Skip to content

GitLab

Commit d0b49a14 authored by Qianfeng Zhang's avatar Qianfeng Zhang
Browse files

Merge branch 'develop' into bnorm_bwd_pr

parents 29026b0e 87fd1152
Hide whitespace changes
Inline Side-by-side
Showing with 627 additions and 156 deletions
CITATION.cff 0 → 100644
View file @ d0b49a14
cff-version: 1.2.0
title: Composable Kernel
message: If you use this software, please cite using the following metadata.
type: software
authors:
- given-names: Chao
family-names: Liu
email: chao.liu2@amd.com
affiliation: AMD
- given-names: Jing
family-names: Zhang
email: jing.zhang3@amd.com
affiliation: AMD
- given-names: Letao
family-names: Qin
email: letao.qin@amd.com
affiliation: AMD
- given-names: Qianfeng
family-names: Zhang
email: qianfeng.zhang@amd.com
affiliation: AMD
- given-names: Liang
family-names: Huang
email: carlus.huang@amd.com
affiliation: AMD
- given-names: Shaojie
family-names: Wang
email: shaojie.wang@amd.com
affiliation: AMD
- given-names: Anthony
family-names: Chang
email: antc@amd.com
affiliation: AMD
- given-names: Chunyu
family-names: Lai
email: chunyu.lai@amd.com
affiliation: AMD
- given-names: Illia
family-names: Silin
email: illia.silin@amd.com
affiliation: AMD
- given-names: Adam
family-names: Osewski
email: adam.osewski@amd.com
affiliation: AMD
- given-names: Poyen
family-names: Chen
email: poyen.chen@amd.com
affiliation: AMD
- given-names: Rosty
family-names: Geyyer
email: rosty.geyyer@amd.com
affiliation: AMD
- given-names: Hanwen
family-names: Chen
- given-names: Tejash
family-names: Shah
- given-names: Xiaoyan
family-names: Zhou
- given-names: Jianfeng
family-names: Yan
repository-code: 'https://github.com/ROCmSoftwarePlatform/composable_kernel'
abstract: Composable Kernel (CK) library aims to provide a programming model for writing performance critical kernels for Machine Learning workloads across multiple architectures including GPUs, CPUs, etc, through general purpose kernel progarmming languages, like HIP C++.
keywords:
- 'CK, Composable Kernel, Tensor Coordinate Transformation'
license: MIT
license-url: https://github.com/ROCmSoftwarePlatform/composable_kernel/blob/7fc3ed761aa35709d87c8fbbe41dd368648b3541/LICENSE
CMakeLists.txt
View file @ d0b49a14
...@@ -247,6 +247,16 @@ message("CMAKE_CXX_FLAGS: ${CMAKE_CXX_FLAGS}") ...@@ -247,6 +247,16 @@ message("CMAKE_CXX_FLAGS: ${CMAKE_CXX_FLAGS}")
add_custom_target(check COMMAND ${CMAKE_CTEST_COMMAND} --output-on-failure -C ${CMAKE_CFG_INTDIR}) add_custom_target(check COMMAND ${CMAKE_CTEST_COMMAND} --output-on-failure -C ${CMAKE_CFG_INTDIR})
file(GLOB_RECURSE INSTANCE_FILES "${PROJECT_SOURCE_DIR}/*/device_*_instance.cpp")
file(GLOB dir_list RELATIVE ${PROJECT_SOURCE_DIR}/library/src/tensor_operation_instance/gpu ${PROJECT_SOURCE_DIR}/library/src/tensor_operation_instance/gpu/*)
set(CK_DEVICE_INSTANCES)
FOREACH(subdir_path ${dir_list})
IF(IS_DIRECTORY "${PROJECT_SOURCE_DIR}/library/src/tensor_operation_instance/gpu/${subdir_path}")
list(APPEND CK_DEVICE_INSTANCES device_${subdir_path}_instance)
ENDIF()
ENDFOREACH()
add_custom_target(instances DEPENDS utility;${CK_DEVICE_INSTANCES} SOURCES ${INSTANCE_FILES})
rocm_package_setup_component(tests rocm_package_setup_component(tests
LIBRARY_NAME composablekernel LIBRARY_NAME composablekernel
PACKAGE_NAME tests # Prevent -static suffix on package name PACKAGE_NAME tests # Prevent -static suffix on package name
......
CONTRIBUTORS.md 0 → 100644
View file @ d0b49a14
# Composable Kernel Developers and Contributors
This is the list of developers and contributors to Composable Kernel library
## Developers
[Chao Liu](https://github.com/asroy), [Jing Zhang](https://github.com/zjing14), 2018-2022
[Letao Qin](https://github.com/ltqin), [Qianfeng Zhang](https://github.com/qianfengz), [Liang Huang](https://github.com/carlushuang), [Shaojie Wang](https://github.com/shaojiewang), 2019-2022
[Anthony Chang](https://github.com/rosenrodt), [Chunyu Lai](https://github.com/rocking5566), [Illia Silin](https://github.com/illsilin), [Adam Osewski](https://github.com/aosewski), [Poyen Chen](https://github.com/poyenc), [Rosty Geyyer](https://github.com/geyyer), 2022
Hanwen Chang, 2019-2021,
Tejash Shah, 2019-2020
Xiaoyan Zhou, 2020
[Jianfeng Yan](https://github.com/j4yan), 2021-2022
## Product Manager
[Jun Liu](https://github.com/junliume)
## Contributors
[Dan Yao](https://github.com/danyao12), [Guangzhao Lu](https://github.com/guangzlu), [Raman Jana](https://github.com/ramjana), [Jehandad Khan](https://github.com/JehandadKhan), [Wen-Heng (Jack) Chung](https://github.com/whchung)
## Acknowledgement
CK team works closely with Meta [AITemplate](https://github.com/facebookincubator/AITemplate) team ([Bing Xu](https://github.com/antinucleon), [Hao Lu](https://github.com/hlu1), [Ying Zhang](https://github.com/ipiszy), etc). Most of the lucrative graph optimization opportunities in ML models were identified by AITemplate team, and we also co-designed many high performance fused kernels for AMD GPUs. Without this collaboration, CK would not reach its current potential.
