using D1DataType = BF16;
using EDataType = BF16; using DsDataType = ck::Tuple<D0DataType, D1DataType>;
using EDataType = BF16;
using ALayout = Row; using ALayout = Row;
using BLayout = Col; using BLayout = Col;
...@@ -36,7 +37,7 @@ using DeviceOpInstance = ck::tensor_operation::device::DeviceGemmMultipleD_Xdl_C ...@@ -36,7 +37,7 @@ using DeviceOpInstance = ck::tensor_operation::device::DeviceGemmMultipleD_Xdl_C
using ReferenceGemmInstance = ck::tensor_operation::host::ReferenceGemm<ADataType, using ReferenceGemmInstance = ck::tensor_operation::host::ReferenceGemm<ADataType,
BDataType, BDataType,
AccDataType, CDataType,
AccDataType, AccDataType,
AElementOp, AElementOp,
BElementOp, BElementOp,
......
example/04_gemm_add_add_fastgelu/gemm_add_add_fastgelu_xdl_fp16.cpp
View file @ aaf4defa
...@@ -7,10 +7,11 @@ using ADataType = F16; ...@@ -7,10 +7,11 @@ using ADataType = F16;
using BDataType = F16; using BDataType = F16;
using AccDataType = F32; using AccDataType = F32;
using CShuffleDataType = F32; using CShuffleDataType = F32;
using D0DataType = F16; using CDataType = F32; // C matrix doesn't exsit in GPU memory, this is used for host verification
using D1DataType = F16; using D0DataType = F16;
using DsDataType = ck::Tuple<D0DataType, D1DataType>; using D1DataType = F16;
using EDataType = F16; using DsDataType = ck::Tuple<D0DataType, D1DataType>;
using EDataType = F16;
using ALayout = Row; using ALayout = Row;
using BLayout = Col; using BLayout = Col;
...@@ -36,7 +37,7 @@ using DeviceOpInstance = ck::tensor_operation::device::DeviceGemmMultipleD_Xdl_C ...@@ -36,7 +37,7 @@ using DeviceOpInstance = ck::tensor_operation::device::DeviceGemmMultipleD_Xdl_C
using ReferenceGemmInstance = ck::tensor_operation::host::ReferenceGemm<ADataType, using ReferenceGemmInstance = ck::tensor_operation::host::ReferenceGemm<ADataType,
BDataType, BDataType,
AccDataType, CDataType,
AccDataType, AccDataType,
AElementOp, AElementOp,
BElementOp, BElementOp,
......
example/04_gemm_add_add_fastgelu/gemm_add_add_fastgelu_xdl_fp32.cpp
View file @ aaf4defa
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved. // Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
#include "common.hpp" #include "common.hpp"
...@@ -7,10 +6,11 @@ using ADataType = F32; ...@@ -7,10 +6,11 @@ using ADataType = F32;
using BDataType = F32; using BDataType = F32;
using AccDataType = F32; using AccDataType = F32;
using CShuffleDataType = F32; using CShuffleDataType = F32;
using D0DataType = F32; using CDataType = F32; // C matrix doesn't exsit in GPU memory, this is used for host verification
using D1DataType = F32; using D0DataType = F32;
using DsDataType = ck::Tuple<D0DataType, D1DataType>; using D1DataType = F32;
using EDataType = F32; using DsDataType = ck::Tuple<D0DataType, D1DataType>;
using EDataType = F32;
using ALayout = Row; using ALayout = Row;
using BLayout = Col; using BLayout = Col;
...@@ -36,7 +36,7 @@ using DeviceOpInstance = ck::tensor_operation::device::DeviceGemmMultipleD_Xdl_C ...@@ -36,7 +36,7 @@ using DeviceOpInstance = ck::tensor_operation::device::DeviceGemmMultipleD_Xdl_C
using ReferenceGemmInstance = ck::tensor_operation::host::ReferenceGemm<ADataType, using ReferenceGemmInstance = ck::tensor_operation::host::ReferenceGemm<ADataType,
BDataType, BDataType,
AccDataType, CDataType,
AccDataType, AccDataType,
AElementOp, AElementOp,
BElementOp, BElementOp,
......
