Skip to content

GitLab

Commit ccaea50e authored by Jing Zhang's avatar Jing Zhang
Browse files

merge navi31_rel

parents 0b914465 10127959
Hide whitespace changes
Inline Side-by-side
Showing with 127 additions and 78 deletions
.github/CODEOWNERS
View file @ ccaea50e
* @zjing14 @asroy @junliume @illsilin @carlushuang @aosewski * @zjing14 @junliume @illsilin @carlushuang @aosewski
# Documentation files # Documentation files
docs/* @saadrahim @LisaDelaney docs/* @ROCm/rocm-documentation
*.md @saadrahim @LisaDelaney *.md @ROCm/rocm-documentation
*.rst @saadrahim @LisaDelaney *.rst @ROCm/rocm-documentation
# Header directory # Header directory for Doxygen documentation
library/include/* @saadrahim @LisaDelaney library/include/* @ROCm/rocm-documentation
Jenkinsfile
View file @ ccaea50e
def rocmnode(name) { def rocmnode(name) {
return '(rocmtest || miopen) && ' + name return '(rocmtest || miopen) && (' + name + ')'
} }
def show_node_info() { def show_node_info() {
...@@ -7,6 +7,7 @@ def show_node_info() { ...@@ -7,6 +7,7 @@ def show_node_info() {
echo "NODE_NAME = \$NODE_NAME" echo "NODE_NAME = \$NODE_NAME"
lsb_release -sd lsb_release -sd
uname -r uname -r
cat /sys/module/amdgpu/version
ls /opt/ -la ls /opt/ -la
""" """
} }
...@@ -33,6 +34,10 @@ def runShell(String command){ ...@@ -33,6 +34,10 @@ def runShell(String command){
def getDockerImageName(){ def getDockerImageName(){
def img def img
if (params.USE_CUSTOM_DOCKER != ""){
img = "${params.USE_CUSTOM_DOCKER}"
}
else{
if (params.ROCMVERSION != "6.0.1"){ if (params.ROCMVERSION != "6.0.1"){
if (params.COMPILER_VERSION == "") { if (params.COMPILER_VERSION == "") {
img = "${env.CK_DOCKERHUB}:ck_ub20.04_rocm${params.ROCMVERSION}" img = "${env.CK_DOCKERHUB}:ck_ub20.04_rocm${params.ROCMVERSION}"
...@@ -61,6 +66,7 @@ def getDockerImageName(){ ...@@ -61,6 +66,7 @@ def getDockerImageName(){
} }
} }
} }
}
return img return img
} }
...@@ -258,18 +264,24 @@ def cmake_build(Map conf=[:]){ ...@@ -258,18 +264,24 @@ def cmake_build(Map conf=[:]){
""") """)
sh cmd3 sh cmd3
} }
def setup_cmd = conf.get("setup_cmd", "${cmake_envs} cmake ${setup_args} .. ")
// reduce parallelism when compiling, clang uses too much memory // reduce parallelism when compiling, clang uses too much memory
def nt = nthreads() def nt = nthreads()
def build_cmd = conf.get("build_cmd", "${build_envs} dumb-init make -j${nt} ${config_targets}") def cmd
def execute_cmd = conf.get("execute_cmd", "") def execute_cmd = conf.get("execute_cmd", "")
if(!setup_args.contains("NO_CK_BUILD")){
def cmd = conf.get("cmd", """ def setup_cmd = conf.get("setup_cmd", "${cmake_envs} cmake ${setup_args} .. ")
def build_cmd = conf.get("build_cmd", "${build_envs} dumb-init make -j${nt} ${config_targets}")
cmd = conf.get("cmd", """
${setup_cmd} ${setup_cmd}
${build_cmd} ${build_cmd}
${execute_cmd} ${execute_cmd}
""") """)
}
else{
cmd = conf.get("cmd", """
${execute_cmd}
""")
}
echo cmd echo cmd
...@@ -365,8 +377,8 @@ def runCKProfiler(Map conf=[:]){ ...@@ -365,8 +377,8 @@ def runCKProfiler(Map conf=[:]){
(retimage, image) = getDockerImage(conf) (retimage, image) = getDockerImage(conf)
withDockerContainer(image: image, args: dockerOpts) { withDockerContainer(image: image, args: dockerOpts) {
timeout(time: 5, unit: 'MINUTES'){ timeout(time: 5, unit: 'MINUTES'){