Dockerfile
View file @ d0b49a14
FROM ubuntu:20.04 FROM ubuntu:20.04
ARG ROCMVERSION=5.2.3 ARG ROCMVERSION=5.3
ARG compiler_version ARG compiler_version="release"
ARG compiler_commit ARG compiler_commit=""
RUN set -xe RUN set -xe
...@@ -19,6 +19,7 @@ RUN sh -c "echo deb http://mirrors.kernel.org/ubuntu focal main universe | tee - ...@@ -19,6 +19,7 @@ RUN sh -c "echo deb http://mirrors.kernel.org/ubuntu focal main universe | tee -
RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y --allow-unauthenticated \ RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y --allow-unauthenticated \
apt-utils \ apt-utils \
build-essential \ build-essential \
ccache \
cmake-data \ cmake-data \
cmake \ cmake \
curl \ curl \
......
Jenkinsfile
View file @ d0b49a14
...@@ -19,10 +19,24 @@ def runShell(String command){ ...@@ -19,10 +19,24 @@ def runShell(String command){
} }
def getDockerImageName(){ def getDockerImageName(){
def img = "${env.CK_DOCKERHUB}:ck_ub20.04_rocm5.2.3_${params.COMPILER_VERSION}" def img = "${env.CK_DOCKERHUB}:ck_ub20.04_rocm${params.ROCMVERSION}_${params.COMPILER_VERSION}"
return img return img
} }
def check_host() {
if ("${env.CK_CCACHE}" != "null"){
def CCACHE_SERVER="${env.CK_CCACHE.split(':')[0]}"
echo "ccache server: ${CCACHE_SERVER}"
sh '''ping -c 1 -p 6379 "${CCACHE_SERVER}" | echo $? > tmp.txt'''
def output = readFile(file: "tmp.txt")
echo "tmp.txt contents: \$output"
return (output != "0")
}
else{
return 1
}
}
def build_compiler(){ def build_compiler(){
def compiler def compiler
if (params.BUILD_COMPILER == "hipcc"){ if (params.BUILD_COMPILER == "hipcc"){
...@@ -43,21 +57,21 @@ def getDockerImage(Map conf=[:]){ ...@@ -43,21 +57,21 @@ def getDockerImage(Map conf=[:]){
env.DOCKER_BUILDKIT=1 env.DOCKER_BUILDKIT=1
def prefixpath = conf.get("prefixpath", "/opt/rocm") // prefix:/opt/rocm def prefixpath = conf.get("prefixpath", "/opt/rocm") // prefix:/opt/rocm
def no_cache = conf.get("no_cache", false) def no_cache = conf.get("no_cache", false)
def dockerArgs = "--build-arg BUILDKIT_INLINE_CACHE=1 --build-arg PREFIX=${prefixpath} --build-arg compiler_version='${params.COMPILER_VERSION}' --build-arg compiler_commit='${params.COMPILER_COMMIT}' " def dockerArgs = "--build-arg BUILDKIT_INLINE_CACHE=1 --build-arg PREFIX=${prefixpath} --build-arg compiler_version='${params.COMPILER_VERSION}' --build-arg compiler_commit='${params.COMPILER_COMMIT}' --build-arg ROCMVERSION='${params.ROCMVERSION}' "
if(env.CCACHE_HOST) echo "ccache server: ${env.CK_CCACHE}"
if(env.CK_CCACHE)
{ {
def check_host = sh(script:"""(printf "PING\r\n";) | nc -N ${env.CCACHE_HOST} 6379 """, returnStdout: true).trim() if(check_host())
if(check_host == "+PONG")
{ {
echo "FOUND CCACHE SERVER: ${CCACHE_HOST}" echo "FOUND CCACHE SERVER: ${env.CK_CCACHE}"
} }
else else
{ {
echo "CCACHE SERVER: ${CCACHE_HOST} NOT FOUND, got ${check_host} response" echo "CCACHE SERVER: ${env.CK_CCACHE} NOT FOUND, got ${check_host} response"
} }
dockerArgs = dockerArgs + " --build-arg CCACHE_SECONDARY_STORAGE='redis://${env.CCACHE_HOST}' --build-arg COMPILER_LAUNCHER='ccache' " dockerArgs = dockerArgs + " --build-arg CCACHE_SECONDARY_STORAGE='redis://${env.CK_CCACHE}' --build-arg COMPILER_LAUNCHER='ccache' "
env.CCACHE_DIR = """/tmp/ccache_store""" env.CCACHE_DIR = """/tmp/ccache_store"""
env.CCACHE_SECONDARY_STORAGE="""redis://${env.CCACHE_HOST}""" env.CCACHE_SECONDARY_STORAGE="""redis://${env.CK_CCACHE}"""
} }
if(no_cache) if(no_cache)
{ {
...@@ -86,21 +100,21 @@ def buildDocker(install_prefix){ ...@@ -86,21 +100,21 @@ def buildDocker(install_prefix){
checkout scm checkout scm
def image_name = getDockerImageName() def image_name = getDockerImageName()
echo "Building Docker for ${image_name}" echo "Building Docker for ${image_name}"
def dockerArgs = "--build-arg BUILDKIT_INLINE_CACHE=1 --build-arg PREFIX=${install_prefix} --build-arg compiler_version='${params.COMPILER_VERSION}' --build-arg compiler_commit='${params.COMPILER_COMMIT}'" def dockerArgs = "--build-arg BUILDKIT_INLINE_CACHE=1 --build-arg PREFIX=${install_prefix} --build-arg compiler_version='${params.COMPILER_VERSION}' --build-arg compiler_commit='${params.COMPILER_COMMIT}' --build-arg ROCMVERSION='${params.ROCMVERSION}' "
if(env.CCACHE_HOST) echo "ccache server: ${env.CK_CCACHE}"
if(env.CK_CCACHE)
{ {
def check_host = sh(script:"""(printf "PING\\r\\n";) | nc -N ${env.CCACHE_HOST} 6379 """, returnStdout: true).trim() if(check_host())
if(check_host == "+PONG")
{ {
echo "FOUND CCACHE SERVER: ${CCACHE_HOST}" echo "FOUND CCACHE SERVER: ${env.CK_CCACHE}"
} }
else else
{ {
echo "CCACHE SERVER: ${CCACHE_HOST} NOT FOUND, got ${check_host} response" echo "CCACHE SERVER: ${env.CK_CCACHE} NOT FOUND, got ${check_host} response"
} }
dockerArgs = dockerArgs + " --build-arg CCACHE_SECONDARY_STORAGE='redis://${env.CCACHE_HOST}' --build-arg COMPILER_LAUNCHER='ccache' " dockerArgs = dockerArgs + " --build-arg CCACHE_SECONDARY_STORAGE='redis://${env.CK_CCACHE}' --build-arg COMPILER_LAUNCHER='ccache' "
env.CCACHE_DIR = """/tmp/ccache_store""" env.CCACHE_DIR = """/tmp/ccache_store"""
env.CCACHE_SECONDARY_STORAGE="""redis://${env.CCACHE_HOST}""" env.CCACHE_SECONDARY_STORAGE="""redis://${env.CK_CCACHE}"""