example/04_gemm_add_add_fastgelu/gemm_add_add_fastgelu_xdl_int4.cpp
View file @ aaf4defa
...@@ -11,10 +11,11 @@ using ADataType = I4; ...@@ -11,10 +11,11 @@ using ADataType = I4;
using BDataType = I4; using BDataType = I4;
using AccDataType = I32; using AccDataType = I32;
using CShuffleDataType = I32; using CShuffleDataType = I32;
using D0DataType = I4; using CDataType = I32; // C matrix doesn't exsit in GPU memory, this is used for host verification
using D1DataType = I4; using D0DataType = I4;
using DsDataType = ck::Tuple<D0DataType, D1DataType>; using D1DataType = I4;
using EDataType = I4; using DsDataType = ck::Tuple<D0DataType, D1DataType>;
using EDataType = I4;
using KernelADataType = I8; using KernelADataType = I8;
using KernelBDataType = I8; using KernelBDataType = I8;
...@@ -47,7 +48,7 @@ using DeviceOpInstance = ck::tensor_operation::device::DeviceGemmMultipleD_Xdl_C ...@@ -47,7 +48,7 @@ using DeviceOpInstance = ck::tensor_operation::device::DeviceGemmMultipleD_Xdl_C
using ReferenceGemmInstance = ck::tensor_operation::host::ReferenceGemm<ADataType, using ReferenceGemmInstance = ck::tensor_operation::host::ReferenceGemm<ADataType,
BDataType, BDataType,
AccDataType, CDataType,
AccDataType, AccDataType,
AElementOp, AElementOp,
BElementOp, BElementOp,
......
example/04_gemm_add_add_fastgelu/gemm_add_add_fastgelu_xdl_int8.cpp
View file @ aaf4defa
...@@ -7,10 +7,11 @@ using ADataType = I8; ...@@ -7,10 +7,11 @@ using ADataType = I8;
using BDataType = I8; using BDataType = I8;
using AccDataType = I32; using AccDataType = I32;
using CShuffleDataType = I32; using CShuffleDataType = I32;
using D0DataType = I8; using CDataType = I32; // C matrix doesn't exsit in GPU memory, this is used for host verification
using D1DataType = I8; using D0DataType = I8;
using DsDataType = ck::Tuple<D0DataType, D1DataType>; using D1DataType = I8;
using EDataType = I8; using DsDataType = ck::Tuple<D0DataType, D1DataType>;
using EDataType = I8;
using ALayout = Row; using ALayout = Row;
using BLayout = Col; using BLayout = Col;
...@@ -36,7 +37,7 @@ using DeviceOpInstance = ck::tensor_operation::device::DeviceGemmMultipleD_Xdl_C ...@@ -36,7 +37,7 @@ using DeviceOpInstance = ck::tensor_operation::device::DeviceGemmMultipleD_Xdl_C
using ReferenceGemmInstance = ck::tensor_operation::host::ReferenceGemm<ADataType, using ReferenceGemmInstance = ck::tensor_operation::host::ReferenceGemm<ADataType,
BDataType, BDataType,
AccDataType, CDataType,
AccDataType, AccDataType,
AElementOp, AElementOp,
BElementOp, BElementOp,
......
example/04_gemm_add_add_fastgelu/run_gemm_add_add_fastgelu_example.inc
View file @ aaf4defa
...@@ -124,7 +124,7 @@ bool run_gemm_add_add_fastgelu(const ProblemSize& problem_size, const ExecutionC ...@@ -124,7 +124,7 @@ bool run_gemm_add_add_fastgelu(const ProblemSize& problem_size, const ExecutionC
if(config.do_verification) if(config.do_verification)
{ {
Tensor<AccDataType> c_m_n({M, N}); Tensor<CDataType> c_m_n({M, N});
auto ref_gemm = ReferenceGemmInstance{}; auto ref_gemm = ReferenceGemmInstance{};
auto ref_invoker = ref_gemm.MakeInvoker(); auto ref_invoker = ref_gemm.MakeInvoker();
......