sh 'PATH="/opt/rocm/opencl/bin:/opt/rocm/opencl/bin/x86_64:$PATH" clinfo | tee clinfo.log' sh 'rocminfo | tee rocminfo.log'
if ( runShell('grep -n "Number of devices:.*. 0" clinfo.log') ){ if ( !runShell('grep -n "gfx" rocminfo.log') ){
throw new Exception ("GPU not found") throw new Exception ("GPU not found")
} }
else{ else{
...@@ -379,20 +391,6 @@ def runCKProfiler(Map conf=[:]){ ...@@ -379,20 +391,6 @@ def runCKProfiler(Map conf=[:]){
echo "The job was cancelled or aborted" echo "The job was cancelled or aborted"
throw e 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: 24, unit: 'HOURS') timeout(time: 24, unit: 'HOURS')
...@@ -473,6 +471,7 @@ def Build_CK(Map conf=[:]){ ...@@ -473,6 +471,7 @@ def Build_CK(Map conf=[:]){
show_node_info() show_node_info()
env.HSA_ENABLE_SDMA=0 env.HSA_ENABLE_SDMA=0
env.DOCKER_BUILDKIT=1
checkout scm checkout scm
def image = getDockerImageName() def image = getDockerImageName()
...@@ -487,25 +486,35 @@ def Build_CK(Map conf=[:]){ ...@@ -487,25 +486,35 @@ def Build_CK(Map conf=[:]){
if (params.COMPILER_VERSION == "amd-staging" || params.COMPILER_VERSION == "amd-mainline-open" || params.COMPILER_COMMIT != ""){ if (params.COMPILER_VERSION == "amd-staging" || params.COMPILER_VERSION == "amd-mainline-open" || params.COMPILER_COMMIT != ""){
dockerOpts = dockerOpts + " --env HIP_CLANG_PATH='/llvm-project/build/bin' " dockerOpts = dockerOpts + " --env HIP_CLANG_PATH='/llvm-project/build/bin' "
} }
def video_id = sh(returnStdout: true, script: 'getent group video | cut -d: -f3')
def render_id = sh(returnStdout: true, script: 'getent group render | cut -d: -f3')
dockerOpts = dockerOpts + " --group-add=${video_id} --group-add=${render_id} "
echo "Docker flags: ${dockerOpts}"
def variant = env.STAGE_NAME def variant = env.STAGE_NAME
def retimage def retimage
def navi_node = 0 def navi_node = 0
def mi300_node = 0
gitStatusWrapper(credentialsId: "${status_wrapper_creds}", gitHubContext: "Jenkins - ${variant}", account: 'ROCm', repo: 'composable_kernel-internal') { gitStatusWrapper(credentialsId: "${status_wrapper_creds}", gitHubContext: "Jenkins - ${variant}", account: 'ROCm', repo: 'composable_kernel-internal') {
try { try {
(retimage, image) = getDockerImage(conf) (retimage, image) = getDockerImage(conf)
withDockerContainer(image: image, args: dockerOpts) { withDockerContainer(image: image, args: dockerOpts) {
timeout(time: 5, unit: 'MINUTES'){ timeout(time: 5, unit: 'MINUTES'){
sh 'PATH="/opt/rocm/opencl/bin:/opt/rocm/opencl/bin/x86_64:$PATH" clinfo | tee clinfo.log' sh 'rocminfo | tee rocminfo.log'
if ( runShell('grep -n "Number of devices:.*. 0" clinfo.log') ){ if ( !runShell('grep -n "gfx" rocminfo.log') ){
throw new Exception ("GPU not found") throw new Exception ("GPU not found")
} }
else{ else{
echo "GPU is OK" echo "GPU is OK"
} }
if ( runShell('grep -n "gfx1030" clinfo.log') || runShell('grep -n "gfx1101" clinfo.log') ){ if ( runShell('grep -n "gfx1030" rocminfo.log') || runShell('grep -n "gfx1101" rocminfo.log') ){
navi_node = 1 navi_node = 1
echo "This is a Navi node"
}
if ( runShell('grep -n "gfx942" rocminfo.log') ){
mi300_node = 1
echo "This is MI300 node"
} }
} }
} }
...@@ -514,23 +523,6 @@ def Build_CK(Map conf=[:]){ ...@@ -514,23 +523,6 @@ def Build_CK(Map conf=[:]){