} }
echo "Build Args: ${dockerArgs}" echo "Build Args: ${dockerArgs}"
...@@ -161,10 +175,11 @@ def cmake_build(Map conf=[:]){ ...@@ -161,10 +175,11 @@ def cmake_build(Map conf=[:]){
}else{ }else{
setup_args = " -DCMAKE_BUILD_TYPE=release" + setup_args setup_args = " -DCMAKE_BUILD_TYPE=release" + setup_args
} }
if(env.CCACHE_HOST) if(env.CK_CCACHE)
{ {
setup_args = " -DCMAKE_CXX_COMPILER_LAUNCHER='ccache' -DCMAKE_C_COMPILER_LAUNCHER='ccache' " + setup_args setup_args = " -DCMAKE_CXX_COMPILER_LAUNCHER='ccache' -DCMAKE_C_COMPILER_LAUNCHER='ccache' " + setup_args
} }
echo "ccache server: ${env.CK_CCACHE}"
def pre_setup_cmd = """ def pre_setup_cmd = """
echo \$HSA_ENABLE_SDMA echo \$HSA_ENABLE_SDMA
...@@ -210,7 +225,7 @@ def buildHipClangJob(Map conf=[:]){ ...@@ -210,7 +225,7 @@ def buildHipClangJob(Map conf=[:]){
if (conf.get("enforce_xnack_on", false)) { if (conf.get("enforce_xnack_on", false)) {
dockerOpts = dockerOpts + " --env HSA_XNACK=1 " dockerOpts = dockerOpts + " --env HSA_XNACK=1 "
} }
def dockerArgs = "--build-arg PREFIX=${prefixpath} --build-arg compiler_version='${params.COMPILER_VERSION}' --build-arg compiler_commit='${params.COMPILER_COMMIT}' " def dockerArgs = "--build-arg PREFIX=${prefixpath} --build-arg compiler_version='${params.COMPILER_VERSION}' --build-arg compiler_commit='${params.COMPILER_COMMIT}' --build-arg ROCMVERSION='${params.ROCMVERSION}' "
if (params.COMPILER_VERSION != "release"){ if (params.COMPILER_VERSION != "release"){
dockerOpts = dockerOpts + " --env HIP_CLANG_PATH='/llvm-project/build/bin' " dockerOpts = dockerOpts + " --env HIP_CLANG_PATH='/llvm-project/build/bin' "
} }
...@@ -218,41 +233,9 @@ def buildHipClangJob(Map conf=[:]){ ...@@ -218,41 +233,9 @@ def buildHipClangJob(Map conf=[:]){
def variant = env.STAGE_NAME def variant = env.STAGE_NAME
def retimage def retimage
(retimage, image) = getDockerImage(conf)
gitStatusWrapper(credentialsId: "${status_wrapper_creds}", gitHubContext: "Jenkins - ${variant}", account: 'ROCmSoftwarePlatform', repo: 'composable_kernel') { gitStatusWrapper(credentialsId: "${status_wrapper_creds}", gitHubContext: "Jenkins - ${variant}", account: 'ROCmSoftwarePlatform', repo: 'composable_kernel') {
try {
(retimage, image) = getDockerImage(conf)
withDockerContainer(image: image, args: dockerOpts) {
timeout(time: 5, unit: 'MINUTES'){
sh 'PATH="/opt/rocm/opencl/bin:/opt/rocm/opencl/bin/x86_64:$PATH" clinfo | tee clinfo.log'
if ( runShell('grep -n "Number of devices:.*. 0" clinfo.log') ){
throw new Exception ("GPU not found")
}
else{
echo "GPU is OK"
}
}
}
}
catch (org.jenkinsci.plugins.workflow.steps.FlowInterruptedException e){
echo "The job was cancelled or aborted"
throw e
}
catch(Exception ex) {
retimage = docker.build("${image}", dockerArgs + " --no-cache .")
withDockerContainer(image: image, args: dockerOpts) {
timeout(time: 5, unit: 'MINUTES'){
sh 'PATH="/opt/rocm/opencl/bin:/opt/rocm/opencl/bin/x86_64:$PATH" clinfo |tee clinfo.log'
if ( runShell('grep -n "Number of devices:.*. 0" clinfo.log') ){
throw new Exception ("GPU not found")
}
else{
echo "GPU is OK"
}
}
}
}
withDockerContainer(image: image, args: dockerOpts + ' -v=/var/jenkins/:/var/jenkins') { withDockerContainer(image: image, args: dockerOpts + ' -v=/var/jenkins/:/var/jenkins') {
timeout(time: 5, unit: 'HOURS') timeout(time: 5, unit: 'HOURS')
{ {
...@@ -297,7 +280,7 @@ def runCKProfiler(Map conf=[:]){ ...@@ -297,7 +280,7 @@ def runCKProfiler(Map conf=[:]){
if (conf.get("enforce_xnack_on", false)) { if (conf.get("enforce_xnack_on", false)) {
dockerOpts = dockerOpts + " --env HSA_XNACK=1 " dockerOpts = dockerOpts + " --env HSA_XNACK=1 "
} }
def dockerArgs = "--build-arg PREFIX=${prefixpath} --build-arg compiler_version='${params.COMPILER_VERSION}' --build-arg compiler_commit='${params.COMPILER_COMMIT}' " def dockerArgs = "--build-arg PREFIX=${prefixpath} --build-arg compiler_version='${params.COMPILER_VERSION}' --build-arg compiler_commit='${params.COMPILER_COMMIT}' --build-arg ROCMVERSION='${params.ROCMVERSION}' "
if (params.COMPILER_VERSION != "release"){ if (params.COMPILER_VERSION != "release"){
dockerOpts = dockerOpts + " --env HIP_CLANG_PATH='/llvm-project/build/bin' " dockerOpts = dockerOpts + " --env HIP_CLANG_PATH='/llvm-project/build/bin' "
} }
...@@ -430,7 +413,7 @@ def Build_CK(Map conf=[:]){ ...@@ -430,7 +413,7 @@ def Build_CK(Map conf=[:]){
if (conf.get("enforce_xnack_on", false)) { if (conf.get("enforce_xnack_on", false)) {
dockerOpts = dockerOpts + " --env HSA_XNACK=1 " dockerOpts = dockerOpts + " --env HSA_XNACK=1 "
} }
def dockerArgs = "--build-arg PREFIX=${prefixpath} --build-arg compiler_version='${params.COMPILER_VERSION}' --build-arg compiler_commit='${params.COMPILER_COMMIT}' " def dockerArgs = "--build-arg PREFIX=${prefixpath} --build-arg compiler_version='${params.COMPILER_VERSION}' --build-arg compiler_commit='${params.COMPILER_COMMIT}' --build-arg ROCMVERSION='${params.ROCMVERSION}' "
if (params.COMPILER_VERSION != "release"){ if (params.COMPILER_VERSION != "release"){
dockerOpts = dockerOpts + " --env HIP_CLANG_PATH='/llvm-project/build/bin' " dockerOpts = dockerOpts + " --env HIP_CLANG_PATH='/llvm-project/build/bin' "
} }
...@@ -566,8 +549,9 @@ def process_results(Map conf=[:]){ ...@@ -566,8 +549,9 @@ def process_results(Map conf=[:]){
} }
} }
//launch develop branch daily at 23:00 in FULL_QA mode //launch develop branch daily at 23:00 UT in FULL_QA mode and at 19:00 UT with latest staging compiler version
CRON_SETTINGS = BRANCH_NAME == "develop" ? '''0 23 * * * % RUN_FULL_QA=true''' : "" CRON_SETTINGS = BRANCH_NAME == "develop" ? '''0 23 * * * % RUN_FULL_QA=true;COMPILER_VERSION=release
0 19 * * * % BUILD_DOCKER=true;COMPILER_VERSION=amd-stg-open''' : ""
pipeline { pipeline {
agent none agent none
...@@ -581,15 +565,19 @@ pipeline { ...@@ -581,15 +565,19 @@ pipeline {
booleanParam( booleanParam(
name: "BUILD_DOCKER", name: "BUILD_DOCKER",
defaultValue: false, defaultValue: false,
description: "Force building docker image (default: false)") description: "Force building docker image (default: false), set to true if docker image needs to be updated.")
string(
name: 'ROCMVERSION',
defaultValue: '5.3',
description: 'Specify which ROCM version to use: 5.2.3, or 5.3 (default), etc.')
string( string(
name: 'COMPILER_VERSION', name: 'COMPILER_VERSION',
defaultValue: 'amd-stg-open', defaultValue: 'release',
description: 'Specify which version of compiler to use: ck-9110, release, or amd-stg-open (default).') description: 'Specify which version of compiler to use: ck-9110, release (default), or amd-stg-open.')
string( string(
name: 'COMPILER_COMMIT', name: 'COMPILER_COMMIT',
defaultValue: '8a82e4eb7ba28521ba9a9424a0315a8a16590424', defaultValue: '',
description: 'Specify which commit of compiler branch to use: leave empty to use the latest commit, or use 10738 commit (default).') description: 'Specify which commit of compiler branch to use: leave empty to use the latest commit (default), or use 8a82e4eb7ba28521ba9a9424a0315a8a16590424 commit of amd-stg-open branch.')
string( string(
name: 'BUILD_COMPILER', name: 'BUILD_COMPILER',
defaultValue: 'hipcc', defaultValue: 'hipcc',
...@@ -627,17 +615,6 @@ pipeline { ...@@ -627,17 +615,6 @@ pipeline {
} }
stage("Static checks") { stage("Static checks") {
parallel{ parallel{
// enable after we move from hipcc to hip-clang
// stage('Tidy') {
// agent{ label rocmnode("nogpu") }
// environment{
// // setup_cmd = "CXX='/opt/rocm/bin/hipcc' cmake -DBUILD_DEV=On .. "
// build_cmd = "make -j\$(nproc) -k analyze"
// }
// steps{
// buildHipClangJobAndReboot(build_cmd: build_cmd, no_reboot:true, prefixpath: '/opt/rocm', build_type: 'debug')
// }
// }
stage('Clang Format') { stage('Clang Format') {
agent{ label rocmnode("nogpu") } agent{ label rocmnode("nogpu") }
environment{ environment{
...@@ -676,28 +653,6 @@ pipeline { ...@@ -676,28 +653,6 @@ pipeline {
} }
} }
/*
//at present this stage only builds binaries.
//we will now build all binaries in a separate stage.
//once we have some tests to run in this stage, we can enable it again.
stage("Client App")
{
parallel
{
stage("Run Client App")
{
agent{ label rocmnode("gfx908")}
environment{
setup_args = "${params.COMPILER_VERSION == "ck-9110" ? """ -DBUILD_DEV=Off -DCMAKE_INSTALL_PREFIX=../install -DGPU_TARGETS="gfx908;gfx90a" -DCMAKE_CXX_FLAGS="-O3 -Xclang -mlink-builtin-bitcode -Xclang /opt/rocm/amdgcn/bitcode/oclc_abi_version_400.bc" """ : """ -DBUILD_DEV=Off -DCMAKE_INSTALL_PREFIX=../install -DGPU_TARGETS="gfx908;gfx90a" -DCMAKE_CXX_FLAGS="-O3 " """ }"
execute_args = "${params.COMPILER_VERSION == "ck-9110" ? """ cd ../client_example && rm -rf build && mkdir build && cd build && cmake -D CMAKE_PREFIX_PATH="${env.WORKSPACE}/install;/opt/rocm" -DGPU_TARGETS="gfx908;gfx90a" -DCMAKE_CXX_FLAGS="-O3 -Xclang -mlink-builtin-bitcode -Xclang /opt/rocm/amdgcn/bitcode/oclc_abi_version_400.bc" -D CMAKE_CXX_COMPILER="${build_compiler()}" .. && make -j """ : """ cd ../client_example && rm -rf build && mkdir build && cd build && cmake -D CMAKE_PREFIX_PATH="${env.WORKSPACE}/install;/opt/rocm" -DGPU_TARGETS="gfx908;gfx90a" -DCMAKE_CXX_FLAGS="-O3" -D CMAKE_CXX_COMPILER="${build_compiler()}" .. && make -j """ }"
}
steps{
buildHipClangJobAndReboot(setup_args: setup_args, config_targets: "install", no_reboot:true, build_type: 'Release', execute_cmd: execute_args, prefixpath: '/usr/local')
}
}
}
}
*/
stage("Performance Tests") stage("Performance Tests")
{ {
parallel parallel
...@@ -746,21 +701,5 @@ pipeline { ...@@ -746,21 +701,5 @@ pipeline {
} }
} }
} }
/* enable after the cmake file supports packaging
stage("Packages") {
when {
expression { params.BUILD_PACKAGES && params.TARGET_NOGPU && params.DATATYPE_NA }
}
parallel {
stage("Package /opt/rocm") {
agent{ label rocmnode("nogpu") }
steps{
buildHipClangJobAndReboot( package_build: "true", prefixpath: '/opt/rocm', gpu_arch: "gfx906;gfx908;gfx90a")
}
}
}
}
*/
} }
} }
README.md
View file @ d0b49a14
## Docker script # Composable Kernel
## Methodology
Composable Kernel (CK) library aims to provide a programming model for writing performance critical kernels for machine learning workloads across multiple architectures including GPUs, CPUs, etc, through general purpose kernel languages, like HIP C++.
CK utilizes two concepts to achieve performance portability and code maintainability:
* A tile-based programming model
* Algorithm complexity reduction for complex ML operators, using innovative technique we call "Tensor Coordinate Transformation".
![ALT](/doc/image/ck_component.png "CK Components")
## Code Structure
Current CK library are structured into 4 layers:
* "Templated Tile Operators" layer
* "Templated Kernel and Invoker" layer
* "Instantiated Kernel and Invoker" layer
* "Client API" layer
![ALT](/doc/image/ck_layer.png "CK Layers")
## Contributors
The list of developers and contributors is here: [Contributors](/CONTRIBUTORS.md)
## Citation
If you use CK, please use following citations:
* CK paper will be freely available on arXiv soon: [Realizing Tensor Operators Using Coordinate Transformations and Tile Based Programming](???)
* [CITATION.cff](/CITATION.cff)
## License
CK is released under the MIT license. [License File](/LICENSE)
# Build CK
## Build docker image
```bash
DOCKER_BUILDKIT=1 docker build -t ck:latest -f Dockerfile .
```
## Launch docker
```bash ```bash
docker run \ docker run \
-it \ -it \
...@@ -6,47 +45,38 @@ docker run \ ...@@ -6,47 +45,38 @@ docker run \
--group-add sudo \ --group-add sudo \
-w /root/workspace \ -w /root/workspace \
-v ${PATH_TO_LOCAL_WORKSPACE}:/root/workspace \ -v ${PATH_TO_LOCAL_WORKSPACE}:/root/workspace \
rocm/tensorflow:rocm5.1-tf2.6-dev \ ck:latest \
/bin/bash /bin/bash
``` ```
# Install newer version of rocm-cmake ## Build CK
https://github.com/RadeonOpenCompute/rocm-cmake
## Build
```bash ```bash
mkdir build && cd build mkdir build && cd build
```
```bash # Need to specify target ID, example below is for gfx908 and gfx90a
# Need to specify target ID, example below is gfx908 and gfx90a cmake \
cmake \ -D CMAKE_PREFIX_PATH=/opt/rocm \
-D BUILD_DEV=OFF \ -D CMAKE_CXX_COMPILER=/opt/rocm/bin/hipcc \
-D CMAKE_BUILD_TYPE=Release \ -D CMAKE_CXX_FLAGS="-O3" \
-D CMAKE_CXX_FLAGS=" --offload-arch=gfx908 --offload-arch=gfx90a -O3" \ -D CMAKE_BUILD_TYPE=Release \
-D CMAKE_CXX_COMPILER=/opt/rocm/bin/hipcc \ -D GPU_TARGETS="gfx908;gfx90a" \
-D CMAKE_PREFIX_PATH=/opt/rocm \
-D CMAKE_INSTALL_PREFIX=${PATH_TO_CK_INSTALL_DIRECTORY} \
.. ..
``` ```
### Build and Run Examples ### Build examples and tests
```bash
make -j examples
```
Instructions for running each individual examples are under ```example/```
## Tests
```bash ```bash
make -j examples tests make -j examples tests
make test make test
``` ```
Instructions for running each individual examples are under [example](/example)
## Build ckProfiler ## Build ckProfiler
```bash ```bash
make -j ckProfiler make -j ckProfiler
``` ```
Instructions for running ckProfiler are under ```profiler/``` Instructions for running ckProfiler are under [profiler](/profiler)
## Install CK ## Install CK
```bash ```bash
...@@ -54,13 +84,13 @@ make install ...@@ -54,13 +84,13 @@ make install
``` ```
## Using CK as pre-built kernel library ## Using CK as pre-built kernel library
Instructions for using CK as a pre-built kernel library are under ```client_example/``` Instructions for using CK as a pre-built kernel library are under [client_example](/client_example)
## Caveat ## Caveat
### Kernel Timing and Verification ### Kernel Timing and Verification
CK's own kernel timer will warn up kernel once, and then run it multiple times CK's own kernel timer will warn up kernel once, and then run it multiple times
to get average kernel time. For some kernels that use atomic add, this will cause to get average kernel time. For some kernels that use atomic add, this will cause
output buffer to be accumulated multiple times, causing verfication failure. output buffer to be accumulated multiple times, causing verification failure.
To work around it, do not use CK's own timer and do verification at the same time. To work around it, do not use CK's own timer and do verification at the same time.
CK's own timer and verification in each example and ckProfiler can be enabled or CK's own timer and verification in each example and ckProfiler can be enabled or
disabled from command line. disabled from command line.
client_example/03_gemm_layernorm/gemm_add_add_layernorm.cpp
View file @ d0b49a14
...@@ -8,7 +8,7 @@ ...@@ -8,7 +8,7 @@
#include "ck/ck.hpp" #include "ck/ck.hpp"
#include "ck/tensor_operation/gpu/device/tensor_layout.hpp" #include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
#include "ck/tensor_operation/gpu/device/device_gemm_reduce.hpp" #include "ck/tensor_operation/gpu/device/device_gemm_reduce.hpp"
#include "ck/tensor_operation/gpu/device/device_elementwise.hpp" #include "ck/tensor_operation/gpu/device/impl/device_elementwise.hpp"
#include "ck/tensor_operation/gpu/element/element_wise_operation.hpp" #include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
#include "ck/library/tensor_operation_instance/gpu/device_elementwise_instance.hpp" #include "ck/library/tensor_operation_instance/gpu/device_elementwise_instance.hpp"
......
client_example/05_layernorm/layernorm2d.cpp
View file @ d0b49a14
...@@ -10,7 +10,7 @@ ...@@ -10,7 +10,7 @@
#include "ck/tensor_operation/gpu/device/device_normalization.hpp" #include "ck/tensor_operation/gpu/device/device_normalization.hpp"
#include "ck/tensor_operation/gpu/element/element_wise_operation.hpp" #include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
#include "ck/library/tensor_operation_instance/gpu/layernorm.hpp" #include "ck/library/tensor_operation_instance/gpu/normalization.hpp"
using XDataType = ck::half_t; using XDataType = ck::half_t;
using GammaDataType = ck::half_t; using GammaDataType = ck::half_t;
...@@ -51,14 +51,14 @@ int main(int argc, char* argv[]) ...@@ -51,14 +51,14 @@ int main(int argc, char* argv[])
SimpleDeviceMem beta_device_buf(sizeof(BetaDataType) * N); SimpleDeviceMem beta_device_buf(sizeof(BetaDataType) * N);
SimpleDeviceMem y_device_buf(sizeof(YDataType) * xy_size); SimpleDeviceMem y_device_buf(sizeof(YDataType) * xy_size);
using DeviceOp = ck::tensor_operation::device::DeviceLayernorm<XDataType, using DeviceOp = ck::tensor_operation::device::DeviceNormalization<XDataType,
GammaDataType, GammaDataType,
BetaDataType, BetaDataType,
AccDataType, AccDataType,
YDataType, YDataType,
PassThrough, PassThrough,
Rank, Rank,
NumReduceDim>; NumReduceDim>;
// get device op instances // get device op instances
const auto op_ptrs = ck::tensor_operation::device::instance::DeviceOperationInstanceFactory< const auto op_ptrs = ck::tensor_operation::device::instance::DeviceOperationInstanceFactory<
......
client_example/07_conv2d_fwd/CMakeLists.txt 0 → 100644
View file @ d0b49a14
add_executable(client_conv2d_fwd conv2d_fwd.cpp)
target_link_libraries(client_conv2d_fwd PRIVATE composable_kernel::device_operations)
client_example/07_conv2d_fwd/conv2d_fwd.cpp 0 → 100644
View file @ d0b49a14
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
#include <iomanip>
#include <iostream>
#include <vector>
#include "ck/ck.hpp"
#include "ck/library/tensor_operation_instance/gpu/convolution_forward.hpp"
#include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
#include "ck/tensor_operation/gpu/device/device_conv_fwd.hpp"
#include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
using InDataType = ck::half_t;
using WeiDataType = ck::half_t;
using OutDataType = ck::half_t;
using InLayout = ck::tensor_layout::convolution::NHWC;
using WeiLayout = ck::tensor_layout::convolution::KYXC;
using OutLayout = ck::tensor_layout::convolution::NHWK;
using PassThrough = ck::tensor_operation::element_wise::PassThrough;
static constexpr ck::index_t NumDimSpatial = 2;
static constexpr ck::index_t N = 16;
static constexpr ck::index_t K = 32;
static constexpr ck::index_t C = 3;
static constexpr ck::index_t Y = 3;
static constexpr ck::index_t X = 3;
static constexpr ck::index_t Hi = 224;
static constexpr ck::index_t Wi = 224;
static constexpr ck::index_t Ho = 113;
static constexpr ck::index_t Wo = 113;
struct SimpleDeviceMem
{
SimpleDeviceMem() = delete;
SimpleDeviceMem(std::size_t mem_size) : p_mem_{}
{
(void)hipMalloc(static_cast<void**>(&p_mem_), mem_size);
}
void* GetDeviceBuffer() { return p_mem_; }
~SimpleDeviceMem() { (void)hipFree(p_mem_); }
void* p_mem_;
};
int main(int argc, char* argv[])
{
std::vector<ck::index_t> in_spatial_lengths{Hi, Wi};
std::vector<ck::index_t> filter_spatial_lengths{Y, X};
std::vector<ck::index_t> out_spatial_lengths{Ho, Wo};
std::vector<ck::index_t> filter_strides{2, 2};
std::vector<ck::index_t> filter_dilations{1, 1};
std::vector<ck::index_t> input_left_pads{2, 2};
std::vector<ck::index_t> input_right_pads{2, 2};
SimpleDeviceMem in(sizeof(InDataType) * N * Hi * Wi * C);
SimpleDeviceMem wei(sizeof(WeiDataType) * K * Y * X * C);
SimpleDeviceMem out(sizeof(OutDataType) * N * Ho * Wo * K);
using DeviceOp = ck::tensor_operation::device::DeviceConvFwd<NumDimSpatial,
InLayout,
WeiLayout,
OutLayout,
InDataType,
WeiDataType,
OutDataType,
PassThrough,
PassThrough,
PassThrough>;
// get device op instances
const auto op_ptrs = ck::tensor_operation::device::instance::DeviceOperationInstanceFactory<
DeviceOp>::GetInstances();
std::cout << "found " << op_ptrs.size() << " instances" << std::endl;
std::string best_op_name;
int best_op_id = -1;
float best_avg_time = std::numeric_limits<float>::max();
float best_gb_per_sec = 0;
float best_tflops = 0;
// profile device operation instances
std::cout << "Run all instances and do timing" << std::endl;
for(int i = 0; i < op_ptrs.size(); ++i)
{
auto& op_ptr = op_ptrs[i];
auto argument_ptr = op_ptr->MakeArgumentPointer(in.GetDeviceBuffer(),
wei.GetDeviceBuffer(),
out.GetDeviceBuffer(),
N,
K,
C,
in_spatial_lengths,
filter_spatial_lengths,
out_spatial_lengths,
filter_strides,
filter_dilations,
input_left_pads,
input_right_pads,
PassThrough{},
PassThrough{},
PassThrough{});
auto invoker_ptr = op_ptr->MakeInvokerPointer();
std::string op_name = op_ptr->GetTypeString();
if(op_ptr->IsSupportedArgument(argument_ptr.get()))
{
float avg_time = invoker_ptr->Run(argument_ptr.get(), StreamConfig{nullptr, true});
std::size_t flop = 2 * N * K * C * Ho * Wo * Y * X;
std::size_t num_bytes = sizeof(InDataType) * N * Hi * Wi * C +
sizeof(WeiDataType) * K * Y * X * C +
sizeof(OutDataType) * N * Ho * Wo * K;
float tflops = static_cast<float>(flop) / 1.E9 / avg_time;
float gb_per_sec = num_bytes / 1.E6 / avg_time;
std::cout << "Perf: " << std::setw(10) << avg_time << " ms, " << tflops << " TFlops, "
<< gb_per_sec << " GB/s, " << op_name << std::endl;
if(tflops > best_tflops)
{
best_op_id = i;
best_op_name = op_name;
best_avg_time = avg_time;
best_gb_per_sec = gb_per_sec;
best_tflops = tflops;
}
}
else
{
std::cout << op_name << " does not support this problem" << std::endl;
}
}
std::cout << "Best Perf: " << std::setw(10) << best_avg_time << " ms, " << best_tflops
<< " TFlops, " << best_gb_per_sec << " GB/s, " << best_op_name << std::endl;
// run the best intance
{
auto& op_ptr = op_ptrs[best_op_id];
std::cout << "Run the best instance without timing: " << op_ptr->GetTypeString()
<< std::endl;
auto argument_ptr = op_ptr->MakeArgumentPointer(in.GetDeviceBuffer(),
wei.GetDeviceBuffer(),
out.GetDeviceBuffer(),
N,
K,
C,
in_spatial_lengths,
filter_spatial_lengths,
out_spatial_lengths,
filter_strides,
filter_dilations,
input_left_pads,
input_right_pads,
PassThrough{},
PassThrough{},
PassThrough{});
auto invoker_ptr = op_ptr->MakeInvokerPointer();
if(op_ptr->IsSupportedArgument(argument_ptr.get()))
{
invoker_ptr->Run(argument_ptr.get(), StreamConfig{nullptr, false});
}
std::cout << "Done" << std::endl;
}
return 0;
}
\ No newline at end of file
client_example/08_fused_attention/CMakeLists.txt 0 → 100644
View file @ d0b49a14
add_executable(client_fused_attention fused_attention.cpp)
target_link_libraries(client_fused_attention PRIVATE composable_kernel::device_operations)
client_example/08_fused_attention/fused_attention.cpp 0 → 100644
View file @ d0b49a14
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
#include <iostream>
#include <vector>
#include "ck/ck.hpp"
#include "ck/library/tensor_operation_instance/gpu/batched_gemm_softmax_gemm_permute.hpp"
#include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
#include "ck/tensor_operation/gpu/device/device_batched_gemm_softmax_gemm_permute.hpp"
#include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
using AElementOp = ck::tensor_operation::element_wise::PassThrough;
using B0ElementOp = ck::tensor_operation::element_wise::PassThrough;
using Acc0ElementOp = ck::tensor_operation::element_wise::Scale;
using B1ElementOp = ck::tensor_operation::element_wise::PassThrough;
using CElementOp = ck::tensor_operation::element_wise::PassThrough;
constexpr static auto MaskingSpec =
ck::tensor_operation::device::MaskingSpecialization::MaskDisabled;
using ADataType = ck::half_t;
using B0DataType = ck::half_t;
using B1DataType = ck::half_t;
using CDataType = ck::half_t;
using AccDataType = float;
struct SimpleDeviceMem
{
SimpleDeviceMem() = delete;
SimpleDeviceMem(std::size_t mem_size) : p_mem_{}
{
(void)hipMalloc(static_cast<void**>(&p_mem_), mem_size);
}
void* GetDeviceBuffer() { return p_mem_; }
~SimpleDeviceMem() { (void)hipFree(p_mem_); }
void* p_mem_;
};
int main(int argc, char* argv[])
{
int G0 = 48;
int G1 = 16;
int M = 1024;
int N = 1024;
int K = 64;
int O = 64;
// A layout [G0, M, G1, K]
std::vector<ck::index_t> a_gs_ms_ks_lengths{G0, G1, M, K};
std::vector<ck::index_t> a_gs_ms_ks_strides{M * G1 * K, K, G1 * K, 1};
// B0 layout [G0, N, G1, K]
std::vector<ck::index_t> b0_gs_ns_ks_lengths{G0, G1, N, K};
std::vector<ck::index_t> b0_gs_ns_ks_strides{N * G1 * K, K, G1 * K, 1};
// B1 layout [G0, N, G1, O]
std::vector<ck::index_t> b1_gs_os_ns_lengths{G0, G1, O, N};
std::vector<ck::index_t> b1_gs_os_ns_strides{N * G1 * O, O, 1, G1 * O};
// C layout [G0, M, G1, O]
std::vector<ck::index_t> c_gs_ms_os_lengths{G0, G1, M, O};
std::vector<ck::index_t> c_gs_ms_os_strides{M * G1 * O, O, G1 * O, 1};
SimpleDeviceMem a_device_buf(sizeof(ADataType) * G0 * G1 * M * K);
SimpleDeviceMem b0_device_buf(sizeof(B0DataType) * G0 * G1 * N * K);
SimpleDeviceMem b1_device_buf(sizeof(B1DataType) * G0 * G1 * O * N);
SimpleDeviceMem c_device_buf(sizeof(CDataType) * G0 * G1 * M * O);
using DeviceOp =
ck::tensor_operation::device::DeviceBatchedGemmSoftmaxGemmPermute<2,
1,
1,
1,
1,
ADataType,
B0DataType,
B1DataType,
CDataType,
ck::Tuple<>,
ck::Tuple<>,
AElementOp,
B0ElementOp,
Acc0ElementOp,
B1ElementOp,
CElementOp,
MaskingSpec>;
// get device op instances
const auto op_ptrs = ck::tensor_operation::device::instance::DeviceOperationInstanceFactory<
DeviceOp>::GetInstances();
std::cout << "found " << op_ptrs.size() << " instances" << std::endl;
std::string best_op_name;
int best_op_id = -1;
float best_ave_time = 0;
float best_tflops = 0;
float best_gb_per_sec = 0;
// profile device op instances
std::cout << "Run all instances and do timing" << std::endl;
for(int i = 0; i < op_ptrs.size(); ++i)
{
auto& op_ptr = op_ptrs[i];
auto argument_ptr = op_ptr->MakeArgumentPointer(a_device_buf.GetDeviceBuffer(),
b0_device_buf.GetDeviceBuffer(),
b1_device_buf.GetDeviceBuffer(),
c_device_buf.GetDeviceBuffer(),
{}, // p_acc0_biases
{}, // p_acc1_biases
a_gs_ms_ks_lengths,
a_gs_ms_ks_strides,
b0_gs_ns_ks_lengths,
b0_gs_ns_ks_strides,
b1_gs_os_ns_lengths,
b1_gs_os_ns_strides,
c_gs_ms_os_lengths,
c_gs_ms_os_strides,
{}, // acc0_biases_gs_ms_ns_lengths
{}, // acc0_biases_gs_ms_ns_strides
{}, // acc1_biases_gs_ms_os_lengths
{}, // acc1_biases_gs_ms_os_strides
AElementOp{},
B0ElementOp{},
Acc0ElementOp{1 / sqrtf(K)},
B1ElementOp{},
CElementOp{});
auto invoker_ptr = op_ptr->MakeInvokerPointer();
std::string op_name = op_ptr->GetTypeString();
if(op_ptr->IsSupportedArgument(argument_ptr.get()))
{
float ave_time = invoker_ptr->Run(argument_ptr.get(), StreamConfig{nullptr, true});
std::size_t flop = (size_t(M) * N * K * 2 + size_t(M) * N * O * 2) * G0 * G1;
std::size_t num_btype = (sizeof(ADataType) * M * K + sizeof(B0DataType) * K * N +
sizeof(B1DataType) * N * O + sizeof(CDataType) * M * O) *
G0 * G1;
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, " << op_name << std::endl;
if(tflops > best_tflops)
{
best_op_id = i;
best_op_name = op_name;
best_tflops = tflops;
best_ave_time = ave_time;
best_gb_per_sec = gb_per_sec;
}
}
else
{
std::cout << op_name << " does not support this problem" << std::endl;
}
}
std::cout << "Best Perf: " << best_ave_time << " ms, " << best_tflops << " TFlops, "
<< best_gb_per_sec << " GB/s, " << best_op_name << std::endl;
// run the best instance
{
auto& op_ptr = op_ptrs[best_op_id];
std::cout << "Run the best instance without timing: " << op_ptr->GetTypeString()
<< std::endl;
auto argument_ptr = op_ptr->MakeArgumentPointer(a_device_buf.GetDeviceBuffer(),
b0_device_buf.GetDeviceBuffer(),
b1_device_buf.GetDeviceBuffer(),
c_device_buf.GetDeviceBuffer(),
{}, // p_acc0_biases
{}, // p_acc1_biases
a_gs_ms_ks_lengths,
a_gs_ms_ks_strides,
b0_gs_ns_ks_lengths,
b0_gs_ns_ks_strides,
b1_gs_os_ns_lengths,
b1_gs_os_ns_strides,
c_gs_ms_os_lengths,
c_gs_ms_os_strides,
{}, // acc0_biases_gs_ms_ns_lengths
{}, // acc0_biases_gs_ms_ns_strides
{}, // acc1_biases_gs_ms_os_lengths
{}, // acc1_biases_gs_ms_os_strides
AElementOp{},
B0ElementOp{},
Acc0ElementOp{1 / sqrtf(K)},
B1ElementOp{},
CElementOp{});
auto invoker_ptr = op_ptr->MakeInvokerPointer();
if(op_ptr->IsSupportedArgument(argument_ptr.get()))
{
invoker_ptr->Run(argument_ptr.get(), StreamConfig{nullptr, false});
}
std::cout << "Done" << std::endl;
}
return 0;
}
example/01_gemm/gemm_dl_fp16.cpp
View file @ d0b49a14
...@@ -3,7 +3,7 @@ ...@@ -3,7 +3,7 @@
#include "common.hpp" #include "common.hpp"
#include "ck/tensor_operation/gpu/device/device_gemm_dl.hpp" #include "ck/tensor_operation/gpu/device/impl/device_gemm_dl.hpp"
using ADataType = ck::half_t; using ADataType = ck::half_t;
using BDataType = ck::half_t; using BDataType = ck::half_t;
......
example/01_gemm/gemm_dl_fp32.cpp
View file @ d0b49a14
...@@ -3,7 +3,7 @@ ...@@ -3,7 +3,7 @@
#include "common.hpp" #include "common.hpp"
#include "ck/tensor_operation/gpu/device/device_gemm_dl.hpp" #include "ck/tensor_operation/gpu/device/impl/device_gemm_dl.hpp"
using ADataType = float; using ADataType = float;
using BDataType = float; using BDataType = float;
......
example/01_gemm/gemm_dl_int4.cpp
View file @ d0b49a14
...@@ -7,7 +7,7 @@ ...@@ -7,7 +7,7 @@
#include "common.hpp" #include "common.hpp"
#include "ck/tensor_operation/gpu/device/device_gemm_dl.hpp" #include "ck/tensor_operation/gpu/device/impl/device_gemm_dl.hpp"
using ADataType = ck::int4_t; using ADataType = ck::int4_t;
using BDataType = ck::int4_t; using BDataType = ck::int4_t;
......
example/01_gemm/gemm_dl_int8.cpp
View file @ d0b49a14
...@@ -3,7 +3,7 @@ ...@@ -3,7 +3,7 @@
#include "common.hpp" #include "common.hpp"
#include "ck/tensor_operation/gpu/device/device_gemm_dl.hpp" #include "ck/tensor_operation/gpu/device/impl/device_gemm_dl.hpp"
using ADataType = int8_t; using ADataType = int8_t;
using BDataType = int8_t; using BDataType = int8_t;
......
example/01_gemm/gemm_xdl_bf16.cpp
View file @ d0b49a14
...@@ -3,7 +3,7 @@ ...@@ -3,7 +3,7 @@
#include "common.hpp" #include "common.hpp"
#include "ck/tensor_operation/gpu/device/device_gemm_xdl_cshuffle.hpp" #include "ck/tensor_operation/gpu/device/impl/device_gemm_xdl_cshuffle.hpp"
using ADataType = ck::bhalf_t; using ADataType = ck::bhalf_t;
using BDataType = ck::bhalf_t; using BDataType = ck::bhalf_t;
......
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