example/14_gemm_quantization/CMakeLists.txt
View file @ aaf4defa
# dlops
add_example_executable(example_gemm_dl_quantization_int8 gemm_dl_quantization_int8.cpp)
# xdlops
add_example_executable(example_gemm_xdl_bias_relu_quantization_int8 gemm_xdl_bias_relu_quantization_int8.cpp) add_example_executable(example_gemm_xdl_bias_relu_quantization_int8 gemm_xdl_bias_relu_quantization_int8.cpp)
add_example_executable(example_gemm_xdl_quantization_int8 gemm_xdl_quantization_int8.cpp) add_example_executable(example_gemm_xdl_quantization_int8 gemm_xdl_quantization_int8.cpp)
\ No newline at end of file
example/14_gemm_quantization/gemm_dl_quantization_int8.cpp 0 → 100644
View file @ aaf4defa
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
#include <iostream>
#include <numeric>
#include <initializer_list>
#include <cstdlib>
#include "ck/ck.hpp"
#include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
#include "ck/tensor_operation/gpu/device/gemm_specialization.hpp"
#include "ck/tensor_operation/gpu/device/impl/device_gemm_multiple_d_dl.hpp"
#include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
#include "ck/library/utility/device_memory.hpp"
#include "ck/library/utility/host_tensor.hpp"
#include "ck/library/utility/host_tensor_generator.hpp"
#include "ck/library/utility/literals.hpp"
#include "ck/library/reference_tensor_operation/cpu/reference_gemm.hpp"
#include "ck/library/utility/check_err.hpp"
template <ck::index_t... Is>
using S = ck::Sequence<Is...>;
using I8 = int8_t;
using I32 = int32_t;
using Row = ck::tensor_layout::gemm::RowMajor;
using Col = ck::tensor_layout::gemm::ColumnMajor;
using PassThrough = ck::tensor_operation::element_wise::PassThrough;
using AElementOp = PassThrough;
using BElementOp = PassThrough;
using ActivationOp = PassThrough;
using CDEElementOp = ck::tensor_operation::element_wise::Activation_Mul_Clamp<ActivationOp>;
using ADataType = I8;
using BDataType = I8;
using AccDataType = I32;
using CShuffleDataType = I32;
using DsDataType = ck::Tuple<>;
using EDataType = I8;
using ALayout = Row;
using BLayout = Col;
using DsLayout = ck::Tuple<>;
using ELayout = Row;
static constexpr auto GemmDefault = ck::tensor_operation::device::GemmSpecialization::Default;
using DeviceGemmInstance = ck::tensor_operation::device::DeviceGemmMultipleD_Dl<
ALayout,
BLayout,
DsLayout,
ELayout,
ADataType,
BDataType,
AccDataType,
DsDataType,
EDataType,
AElementOp,
BElementOp,
CDEElementOp,
GemmDefault,
256, // BlockSize
128, // MPerBlock
128, // NPerBlock
16, // K0PerBlock
4, // K1
4, // M1PerThread
4, // N1PerThread
1, // KPerThread
S<8, 2>, // M1N1ThreadClusterM1Xs
S<8, 2>, // M1N1ThreadClusterN1Xs
S<8, 1, 1, 4>, // ABlockTransferThreadSliceLengths_K0_M0_M1_K1
S<2, 1, 128, 1>, // ABlockTransferThreadClusterLengths_K0_M0_M1_K1
S<1, 2, 0, 3>, // ABlockTransferThreadClusterArrangeOrder