echo "The job was cancelled or aborted" echo "The job was cancelled or aborted"
throw e 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"
}
if ( runShell('grep -n "gfx1030" clinfo.log') || runShell('grep -n "gfx1101" clinfo.log') ){
navi_node = 1
}
}
}
}
withDockerContainer(image: image, args: dockerOpts + ' -v=/var/jenkins/:/var/jenkins') { withDockerContainer(image: image, args: dockerOpts + ' -v=/var/jenkins/:/var/jenkins') {
timeout(time: 24, unit: 'HOURS') timeout(time: 24, unit: 'HOURS')
{ {
...@@ -544,8 +536,8 @@ def Build_CK(Map conf=[:]){ ...@@ -544,8 +536,8 @@ def Build_CK(Map conf=[:]){
sh 'tar -zcvf ckProfiler.tar.gz bin/ckProfiler' sh 'tar -zcvf ckProfiler.tar.gz bin/ckProfiler'
stash "ckProfiler.tar.gz" stash "ckProfiler.tar.gz"
} }
if (params.RUN_FULL_QA){ if (params.RUN_FULL_QA && mi300_node == 0 ){
// build deb packages // build deb packages for all MI100/200/300 targets and prepare to export
sh 'make -j package' sh 'make -j package'
archiveArtifacts artifacts: 'composablekernel-ckprofiler_*.deb' archiveArtifacts artifacts: 'composablekernel-ckprofiler_*.deb'
archiveArtifacts artifacts: 'composablekernel-tests_*.deb' archiveArtifacts artifacts: 'composablekernel-tests_*.deb'
...@@ -669,6 +661,10 @@ pipeline { ...@@ -669,6 +661,10 @@ pipeline {
name: "BUILD_DOCKER", name: "BUILD_DOCKER",
defaultValue: false, defaultValue: false,
description: "Force building docker image (default: false), set to true if docker image needs to be updated.") description: "Force building docker image (default: false), set to true if docker image needs to be updated.")
string(
name: 'USE_CUSTOM_DOCKER',
defaultValue: '',
description: 'If you want to use a custom docker image, please specify it here (default: leave blank).')
string( string(
name: 'ROCMVERSION', name: 'ROCMVERSION',
defaultValue: '6.0', defaultValue: '6.0',
...@@ -713,6 +709,10 @@ pipeline { ...@@ -713,6 +709,10 @@ pipeline {
name: "RUN_PERFORMANCE_TESTS", name: "RUN_PERFORMANCE_TESTS",
defaultValue: false, defaultValue: false,
description: "Run the performance tests (default: OFF)") description: "Run the performance tests (default: OFF)")
booleanParam(
name: "RUN_CODEGEN_TESTS",
defaultValue: true,
description: "Run the codegen tests (default: ON)")
} }
environment{ environment{
dbuser = "${dbuser}" dbuser = "${dbuser}"
...@@ -791,7 +791,34 @@ pipeline { ...@@ -791,7 +791,34 @@ pipeline {
} }
} }
} }
stage("Run Codegen Tests")
{
parallel
{
stage("Run Codegen Tests on MI100/MI200")
{
when {
beforeAgent true
expression { params.RUN_CODEGEN_TESTS.toBoolean() }
}
options { retry(2) }
agent{ label rocmnode("gfx908 || gfx90a")}
environment{
setup_args = "NO_CK_BUILD"
execute_args = """ cd ../codegen && rm -rf build && mkdir build && cd build && \
cmake -D CMAKE_PREFIX_PATH=/opt/rocm \
-D CMAKE_CXX_COMPILER=/opt/rocm/llvm/bin/clang++ \
-D CMAKE_BUILD_TYPE=Release \
-D GPU_TARGETS="gfx908;gfx90a" \
-DCMAKE_CXX_FLAGS=" -O3 " .. && make -j check"""
}
steps{
buildHipClangJobAndReboot(setup_args:setup_args, no_reboot:true, build_type: 'Release', execute_cmd: execute_args)
cleanWs()
}
}
}
}
stage("Build CK and run Tests") stage("Build CK and run Tests")
{ {
parallel parallel
...@@ -819,6 +846,26 @@ pipeline { ...@@ -819,6 +846,26 @@ pipeline {
cleanWs() cleanWs()
} }
} }
stage("Build CK and run Tests on MI300")
{
when {
beforeAgent true
expression { params.RUN_FULL_QA.toBoolean() }
}
agent{ label rocmnode("gfx942") }
environment{
setup_args = """ -DCMAKE_INSTALL_PREFIX=../install -DGPU_TARGETS="gfx942" -DCMAKE_CXX_FLAGS=" -O3 " """
execute_args = """ cd ../client_example && rm -rf build && mkdir build && cd build && \
cmake -DCMAKE_PREFIX_PATH="${env.WORKSPACE}/install;/opt/rocm" \
-DGPU_TARGETS="gfx942" \
-DCMAKE_CXX_COMPILER="${build_compiler()}" \
-DCMAKE_CXX_FLAGS=" -O3 " .. && make -j """
}
steps{
Build_CK_and_Reboot(setup_args: setup_args, config_targets: "install", no_reboot:true, build_type: 'Release', execute_cmd: execute_args, prefixpath: '/usr/local')
cleanWs()
}
}
stage("Build CK and run Tests on MI100/MI200") stage("Build CK and run Tests on MI100/MI200")
{ {
when { when {
......
client_example/01_gemm/gemm.cpp
View file @ ccaea50e
// SPDX-License-Identifier: MIT // SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved. // Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
#include <iomanip> #include <iomanip>
#include <vector> #include <vector>
...@@ -83,7 +83,7 @@ int main(int argc, char* argv[]) ...@@ -83,7 +83,7 @@ int main(int argc, char* argv[])
[](std::size_t nRow, std::size_t nCol, std::size_t stride, auto layout) { [](std::size_t nRow, std::size_t nCol, std::size_t stride, auto layout) {
using Layout = decltype(layout); using Layout = decltype(layout);
if constexpr(std::is_same<Layout, ck::tensor_layout::gemm::RowMajor>::value) if constexpr(std::is_same<Layout, Row>::value)
{ {
return (nRow - 1) * stride + nCol; return (nRow - 1) * stride + nCol;
} }
......
client_example/02_gemm_add_add_fastgelu/gemm_add_add_fastgelu.cpp
View file @ ccaea50e
// SPDX-License-Identifier: MIT // SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved. // Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
#include <iomanip> #include <iomanip>
#include <vector> #include <vector>
...@@ -92,7 +92,7 @@ int main(int argc, char* argv[]) ...@@ -92,7 +92,7 @@ int main(int argc, char* argv[])
[](std::size_t nRow, std::size_t nCol, std::size_t stride, auto layout) { [](std::size_t nRow, std::size_t nCol, std::size_t stride, auto layout) {
using Layout = decltype(layout); using Layout = decltype(layout);
if constexpr(std::is_same<Layout, ck::tensor_layout::gemm::RowMajor>::value) if constexpr(std::is_same<Layout, Row>::value)
{ {
return (nRow - 1) * stride + nCol; return (nRow - 1) * stride + nCol;
} }
......
client_example/02_gemm_add_add_fastgelu/gemm_add_add_fastgelu_generic.cpp
View file @ ccaea50e
// SPDX-License-Identifier: MIT // SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved. // Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
#include <iomanip> #include <iomanip>
#include <vector> #include <vector>
...@@ -93,7 +93,7 @@ int main(int argc, char* argv[]) ...@@ -93,7 +93,7 @@ int main(int argc, char* argv[])
[](std::size_t nRow, std::size_t nCol, std::size_t stride, auto layout) { [](std::size_t nRow, std::size_t nCol, std::size_t stride, auto layout) {
using Layout = decltype(layout); using Layout = decltype(layout);
if constexpr(std::is_same<Layout, ck::tensor_layout::gemm::RowMajor>::value) if constexpr(std::is_same<Layout, Row>::value)
{ {
return (nRow - 1) * stride + nCol; return (nRow - 1) * stride + nCol;
} }
......
client_example/02_gemm_add_add_fastgelu/gemm_add_fastgelu.cpp
View file @ ccaea50e
// SPDX-License-Identifier: MIT // SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved. // Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
#include <iomanip> #include <iomanip>
#include <vector> #include <vector>
...@@ -88,7 +88,7 @@ int main(int argc, char* argv[]) ...@@ -88,7 +88,7 @@ int main(int argc, char* argv[])
[](std::size_t nRow, std::size_t nCol, std::size_t stride, auto layout) { [](std::size_t nRow, std::size_t nCol, std::size_t stride, auto layout) {
using Layout = decltype(layout); using Layout = decltype(layout);
if constexpr(std::is_same<Layout, ck::tensor_layout::gemm::RowMajor>::value) if constexpr(std::is_same<Layout, Row>::value)
{ {
return (nRow - 1) * stride + nCol; return (nRow - 1) * stride + nCol;
} }
......
client_example/02_gemm_add_add_fastgelu/gemm_add_fastgelu_generic.cpp
View file @ ccaea50e
// SPDX-License-Identifier: MIT // SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved. // Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
#include <iomanip> #include <iomanip>
#include <vector> #include <vector>
...@@ -89,7 +89,7 @@ int main(int argc, char* argv[]) ...@@ -89,7 +89,7 @@ int main(int argc, char* argv[])
[](std::size_t nRow, std::size_t nCol, std::size_t stride, auto layout) { [](std::size_t nRow, std::size_t nCol, std::size_t stride, auto layout) {
using Layout = decltype(layout); using Layout = decltype(layout);
if constexpr(std::is_same<Layout, ck::tensor_layout::gemm::RowMajor>::value) if constexpr(std::is_same<Layout, Row>::value)
{ {
return (nRow - 1) * stride + nCol; return (nRow - 1) * stride + nCol;
} }
......
client_example/02_gemm_add_add_fastgelu/gemm_fastgelu.cpp
View file @ ccaea50e
// SPDX-License-Identifier: MIT // SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved. // Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
#include <iomanip> #include <iomanip>
#include <vector> #include <vector>
...@@ -84,7 +84,7 @@ int main(int argc, char* argv[]) ...@@ -84,7 +84,7 @@ int main(int argc, char* argv[])
[](std::size_t nRow, std::size_t nCol, std::size_t stride, auto layout) { [](std::size_t nRow, std::size_t nCol, std::size_t stride, auto layout) {
using Layout = decltype(layout); using Layout = decltype(layout);
if constexpr(std::is_same<Layout, ck::tensor_layout::gemm::RowMajor>::value) if constexpr(std::is_same<Layout, Row>::value)
{ {
return (nRow - 1) * stride + nCol; return (nRow - 1) * stride + nCol;
} }
......
client_example/02_gemm_add_add_fastgelu/gemm_fastgelu_generic.cpp
View file @ ccaea50e
// SPDX-License-Identifier: MIT // SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved. // Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
#include <iomanip> #include <iomanip>
#include <vector> #include <vector>
...@@ -85,7 +85,7 @@ int main(int argc, char* argv[]) ...@@ -85,7 +85,7 @@ int main(int argc, char* argv[])
[](std::size_t nRow, std::size_t nCol, std::size_t stride, auto layout) { [](std::size_t nRow, std::size_t nCol, std::size_t stride, auto layout) {
using Layout = decltype(layout); using Layout = decltype(layout);
if constexpr(std::is_same<Layout, ck::tensor_layout::gemm::RowMajor>::value) if constexpr(std::is_same<Layout, Row>::value)
{ {
return (nRow - 1) * stride + nCol; return (nRow - 1) * stride + nCol;
} }
......
client_example/03_gemm_layernorm/gemm_add_add_layernorm_naive.cpp
View file @ ccaea50e
// SPDX-License-Identifier: MIT // SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved. // Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
#include <iomanip> #include <iomanip>
#include <vector> #include <vector>
...@@ -17,6 +17,8 @@ ...@@ -17,6 +17,8 @@
using F16 = ck::half_t; using F16 = ck::half_t;
using F32 = float; using F32 = float;
using Row = ck::tensor_layout::gemm::RowMajor;
using ADataType = F16; using ADataType = F16;
using BDataType = F16; using BDataType = F16;
using BiasDataType = F32; using BiasDataType = F32;
...@@ -191,7 +193,7 @@ int main() ...@@ -191,7 +193,7 @@ int main()
[](std::size_t nRow, std::size_t nCol, std::size_t stride, auto layout) { [](std::size_t nRow, std::size_t nCol, std::size_t stride, auto layout) {
using Layout = decltype(layout); using Layout = decltype(layout);
if constexpr(std::is_same<Layout, ck::tensor_layout::gemm::RowMajor>::value) if constexpr(std::is_same<Layout, Row>::value)
{ {
return (nRow - 1) * stride + nCol; return (nRow - 1) * stride + nCol;
} }
......
client_example/03_gemm_layernorm/gemm_add_relu_add_layernorm_welford.cpp
View file @ ccaea50e
// SPDX-License-Identifier: MIT // SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved. // Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
#include <iomanip> #include <iomanip>
#include <iostream> #include <iostream>
...@@ -78,7 +78,7 @@ int main(int argc, char* argv[]) ...@@ -78,7 +78,7 @@ int main(int argc, char* argv[])
[](std::size_t nRow, std::size_t nCol, std::size_t stride, auto layout) { [](std::size_t nRow, std::size_t nCol, std::size_t stride, auto layout) {
using Layout = decltype(layout); using Layout = decltype(layout);
if constexpr(std::is_same<Layout, ck::tensor_layout::gemm::RowMajor>::value) if constexpr(std::is_same<Layout, Row>::value)
{ {
return (nRow - 1) * stride + nCol; return (nRow - 1) * stride + nCol;
} }
......
client_example/04_contraction/contraction_bilinear_fp32.cpp
View file @ ccaea50e
// SPDX-License-Identifier: MIT // SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved. // Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
#include <iomanip> #include <iomanip>
#include <numeric> #include <numeric>
......
client_example/04_contraction/contraction_bilinear_fp64.cpp
View file @ ccaea50e
// SPDX-License-Identifier: MIT // SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved. // Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
#include <iomanip> #include <iomanip>
#include <numeric> #include <numeric>
......
client_example/04_contraction/contraction_g1m2n3k1_add_xdl_fp16.cpp
View file @ ccaea50e
// SPDX-License-Identifier: MIT // SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved. // Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
#include <iomanip> #include <iomanip>
#include <numeric> #include <numeric>
......
client_example/04_contraction/contraction_scale_fp32.cpp
View file @ ccaea50e
// SPDX-License-Identifier: MIT // SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved. // Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
#include <iomanip> #include <iomanip>
#include <numeric> #include <numeric>
......
client_example/04_contraction/contraction_scale_fp64.cpp
View file @ ccaea50e
// SPDX-License-Identifier: MIT // SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved. // Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
#include <iomanip> #include <iomanip>
#include <numeric> #include <numeric>
......
client_example/05_layernorm/layernorm2d_bwd_data.cpp
View file @ ccaea50e
// SPDX-License-Identifier: MIT // SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved. // Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
#include <iomanip> #include <iomanip>
#include <vector> #include <vector>
......
client_example/05_layernorm/layernorm2d_bwd_gamma_beta.cpp
View file @ ccaea50e
// SPDX-License-Identifier: MIT // SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved. // Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
#include <iomanip> #include <iomanip>
#include <vector> #include <vector>
......
client_example/05_layernorm/layernorm2d_fwd.cpp
View file @ ccaea50e
// SPDX-License-Identifier: MIT // SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved. // Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
#include <iomanip> #include <iomanip>
#include <vector> #include <vector>
......
client_example/05_layernorm/layernorm4d_fwd.cpp
View file @ ccaea50e
// SPDX-License-Identifier: MIT // SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved. // Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
#include <iomanip> #include <iomanip>
#include <vector> #include <vector>
......
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