S<1, 2, 0, 3>, // ABlockTransferSrcAccessOrder
S<4, 1, 1, 4>, // ABlockTransferSrcVectorTensorLengths_K0_M0_M1_K1
S<1, 2, 0, 3>, // ABlockTransferSrcVectorTensorContiguousDimOrder
S<1, 1, 1, 4>, // ABlockTransferDstVectorTensorLengths_K0_M0_M1_K1
S<8, 1, 1, 4>, // BBlockTransferThreadSliceLengths_K0_N0_N1_K1
S<2, 1, 128, 1>, // BBlockTransferThreadClusterLengths_K0_N0_N1_K1
S<1, 2, 0, 3>, // BBlockTransferThreadClusterArrangeOrder
S<1, 2, 0, 3>, // BBlockTransferSrcAccessOrder
S<4, 1, 1, 4>, // BBlockTransferSrcVectorTensorLengths_K0_N0_N1_K1
S<1, 2, 0, 3>, // BBlockTransferSrcVectorTensorContiguousDimOrder
S<1, 1, 1, 4>, // BBlockTransferDstVectorTensorLengths_K0_N0_N1_K1
S<0, 1, 2, 3, 4, 5>, // CThreadTransferSrcDstAccessOrder
5, // CThreadTransferSrcDstVectorDim
4>; // CThreadTransferDstScalarPerVector
using ReferenceGemmInstance = ck::tensor_operation::host::
ReferenceGemm<ADataType, BDataType, EDataType, float, PassThrough, PassThrough, CDEElementOp>;
int main()
{
bool do_verification = true;
bool time_kernel = false;
// GEMM shape
ck::index_t M = 1024;
ck::index_t N = 1024;
ck::index_t K = 1024;
ck::index_t StrideA = 1024;
ck::index_t StrideB = 1024;
ck::index_t StrideE = 1024;
float requant_scale = 0.03;
auto f_host_tensor_descriptor =
[](std::size_t row, std::size_t col, std::size_t stride, auto layout) {
using namespace ck::literals;
if(std::is_same<decltype(layout), ck::tensor_layout::gemm::RowMajor>::value)
{
return HostTensorDescriptor(std::vector<std::size_t>({row, col}),
std::vector<std::size_t>({stride, 1_uz}));
}
else
{
return HostTensorDescriptor(std::vector<std::size_t>({row, col}),
std::vector<std::size_t>({1_uz, stride}));
}
};
Tensor<ADataType> a_m_k(f_host_tensor_descriptor(M, K, StrideA, ALayout{}));
Tensor<BDataType> b_k_n(f_host_tensor_descriptor(K, N, StrideB, BLayout{}));
Tensor<EDataType> e_m_n_host_result(f_host_tensor_descriptor(M, N, StrideE, ELayout{}));
Tensor<EDataType> e_m_n_device_result(f_host_tensor_descriptor(M, N, StrideE, ELayout{}));
std::cout << "a_m_k: " << a_m_k.mDesc << std::endl;
std::cout << "b_k_n: " << b_k_n.mDesc << std::endl;
std::cout << "e_m_n: " << e_m_n_host_result.mDesc << std::endl;
a_m_k.GenerateTensorValue(GeneratorTensor_2<ADataType>{-5, 5});
b_k_n.GenerateTensorValue(GeneratorTensor_2<BDataType>{-5, 5});
DeviceMem a_device_buf(sizeof(ADataType) * a_m_k.mDesc.GetElementSpaceSize());
DeviceMem b_device_buf(sizeof(BDataType) * b_k_n.mDesc.GetElementSpaceSize());
DeviceMem e_device_buf(sizeof(EDataType) * e_m_n_device_result.mDesc.GetElementSpaceSize());
a_device_buf.ToDevice(a_m_k.mData.data());
b_device_buf.ToDevice(b_k_n.mData.data());
auto a_element_op = AElementOp{};
auto b_element_op = BElementOp{};
auto cde_element_op = CDEElementOp{requant_scale, ActivationOp{}};
// do GEMM
auto gemm = DeviceGemmInstance{};
auto invoker = gemm.MakeInvoker();
auto argument = gemm.MakeArgument(a_device_buf.GetDeviceBuffer(),
b_device_buf.GetDeviceBuffer(),
{},
e_device_buf.GetDeviceBuffer(),
M,
N,
K,
StrideA,
StrideB,
{},
StrideE,
a_element_op,
b_element_op,
cde_element_op);
if(!gemm.IsSupportedArgument(argument))
{
throw std::runtime_error(
"wrong! device_gemm with the specified compilation parameters does "
"not support this GEMM problem");
}
float ave_time = invoker.Run(argument, StreamConfig{nullptr, time_kernel});
std::size_t flop = std::size_t(2) * M * N * K;
std::size_t num_btype =
sizeof(ADataType) * M * K + sizeof(BDataType) * K * N + sizeof(EDataType) * M * N;
float tflops = static_cast<float>(flop) / 1.E9 / ave_time;
float gb_per_sec = num_btype / 1.E6 / ave_time;
std::cout << "Perf: " << ave_time << " ms, " << tflops << " TFlops, " << gb_per_sec << " GB/s, "
<< gemm.GetTypeString() << std::endl;
e_device_buf.FromDevice(e_m_n_device_result.mData.data());
if(do_verification)
{
auto ref_gemm = ReferenceGemmInstance{};
auto ref_invoker = ref_gemm.MakeInvoker();
auto ref_argument = ref_gemm.MakeArgument(
a_m_k, b_k_n, e_m_n_host_result, a_element_op, b_element_op, cde_element_op);
ref_invoker.Run(ref_argument);
return ck::utils::check_err(e_m_n_device_result, e_m_n_host_result) ? 0 : 1;
}
return 0;
}
example/15_grouped_gemm/CMakeLists.txt
View file @ aaf4defa
...@@ -4,12 +4,15 @@ add_example_executable(example_grouped_gemm_xdl_fp32 grouped_gemm_xdl_fp32.cpp) ...@@ -4,12 +4,15 @@ add_example_executable(example_grouped_gemm_xdl_fp32 grouped_gemm_xdl_fp32.cpp)
add_example_executable(example_grouped_gemm_xdl_fp16 grouped_gemm_xdl_fp16.cpp) add_example_executable(example_grouped_gemm_xdl_fp16 grouped_gemm_xdl_fp16.cpp)
add_example_executable(example_grouped_gemm_xdl_bfp16 grouped_gemm_xdl_bfp16.cpp) add_example_executable(example_grouped_gemm_xdl_bfp16 grouped_gemm_xdl_bfp16.cpp)
add_example_executable(example_grouped_gemm_xdl_int8 grouped_gemm_xdl_int8.cpp) add_example_executable(example_grouped_gemm_xdl_int8 grouped_gemm_xdl_int8.cpp)
add_example_executable(example_grouped_gemm_multiple_d_dl_fp16 grouped_gemm_multiple_d_dl_fp16.cpp)
add_dependencies(example_grouped_gemm_xdl add_dependencies(example_grouped_gemm_xdl
example_grouped_gemm_xdl_fp32 example_grouped_gemm_xdl_fp32
example_grouped_gemm_xdl_fp16 example_grouped_gemm_xdl_fp16
example_grouped_gemm_xdl_bfp16 example_grouped_gemm_xdl_bfp16
example_grouped_gemm_xdl_int8) example_grouped_gemm_xdl_int8
example_grouped_gemm_multiple_d_dl_fp16)
if(USE_BITINT_EXTENSION_INT4) if(USE_BITINT_EXTENSION_INT4)
add_example_executable(example_grouped_gemm_xdl_int4 grouped_gemm_xdl_int4.cpp) add_example_executable(example_grouped_gemm_xdl_int4 grouped_gemm_xdl_int4.cpp)
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment