Compare commits

..

33 Commits

Author SHA1 Message Date
Sam Wu
f93799f9c9 Merge pull request #3828 from samjwu/tag
Update tags to 6.2.2 (#3827)
2024-09-27 13:12:19 -08:00
Sam Wu
a51979a2a5 Update tags to 6.2.2 (#3827) 2024-09-27 15:00:53 -06:00
Sam Wu
912fab238c Merge pull request #3823 from samjwu/rel62
Update 6.2 release branch
2024-09-27 12:03:39 -08:00
Peter Park
a34e7a381a Merge branch 'roc-6.2.x' into rel62 2024-09-27 15:21:11 -04:00
Sam Wu
e691307d60 Sync develop branch 2024-09-27 12:57:18 -06:00
Daniel Su
681a98a59d External CI: omnitrace tests (#3822) 2024-09-27 13:13:53 -04:00
Yanyao Wang
15d5040517 Merge pull request #3821 from ROCm/JeniferC99-patch-2
Update default.xml - Change 6.2.1 to 6.2.2
2024-09-27 11:15:08 -05:00
JeniferC99
41db131ada Update default.xml - Change 6.2.1 to 6.2.2 2024-09-27 08:44:36 -07:00
Sam Wu
17ad66f3d4 Update version list (#182)
* Update version list

* Update GA date in release notes

---------

Co-authored-by: Peter Park <peter.park@amd.com>
2024-09-27 10:39:53 -04:00
Daniel Su
5d1204ecdc External CI: Fix/add missing pipeline IDs (#3818) 2024-09-27 09:25:00 -04:00
Joseph Macaranas
c542ee1363 External CI: Build pipeline for rocprofiler-sdk (#3819)
* External CI: Pipeline for rocprofiler-sdk

* Add rocprofiler dependency

* External CI: rocprofiler-sdk build pipeline

---------

Co-authored-by: Daniel Su <danielsu@amd.com>
2024-09-26 16:42:53 -04:00
Daniel Su
ed4ffafec5 External CI: rpp tests (#3816) 2024-09-26 10:43:09 -04:00
Daniel Su
3fe8e17e56 External CI: test fixes w/ roctracer, list omniperf as partially succeeding (#3815) 2024-09-26 10:38:45 -04:00
Joseph Macaranas
fa31ee35e3 External CI: AMDMIGraphX Build Fix (#3814)
- Swap to default gcc on OS to resolve build errors from recent commits.
- Added libdnnl-dev dependency from iterative attempts with compiler change.
- Referred to the passing GitHub checks to observe the compilers that was used.
- Build CK jit lib and include in AMDMIGraphX build.
2024-09-26 10:20:10 -04:00
Joseph Macaranas
4b737a1cda External CI: omnitrace build pipeline (#3812)
* External CI: omnitrace build pipeline starter

- Adding initial set of dependencies and build flags.

* External CI: omnitrace build pipeline

- Add bison, rccl, texinfo dependencies based on build failures.
- Add AMDGPU_TARGETS flag
- Add ROCm binaries to PATH for clang-format and other tools used.

* Fix indentation

---------

Co-authored-by: Daniel Su <danielsu@amd.com>
2024-09-26 10:19:28 -04:00
Joseph Macaranas
ff0d863a0d External CI: torch and torchvision smoke tests (#3810)
* External CI: torch and torchvision smoke tests

- Fixed issues with package name and version for the vision wheel that prevented it from installing. A patch is used until my pull request in vision repo is merged.
- Referred to rocAutomation scripts to pick which test scripts to run out of the many in the torch and vision repo, and iteratively tested suggested scripts to see which ones completed in a timely manner.
- Leveraging pytest-azurepipelines module to automatically parse and graph results from these tests.
2024-09-25 15:25:36 -04:00
Daniel Su
b203a7cb74 External CI: default branches for hip-tests, omniperf (#3811) 2024-09-25 14:43:15 -04:00
Daniel Su
7d0e1aa06c External CI: enable ROCR-Runtime tests (#3809) 2024-09-25 14:19:45 -04:00
Peter Park
920efcd0b1 change release notes date to 9/25 (#181) 2024-09-25 10:05:41 -04:00
randyh62
ff26ae7c18 remove llvm-alt license (#3727)
* remove llvm-alt license

* fix linting error
2024-09-24 16:05:28 -07:00
Daniel Su
f9f92037cb External CI: fixes for rocMLIR and nightly build (#3800)
* External CI: fix symlinks for rocMLIR and nightly build

* add pipeline IDs for hip-tests

* fix hip-test ID typo
2024-09-24 16:39:48 -04:00
Yanyao Wang
4f9aadd6b6 Merge pull request #3802 from WBobby/roc-6.2.1
Adding ROCm6.2.1 manifest file and update the build script of rocAL
2024-09-24 13:35:41 -05:00
Peter Park
3ab3782a7f Update compat matrix for ROCm 6.2.2 (#179)
* update current matrix for 6.2.2

* update history compat

* fix typo

* fixed missed 60201s

* fix missed rocm-6.2.1

* Add additional column to compatibility-matrix-historical-6.0, so it includes it correctly

Also, fixing a few 6.2.2 footnote references

* add oracle linux 8.9 under 6.2.2 in historical

* rm widths in historical table

* lowercase a letter

* Fix version numbers for 6.2.2

* Minor updates to historical matrix

* add ubuntu 24.04.1

* Docs: Add Ubuntu 24.04.1 (#3801)

* add ubuntu 24.04.1

* add 24.04.1 to bottom os section

* fix heading and template

* Update compatibility-matrix.rst for OpenMP version

* Update compatibility-matrix-historical-6.0.csv for OpenMP version

* rm ubuntu 24.04.1 from 6.2.0

* Update docs/compatibility/compatibility-matrix.rst

Co-authored-by: Young Hui - AMD <145490163+yhuiYH@users.noreply.github.com>

* rm duplicate ubuntu in historical

---------

Co-authored-by: Young Hui - AMD <145490163+yhuiYH@users.noreply.github.com>

* add overwritten ubuntu 24.04.1

* fix wrong versions and extra comma

---------

Co-authored-by: Young Hui - AMD <145490163+yhuiYH@users.noreply.github.com>
2024-09-24 14:22:00 -04:00
Peter Park
516b6c8ebc Add ROCm 6.2.2 release notes (#178)
* rm extra file

* sync release.md w/ public version

* update versions to 6.2.2

* update version, release, and ga date

* Revert "update versions to 6.2.2"

This reverts commit c08e51d1acc8e773deefef96946c31fd368a09dd.

* stack 6.2.2 and 6.2.1

* fix word

* bump 6.2.1 headings

* add explanatory note

remove 'please'

rm caps

* add fixed issue highlight

* update mode-2 fix

* add clarification

* add ubuntu 24.04.1 note

* update autotag templates for 6.2.x

* rm Ubuntu 24.04.1 from 6.2.2 (have it in 6.2.1 only)

* add horizontal rule for visual separation between 6.2.2 and 6.2.1 release notes

* remove extra templ

* spellcheck

* Docs: Add Ubuntu 24.04.1 (#3801)

* add ubuntu 24.04.1

* add 24.04.1 to bottom os section

* fix heading and template

* Update compatibility-matrix.rst for OpenMP version

* Update compatibility-matrix-historical-6.0.csv for OpenMP version

* rm ubuntu 24.04.1 from 6.2.0

* Update docs/compatibility/compatibility-matrix.rst

Co-authored-by: Young Hui - AMD <145490163+yhuiYH@users.noreply.github.com>

* rm duplicate ubuntu in historical

---------

Co-authored-by: Young Hui - AMD <145490163+yhuiYH@users.noreply.github.com>

---------

Co-authored-by: Young Hui - AMD <145490163+yhuiYH@users.noreply.github.com>
2024-09-24 14:07:40 -04:00
Sam Wu
508f3e1afb Merge pull request #3803 from peterjunpark/roc-6.2.x
roc-6.2.x: Add Ubuntu 24.04.1 (#3801)
2024-09-24 10:02:44 -08:00
Peter Park
8d6d82d65f Docs: Add Ubuntu 24.04.1 (#3801)
* add ubuntu 24.04.1

* add 24.04.1 to bottom os section

* fix heading and template

* Update compatibility-matrix.rst for OpenMP version

* Update compatibility-matrix-historical-6.0.csv for OpenMP version

* rm ubuntu 24.04.1 from 6.2.0

* Update docs/compatibility/compatibility-matrix.rst

Co-authored-by: Young Hui - AMD <145490163+yhuiYH@users.noreply.github.com>

* rm duplicate ubuntu in historical

---------

Co-authored-by: Young Hui - AMD <145490163+yhuiYH@users.noreply.github.com>
2024-09-24 13:33:48 -04:00
Peter Park
c5d0fb8477 Docs: Add Ubuntu 24.04.1 (#3801)
* add ubuntu 24.04.1

* add 24.04.1 to bottom os section

* fix heading and template

* Update compatibility-matrix.rst for OpenMP version

* Update compatibility-matrix-historical-6.0.csv for OpenMP version

* rm ubuntu 24.04.1 from 6.2.0

* Update docs/compatibility/compatibility-matrix.rst

Co-authored-by: Young Hui - AMD <145490163+yhuiYH@users.noreply.github.com>

* rm duplicate ubuntu in historical

---------

Co-authored-by: Young Hui - AMD <145490163+yhuiYH@users.noreply.github.com>
2024-09-24 13:32:08 -04:00
pbhandar-amd
3dc6d89297 Update versions.md (#180) 2024-09-24 11:38:06 -04:00
Daniel Su
7c0c197c91 External CI: fix hip-tests symlink creation (#3799) 2024-09-24 10:32:05 -04:00
Wang, Yanyao
055511ed33 Add ROCm6.2.1 manifest file 2024-09-23 21:17:44 -05:00
Wang, Yanyao
210be113e4 fix build rocal for ROCm6.2.1 2024-09-23 21:14:57 -05:00
Joseph Macaranas
7963d27853 External CI: Publish wheel as artifact for rocPyDecode (#3796) 2024-09-23 17:52:24 -04:00
Sam Wu
a493148cbc Merge pull request #3777 from samjwu/rcclki
Cherry pick to 6.2 release branch: Update RCCL known issue wording (#3775)
2024-09-20 16:08:12 -08:00
55 changed files with 1349 additions and 730 deletions

View File

@@ -8,20 +8,21 @@ parameters:
- name: aptPackages
type: object
default:
- git
- cmake
- ninja-build
- git
- libdnnl-dev
- libdrm-dev
- libnuma-dev
- python3-pip
- python3-venv
- libtbb-dev
- nlohmann-json3-dev
- libmsgpack-dev
- libsqlite3-dev
- libnuma-dev
- libprotobuf-dev
- libsqlite3-dev
- libtbb-dev
- ninja-build
- nlohmann-json3-dev
- protobuf-compiler
- python3-pip
- python3-pybind11
- python3-venv
- name: pipModules
type: object
default:
@@ -95,14 +96,12 @@ jobs:
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/build-cmake.yml
parameters:
extraBuildFlags: >-
-DCMAKE_CXX_COMPILER=$(Agent.BuildDirectory)/rocm/llvm/bin/clang++
-DCMAKE_C_COMPILER=$(Agent.BuildDirectory)/rocm/llvm/bin/clang
-DCMAKE_BUILD_TYPE=Release
-DGPU_TARGETS=$(JOB_GPU_TARGET)
-DAMDGPU_TARGETS=$(JOB_GPU_TARGET)
-DCMAKE_MODULE_PATH=$(Agent.BuildDirectory)/rocm/lib/cmake/hip
-DCMAKE_PREFIX_PATH=$(Agent.BuildDirectory)/rocm/llvm;$(Agent.BuildDirectory)/rocm
-DHALF_INCLUDE_DIR=$(Agent.BuildDirectory)/rocm/include
-DMIGRAPHX_USE_COMPOSABLEKERNEL=OFF
-DBUILD_TESTING=ON
-GNinja
# REFERENCE: https://github.com/ROCm/composable_kernel/issues/782

View File

@@ -8,15 +8,21 @@ parameters:
- name: aptPackages
type: object
default:
- libelf-dev
- g++
- libdrm-dev
- libelf-dev
- libnuma-dev
- name: rocmDependencies
type: object
default:
- llvm-project
- rocprofiler-register
- name: rocmTestDependencies
type: object
default:
- llvm-project
- rocm_smi_lib
- rocprofiler-register
jobs:
- job: ROCR_Runtime
@@ -50,3 +56,120 @@ jobs:
-DCMAKE_PREFIX_PATH=$(Agent.BuildDirectory)/rocm
-DBUILD_SHARED_LIBS=ON
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/artifact-upload.yml
- job: ROCR_Runtime_testing
dependsOn: ROCR_Runtime
condition: succeeded()
variables:
- group: common
- template: /.azuredevops/variables-global.yml
pool: $(JOB_TEST_POOL)
workspace:
clean: all
strategy:
matrix:
gfx942:
JOB_GPU_TARGET: gfx942
JOB_TEST_POOL: ${{ variables.GFX942_TEST_POOL }}
steps:
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/dependencies-other.yml
parameters:
aptPackages: ${{ parameters.aptPackages }}
- task: Bash@3
displayName: Install libhwloc5
inputs:
targetType: 'inline'
script: |
wget http://ftp.us.debian.org/debian/pool/main/h/hwloc/libhwloc5_1.11.12-3_amd64.deb
wget http://ftp.us.debian.org/debian/pool/main/h/hwloc/libhwloc-dev_1.11.12-3_amd64.deb
sudo apt install -y --allow-downgrades ./libhwloc5_1.11.12-3_amd64.deb ./libhwloc-dev_1.11.12-3_amd64.deb
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/preamble.yml
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/local-artifact-download.yml
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/dependencies-aqlprofile.yml
parameters:
${{ if eq(parameters.checkoutRef, '') }}:
dependencySource: staging
${{ elseif ne(parameters.checkoutRef, '') }}:
dependencySource: tag-builds
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/dependencies-rocm.yml
parameters:
dependencyList: ${{ parameters.rocmTestDependencies }}
gpuTarget: $(JOB_GPU_TARGET)
${{ if eq(parameters.checkoutRef, '') }}:
dependencySource: staging
${{ elseif ne(parameters.checkoutRef, '') }}:
dependencySource: tag-builds
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/checkout.yml
parameters:
checkoutRepo: ${{ parameters.checkoutRepo }}
- task: Bash@3
displayName: Build kfdtest
inputs:
targetType: 'inline'
workingDirectory: $(Build.SourcesDirectory)/libhsakmt/tests/kfdtest
script: |
mkdir build && cd build
cmake -DCMAKE_PREFIX_PATH=$(Agent.BuildDirectory)/rocm ..
make
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/test.yml
parameters:
componentName: kfdtest
testExecutable: BIN_DIR=$(Build.SourcesDirectory)/libhsakmt/tests/kfdtest/build ./run_kfdtest.sh
testParameters: '-p core --gtest_output=xml:./test_output.xml --gtest_color=yes'
testDir: $(Build.SourcesDirectory)/libhsakmt/tests/kfdtest/scripts
- task: Bash@3
displayName: Build rdmatest app
inputs:
targetType: 'inline'
workingDirectory: $(Build.SourcesDirectory)/libhsakmt/tests/rdma/simple/app
script: |
cmake -DCMAKE_PREFIX_PATH=$(Agent.BuildDirectory)/rocm -DLIBHSAKMT_PATH=$(Agent.BuildDirectory)/rocm -DDRM_AMDGPU_INCLUDE_DIRS=$(Agent.BuildDirectory)/rocm/include .
cmake --build .
- task: Bash@3
displayName: Build rdmatest driver
inputs:
targetType: 'inline'
workingDirectory: $(Build.SourcesDirectory)/libhsakmt/tests/rdma/simple/drv
script: |
sed -i 's/HSAKMT_PAGE_SHIFT/PAGE_SHIFT/g' amdp2ptest.c
sed -i 's/"MIT"/"GPL"/' amdp2ptest.c
RDMA_HEADER_DIR=/usr/src/amdgpu-*/include make all
- task: Bash@3
displayName: Install rdmatest driver
inputs:
targetType: 'inline'
workingDirectory: $(Build.SourcesDirectory)/libhsakmt/tests/rdma/simple/drv
script: |
sudo rmmod amdp2ptest.ko
sudo insmod amdp2ptest.ko
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/test.yml
parameters:
componentName: rdmatest
testExecutable: yes | ./rdma_test
testParameters: ''
testDir: $(Build.SourcesDirectory)/libhsakmt/tests/rdma/simple/app
testPublishResults: false
- task: Bash@3
displayName: Build rocrtst
inputs:
targetType: 'inline'
workingDirectory: $(Build.SourcesDirectory)/rocrtst/suites/test_common
script: |
sudo rm -rf $(Agent.BuildDirectory)/external/llvm-project
mkdir -p $(Agent.BuildDirectory)/external/llvm-project/clang/lib
sudo ln -s $(Agent.BuildDirectory)/rocm/llvm/lib/clang/20/include $(Agent.BuildDirectory)/external/llvm-project/clang/lib/Headers
mkdir build && cd build
cmake .. \
-DCMAKE_PREFIX_PATH=$(Agent.BuildDirectory)/rocm \
-DTARGET_DEVICES=$(JOB_GPU_TARGET) \
-DROCM_DIR=$(Agent.BuildDirectory)/rocm \
-DLLVM_DIR=$(Agent.BuildDirectory)/rocm/llvm/bin \
-DOPENCL_DIR=$(Agent.BuildDirectory)/rocm/llvm/bin
make
make rocrtst_kernels
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/test.yml
parameters:
componentName: rocrtst
testExecutable: ./rocrtst64
testParameters: '--gtest_filter="-rocrtstNeg.Memory_Negative_Tests" --gtest_output=xml:./test_output.xml --gtest_color=yes'
testDir: $(Build.SourcesDirectory)/rocrtst/suites/test_common/build/$(JOB_GPU_TARGET)

View File

@@ -43,6 +43,7 @@ parameters:
- rocprofiler-register
- ROCR-Runtime
- rocRAND
- roctracer
jobs:
- job: ROCmValidationSuite

View File

@@ -90,6 +90,8 @@ jobs:
-DCMAKE_C_COMPILER_LAUNCHER=ccache
-DCMAKE_HIP_FLAGS="-Wno-missing-include-dirs"
-DCMAKE_PREFIX_PATH=$(Agent.BuildDirectory)/rocm
-DCK_BUILD_JIT_LIB=ON
-DCMAKE_POSITION_INDEPENDENT_CODE=ON
-DCMAKE_BUILD_TYPE=Release
-DGPU_TARGETS=$(JOB_GPU_TARGET)
-GNinja

View File

@@ -89,7 +89,7 @@ jobs:
gpuTarget: $(JOB_GPU_TARGET)
- job: hip_tests_testing
timeoutInMinutes: 180
timeoutInMinutes: 240
dependsOn: hip_tests
condition: succeeded()
variables:
@@ -130,6 +130,7 @@ jobs:
inputs:
targetType: inline
script: |
sudo rm -rf /opt/rocm
sudo mkdir -p /opt/rocm/bin
sudo ln -s $(Agent.BuildDirectory)/rocm/bin/rocm_agent_enumerator /opt/rocm/bin/rocm_agent_enumerator
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/test.yml

View File

@@ -37,6 +37,7 @@ parameters:
- rocprofiler-register
- ROCR-Runtime
- rocSPARSE
- roctracer
jobs:
- job: hipSPARSE

View File

@@ -0,0 +1,141 @@
# largely referenced from: https://github.com/ROCm/omnitrace/blob/main/.github/workflows/ubuntu-jammy.yml
parameters:
- name: checkoutRepo
type: string
default: 'self'
- name: checkoutRef
type: string
default: ''
- name: aptPackages
type: object
default:
- autoconf
- autotools-dev
- bison
- build-essential
- bzip2
- clang
- cmake
- environment-modules
- g++-12
- libdrm-dev
- libfabric-dev
- libiberty-dev
- libpapi-dev
- libpfm4-dev
- libtool
- libopenmpi-dev
- m4
- openmpi-bin
- software-properties-common
- python3-pip
- texinfo
- zlib1g-dev
- name: pipModules
type: object
default:
- numpy
- perfetto
- dataclasses
- name: rocmDependencies
type: object
default:
- clr
- llvm-project
- rccl
- rocm-core
- rocm_smi_lib
- rocminfo
- ROCR-Runtime
- rocprofiler
- rocprofiler-register
- roctracer
jobs:
- job: omnitrace
variables:
- group: common
- template: /.azuredevops/variables-global.yml
pool: ${{ variables.MEDIUM_BUILD_POOL }}
workspace:
clean: all
strategy:
matrix:
gfx942:
JOB_GPU_TARGET: gfx942
gfx90a:
JOB_GPU_TARGET: gfx90a
steps:
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/dependencies-other.yml
parameters:
aptPackages: ${{ parameters.aptPackages }}
pipModules: ${{ parameters.pipModules }}
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/preamble.yml
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/checkout.yml
parameters:
checkoutRepo: ${{ parameters.checkoutRepo }}
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/dependencies-rocm.yml
parameters:
dependencyList: ${{ parameters.rocmDependencies }}
gpuTarget: $(JOB_GPU_TARGET)
# CI case: download latest default branch build
${{ if eq(parameters.checkoutRef, '') }}:
dependencySource: staging
# manual build case: triggered by ROCm/ROCm repo
${{ elseif ne(parameters.checkoutRef, '') }}:
dependencySource: tag-builds
- task: Bash@3
displayName: ROCm symbolic link
inputs:
targetType: inline
script: |
sudo rm -rf /opt/rocm
sudo ln -s $(Agent.BuildDirectory)/rocm /opt/rocm
- task: Bash@3
displayName: Add ROCm binaries to PATH
inputs:
targetType: inline
script: echo "##vso[task.prependpath]$(Agent.BuildDirectory)/rocm/bin"
- task: Bash@3
displayName: Add ROCm compilers to PATH
inputs:
targetType: inline
script: echo "##vso[task.prependpath]$(Agent.BuildDirectory)/rocm/llvm/bin"
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/build-cmake.yml
parameters:
# build flags reference: https://rocm.docs.amd.com/projects/omnitrace/en/latest/install/install.html
extraBuildFlags: >-
-DCMAKE_PREFIX_PATH=$(Agent.BuildDirectory)/rocm
-DOMNITRACE_BUILD_TESTING=ON
-DOMNITRACE_BUILD_DYNINST=ON
-DOMNITRACE_BUILD_LIBUNWIND=ON
-DDYNINST_BUILD_TBB=ON
-DDYNINST_BUILD_ELFUTILS=ON
-DDYNINST_BUILD_LIBIBERTY=ON
-DDYNINST_BUILD_BOOST=ON
-DOMNITRACE_USE_PAPI=ON
-DOMNITRACE_USE_MPI=ON
-DAMDGPU_TARGETS=$(JOB_GPU_TARGET)
multithreadFlag: -- -j32
- task: Bash@3
displayName: Set up omnitrace env
inputs:
targetType: inline
script: source share/omnitrace/setup-env.sh
workingDirectory: build
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/test.yml
parameters:
componentName: omnitrace
- task: Bash@3
displayName: Remove ROCm binaries from PATH
inputs:
targetType: inline
script: echo "##vso[task.setvariable variable=PATH]$(echo $PATH | sed -e 's;:$(Agent.BuildDirectory)/rocm/bin;;' -e 's;^/;;' -e 's;/$;;')"
- task: Bash@3
displayName: Remove ROCm compilers from PATH
inputs:
targetType: inline
script: echo "##vso[task.setvariable variable=PATH]$(echo $PATH | sed -e 's;:$(Agent.BuildDirectory)/rocm/llvm/bin;;' -e 's;^/;;' -e 's;/$;;')"
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/artifact-upload.yml
parameters:
gpuTarget: $(JOB_GPU_TARGET)

View File

@@ -43,6 +43,7 @@ parameters:
- ROCR-Runtime
- rocRAND
- rocSPARSE
- roctracer
jobs:
- job: rocALUTION

View File

@@ -52,6 +52,7 @@ parameters:
- rocprofiler-register
- rocm_smi_lib
- ROCR-Runtime
- roctracer
jobs:
- job: rocBLAS

View File

@@ -105,7 +105,9 @@ jobs:
displayName: ROCm symbolic link
inputs:
targetType: inline
script: sudo ln -s $(Agent.BuildDirectory)/rocm /opt/rocm
script: |
sudo rm -rf /opt/rocm
sudo ln -s $(Agent.BuildDirectory)/rocm /opt/rocm
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/build-cmake.yml
parameters:
installEnabled: false

View File

@@ -73,6 +73,13 @@ jobs:
# manual build case: triggered by ROCm/ROCm repo
${{ elseif ne(parameters.checkoutRef, '') }}:
dependencySource: tag-builds
- task: Bash@3
displayName: 'ROCm symbolic link'
inputs:
targetType: inline
script: |
sudo rm -rf /opt/rocm
sudo ln -s $(Agent.BuildDirectory)/rocm /opt/rocm
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/build-cmake.yml
parameters:
extraBuildFlags: >-
@@ -85,6 +92,24 @@ jobs:
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/artifact-upload.yml
parameters:
gpuTarget: $(JOB_GPU_TARGET)
publish: false
- task: Bash@3
displayName: Create wheel file
inputs:
targetType: inline
script: python setup.py bdist_wheel
workingDirectory: $(Build.SourcesDirectory)
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/artifact-prepare-package.yml
parameters:
sourceDir: $(Build.SourcesDirectory)/dist
contentsString: '*.whl'
targetDir: $(Build.ArtifactStagingDirectory)
clean: false
- task: PublishPipelineArtifact@1
displayName: 'wheel file Publish'
retryCountOnTaskFailure: 3
inputs:
targetPath: $(Build.ArtifactStagingDirectory)
- job: rocPyDecode_testing
dependsOn: rocPyDecode

View File

@@ -40,6 +40,7 @@ parameters:
- rocminfo
- rocprofiler-register
- ROCR-Runtime
- roctracer
jobs:
- job: rocSPARSE

View File

@@ -0,0 +1,95 @@
parameters:
- name: checkoutRepo
type: string
default: 'self'
- name: checkoutRef
type: string
default: ''
- name: aptPackages
type: object
default:
- cmake
- python3-pip
- libdrm-dev
- libdw-dev
- libelf-dev
- name: pipModules
type: object
default:
- black
- clang-format
- clang-tidy
- cmake
- cmake-format
- dataclasses
- numpy
- otf2
- pandas
- perfetto
- pycobertura
- pytest
- pyyaml
- name: rocmDependencies
type: object
default:
- clr
- llvm-project
- rccl
- rocm-cmake
- rocm-core
- rocminfo
- ROCR-Runtime
- rocprofiler-register
- roctracer
jobs:
- job: rocprofilersdk
variables:
- group: common
- template: /.azuredevops/variables-global.yml
pool:
vmImage: ${{ variables.BASE_BUILD_POOL }}
workspace:
clean: all
strategy:
matrix:
gfx942:
JOB_GPU_TARGET: gfx942
gfx90a:
JOB_GPU_TARGET: gfx90a
steps:
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/dependencies-other.yml
parameters:
aptPackages: ${{ parameters.aptPackages }}
pipModules: ${{ parameters.pipModules }}
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/preamble.yml
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/checkout.yml
parameters:
checkoutRepo: ${{ parameters.checkoutRepo }}
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/dependencies-aqlprofile.yml
parameters:
${{ if eq(parameters.checkoutRef, '') }}:
dependencySource: staging
${{ elseif ne(parameters.checkoutRef, '') }}:
dependencySource: tag-builds
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/dependencies-rocm.yml
parameters:
dependencyList: ${{ parameters.rocmDependencies }}
gpuTarget: $(JOB_GPU_TARGET)
# CI case: download latest default branch build
${{ if eq(parameters.checkoutRef, '') }}:
dependencySource: staging
# manual build case: triggered by ROCm/ROCm repo
${{ elseif ne(parameters.checkoutRef, '') }}:
dependencySource: tag-builds
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/build-cmake.yml
parameters:
extraBuildFlags: >-
-DCMAKE_PREFIX_PATH=$(Agent.BuildDirectory)/rocm
-DROCPROFILER_BUILD_TESTS=ON
-DROCPROFILER_BUILD_SAMPLES=OFF
-DAMDGPU_TARGETS=$(JOB_GPU_TARGET)
multithreadFlag: -- -j2
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/artifact-upload.yml
parameters:
gpuTarget: $(JOB_GPU_TARGET)

View File

@@ -9,8 +9,18 @@ parameters:
type: object
default:
- cmake
- libopencv-dev
- libsndfile1-dev
- libstdc++-12-dev
- imagemagick
- ninja-build
- clang
- name: pipModules
type: object
default:
- openpyxl
- pandas
- sphinx
- name: rocmDependencies
type: object
default:
@@ -20,6 +30,18 @@ parameters:
- llvm-project
- rocminfo
- ROCR-Runtime
- name: rocmTestDependencies
type: object
default:
- aomp
- clr
- half
- hipTensor
- llvm-project
- rocm-cmake
- rocminfo
- rocprofiler-register
- ROCR-Runtime
jobs:
- job: rpp
@@ -68,3 +90,92 @@ jobs:
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/artifact-upload.yml
parameters:
gpuTarget: $(JOB_GPU_TARGET)
- job: rpp_testing
dependsOn: rpp
condition: succeeded()
variables:
- group: common
- template: /.azuredevops/variables-global.yml
- name: LD_LIBRARY_PATH
value: $(Agent.BuildDirectory)/rocm/lib;$(Agent.BuildDirectory)/rocm/llvm/lib
pool: $(JOB_TEST_POOL)
workspace:
clean: all
strategy:
matrix:
gfx942:
JOB_GPU_TARGET: gfx942
JOB_TEST_POOL: ${{ variables.GFX942_TEST_POOL }}
steps:
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/dependencies-other.yml
parameters:
aptPackages: ${{ parameters.aptPackages }}
pipModules: ${{ parameters.pipModules }}
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/preamble.yml
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/local-artifact-download.yml
parameters:
gpuTarget: $(JOB_GPU_TARGET)
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/dependencies-aqlprofile.yml
parameters:
${{ if eq(parameters.checkoutRef, '') }}:
dependencySource: staging
${{ elseif ne(parameters.checkoutRef, '') }}:
dependencySource: tag-builds
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/dependencies-rocm.yml
parameters:
dependencyList: ${{ parameters.rocmTestDependencies }}
gpuTarget: $(JOB_GPU_TARGET)
${{ if eq(parameters.checkoutRef, '') }}:
dependencySource: staging
${{ elseif ne(parameters.checkoutRef, '') }}:
dependencySource: tag-builds
# Dependencies from: https://github.com/ROCm/rpp/blob/develop/utilities/test_suite/README.md
- task: Bash@3
displayName: Build and install Turbo JPEG
inputs:
targetType: 'inline'
script: |
sudo apt-get install nasm
sudo apt-get install wget
git clone -b 3.0.2 https://github.com/libjpeg-turbo/libjpeg-turbo.git
cd libjpeg-turbo
mkdir build
cd build
cmake -DCMAKE_INSTALL_PREFIX=/usr \
-DCMAKE_BUILD_TYPE=RELEASE \
-DENABLE_STATIC=FALSE \
-DCMAKE_INSTALL_DEFAULT_LIBDIR=lib \
-DWITH_JPEG8=TRUE \
..
make -j$nproc
sudo make install
- task: Bash@3
displayName: Build and install Nifti
inputs:
targetType: 'inline'
script: |
git clone https://github.com/NIFTI-Imaging/nifti_clib.git
cd nifti_clib
mkdir build
cd build
cmake ..
sudo make -j$nproc install
- task: Bash@3
displayName: Build rpp tests
inputs:
targetType: 'inline'
script: |
sudo rm -rf /opt/rocm
sudo ln -s $(Agent.BuildDirectory)/rocm /opt/rocm
mkdir rpp-tests
cd rpp-tests
cmake /opt/rocm/share/rpp/test \
-DCMAKE_CXX_COMPILER=$(Agent.BuildDirectory)/rocm/llvm/bin/amdclang++ \
-DCMAKE_C_COMPILER=$(Agent.BuildDirectory)/rocm/llvm/bin/amdclang
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/test.yml
parameters:
componentName: rpp
testExecutable: 'export PATH=$(Agent.BuildDirectory)/rocm/llvm/bin:$PATH; CC=$(Agent.BuildDirectory)/rocm/llvm/bin/amdclang CMAKE_VERBOSE_MAKEFILE=ON VERBOSE=1 ctest'
testDir: 'rpp-tests'
- script: sudo rm /opt/rocm

View File

@@ -62,6 +62,8 @@ parameters:
- ffmpeg
- libopenblas-dev
- liblapack-dev
- libswscale-dev
- libavformat-dev
- name: pipModules
type: object
default:
@@ -87,6 +89,9 @@ parameters:
# list for vision
- auditwheel
- future
- pytest
- pytest-azurepipelines
- pillow
# list from https://github.com/pytorch/builder/blob/main/manywheel/build_rocm.sh
- name: rocmDependencies
type: object
@@ -117,6 +122,24 @@ parameters:
- hipCUB
- rocThrust
- hipBLAS-common
- name: rocmTestDependencies
type: object
default:
- rocminfo
# Reference on what tests to run for torchvision found in private repo:
# https://github.com/ROCm/rocAutomation/blob/jenkins-pipelines/pytorch/pytorch_ci/test_pytorch_test1.sh#L54
# Will iterate through this list using pytest
- name: torchTestList
type: object
default:
- nn
- torch
# - cuda seg faults and might need cuda installed on test system
# - ops takes too long
- unary_ufuncs
- binary_ufuncs
- autograd
# - inductor/torchinductor takes too long
trigger: none
pr: none
@@ -295,12 +318,22 @@ jobs:
targetType: inline
script: git clone https://github.com/pytorch/vision.git --depth=1 --recurse-submodules
workingDirectory: $(Build.SourcesDirectory)
- task: Bash@3
displayName: Apply vision patch
inputs:
targetType: inline
script: |
git apply $(Build.SourcesDirectory)/.azuredevops/patches/torchvision-package-name.patch
workingDirectory: $(Build.SourcesDirectory)/vision
- task: Bash@3
displayName: Build vision
inputs:
targetType: inline
script: >-
BUILD_VERSION=$(ROCM_BRANCH).$(JOB_GPU_TARGET)-$(cat $(Build.SourcesDirectory)/pytorch/version.txt | cut -da -f1)$(date -u +%Y%m%d)
TORCH_PACKAGE_NAME=torch.$(ROCM_BRANCH).$(JOB_GPU_TARGET)
TORCHVISION_PACKAGE_NAME=torchvision.$(ROCM_BRANCH).$(JOB_GPU_TARGET)
PYTORCH_VERSION=$(cat $(Build.SourcesDirectory)/pytorch/version.txt | cut -da -f1)post$(date -u +%Y%m%d)
BUILD_VERSION=$(cat $(Build.SourcesDirectory)/vision/version.txt | cut -da -f1)post$(date -u +%Y%m%d)
python3 setup.py bdist_wheel
workingDirectory: $(Build.SourcesDirectory)/vision
- task: Bash@3
@@ -319,3 +352,192 @@ jobs:
retryCountOnTaskFailure: 3
inputs:
targetPath: $(Build.BinariesDirectory)
- job: torchvision_testing
dependsOn: pytorch
condition: succeeded()
variables:
- group: common
- template: /.azuredevops/variables-global.yml
- name: PYTORCH_TEST_WITH_ROCM
value: 1
pool: $(JOB_TEST_POOL)
workspace:
clean: all
strategy:
matrix:
gfx942:
JOB_GPU_TARGET: gfx942
JOB_TEST_POOL: ${{ variables.GFX942_TEST_POOL }}
steps:
- task: Bash@3
displayName: 'Register libjpeg-turbo packages'
inputs:
targetType: inline
script: |
sudo mkdir --parents --mode=0755 /etc/apt/keyrings
wget -q -O- https://packagecloud.io/dcommander/libjpeg-turbo/gpgkey | gpg --dearmor | sudo tee /etc/apt/trusted.gpg.d/libjpeg-turbo.gpg > /dev/null
echo "deb [signed-by=/etc/apt/trusted.gpg.d/libjpeg-turbo.gpg] https://packagecloud.io/dcommander/libjpeg-turbo/any/ any main" | sudo tee /etc/apt/sources.list.d/libjpeg-turbo.list
sudo apt update
apt-cache show libjpeg-turbo-official | grep Version
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/dependencies-other.yml
parameters:
aptPackages: ${{ parameters.aptPackages }}
pipModules: ${{ parameters.pipModules }}
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/dependencies-cmake-latest.yml
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/preamble.yml
- task: DownloadPipelineArtifact@2
displayName: 'Download Pipeline Wheel Files'
inputs:
itemPattern: '**/*$(JOB_GPU_TARGET)*.whl'
targetPath: $(Agent.BuildDirectory)
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/dependencies-aqlprofile.yml
parameters:
dependencySource: staging
- template: ${{ variables.CI_TEMPLATE_PATH }}/steps/dependencies-rocm.yml
parameters:
dependencyList: ${{ parameters.rocmTestDependencies }}
gpuTarget: $(JOB_GPU_TARGET)
dependencySource: staging
skipLlvmSymlink: true
# get sources to run test scripts
- task: Bash@3
displayName: git clone upstream pytorch
inputs:
targetType: inline
script: git clone https://github.com/pytorch/pytorch.git --depth=1 --recurse-submodules
workingDirectory: $(Build.SourcesDirectory)
- task: Bash@3
displayName: git clone pytorch vision
inputs:
targetType: inline
script: git clone https://github.com/pytorch/vision.git --depth=1 --recurse-submodules
workingDirectory: $(Build.SourcesDirectory)
- task: Bash@3
displayName: Install Wheel Files
inputs:
targetType: inline
script: find . -name "*.whl" -exec pip install --no-index --find-links=. --no-dependencies -v {} \;
workingDirectory: $(Agent.BuildDirectory)
- task: Bash@3
displayName: Show Updated pip List
inputs:
targetType: inline
script: pip list -v
workingDirectory: $(Agent.BuildDirectory)
- task: Bash@3
displayName: Add ROCm binaries to PATH
inputs:
targetType: inline
script: echo "##vso[task.prependpath]$(Agent.BuildDirectory)/rocm/bin"
- task: Bash@3
displayName: Add Python site-packages binaries to path
inputs:
targetType: inline
script: |
USER_BASE=$(python3 -m site --user-base)
echo "##vso[task.prependpath]$USER_BASE/bin"
- task: Bash@3
displayName: Add torch libs to ldconfig
inputs:
targetType: inline
script: |
echo $(python3 -m site --user-site)/torch/lib | sudo tee /etc/ld.so.conf.d/torch.conf
sudo ldconfig -v
ldconfig -p
# https://pytorch.org/get-started/locally/#linux-verification
# https://rocm.docs.amd.com/projects/install-on-linux/en/latest/how-to/3rd-party/pytorch-install.html#testing-the-pytorch-installation
- task: Bash@3
displayName: Simple Import Torch Tests
inputs:
targetType: inline
script: |
python3 -c 'import torch' 2> /dev/null && echo 'Success' || echo 'Failure'
python3 -c 'import torch; print(torch.cuda.is_available())'
python3 -c 'import torch; x = torch.rand(5, 3); print(x)'
# Test artifact build script has too many if statements for different environments
# Based off the snippet of interest for this environment, with some adjustments
# https://github.com/pytorch/pytorch/blob/main/.ci/pytorch/build.sh#L335-L375
# Removing in-line comments since it does not fit with the yaml markup
- task: Bash@3
displayName: Build Pytorch Test Artifacts
continueOnError: true
inputs:
targetType: inline
script: |
CUSTOM_TEST_ARTIFACT_BUILD_DIR="build/custom_test_artifacts"
CUSTOM_TEST_USE_ROCM=ON
CUSTOM_TEST_MODULE_PATH="${PWD}/cmake/public"
mkdir -pv "${CUSTOM_TEST_ARTIFACT_BUILD_DIR}"
CUSTOM_OP_BUILD="${CUSTOM_TEST_ARTIFACT_BUILD_DIR}/custom-op-build"
CUSTOM_OP_TEST="${PWD}/test/custom_operator"
python --version
SITE_PACKAGES="$(python -c 'import site; print(";".join([x for x in site.getsitepackages()] + [x + "/torch" for x in site.getsitepackages()]))')"
mkdir -p "$CUSTOM_OP_BUILD"
pushd "$CUSTOM_OP_BUILD"
cmake "$CUSTOM_OP_TEST" -DCMAKE_PREFIX_PATH="$SITE_PACKAGES" -DPython_EXECUTABLE="$(which python)" \
-DCMAKE_MODULE_PATH="$CUSTOM_TEST_MODULE_PATH" -DUSE_ROCM="$CUSTOM_TEST_USE_ROCM"
make VERBOSE=1
popd
JIT_HOOK_BUILD="${CUSTOM_TEST_ARTIFACT_BUILD_DIR}/jit-hook-build"
JIT_HOOK_TEST="$PWD/test/jit_hooks"
python --version
SITE_PACKAGES="$(python -c 'import site; print(";".join([x for x in site.getsitepackages()] + [x + "/torch" for x in site.getsitepackages()]))')"
mkdir -p "$JIT_HOOK_BUILD"
pushd "$JIT_HOOK_BUILD"
cmake "$JIT_HOOK_TEST" -DCMAKE_PREFIX_PATH="$SITE_PACKAGES" -DPython_EXECUTABLE="$(which python)" \
-DCMAKE_MODULE_PATH="$CUSTOM_TEST_MODULE_PATH" -DUSE_ROCM="$CUSTOM_TEST_USE_ROCM"
make VERBOSE=1
popd
CUSTOM_BACKEND_BUILD="${CUSTOM_TEST_ARTIFACT_BUILD_DIR}/custom-backend-build"
CUSTOM_BACKEND_TEST="${PWD}/test/custom_backend"
python --version
mkdir -p "$CUSTOM_BACKEND_BUILD"
pushd "$CUSTOM_BACKEND_BUILD"
cmake "$CUSTOM_BACKEND_TEST" -DCMAKE_PREFIX_PATH="$SITE_PACKAGES" -DPython_EXECUTABLE="$(which python)" \
-DCMAKE_MODULE_PATH="$CUSTOM_TEST_MODULE_PATH" -DUSE_ROCM="$CUSTOM_TEST_USE_ROCM"
make VERBOSE=1
popd
workingDirectory: $(Build.SourcesDirectory)/pytorch
- ${{ each torchTest in parameters.torchTestList }}:
- task: Bash@3
displayName: Test ${{ torchTest }}
continueOnError: true
inputs:
targetType: inline
workingDirectory: $(Build.SourcesDirectory)/pytorch
${{ if contains(torchTest, '/') }}:
script: pytest test/${{ split(torchTest, '/')[0] }}/test_${{ split(torchTest, '/')[1] }}.py
${{ else }}:
script: pytest test/test_${{ torchTest }}.py
# Reference on what tests to run for torchvision found in private repo:
# https://github.com/ROCm/rocAutomation/blob/jenkins-pipelines/pytorch/pytorch_ci/test_torchvision.sh#L51
- task: Bash@3
displayName: Test vision/transforms
continueOnError: true
inputs:
targetType: inline
script: pytest test/test_transforms.py
workingDirectory: $(Build.SourcesDirectory)/vision
- task: Bash@3
displayName: Uninstall Wheel Files
inputs:
targetType: inline
script: find . -name "*.whl" -exec pip uninstall -y {} \;
workingDirectory: $(Agent.BuildDirectory)
- task: Bash@3
displayName: Remove Python site-packages binaries from path
inputs:
targetType: inline
script: |
USER_BASE=$(python3 -m site --user-base)
echo "##vso[task.setvariable variable=PATH]$(echo $PATH | sed -e 's;:$USER_BASE/bin;;' -e 's;^/;;' -e 's;/$;;')"
- task: Bash@3
displayName: Remove ROCm binaries from PATH
inputs:
targetType: inline
script: echo "##vso[task.setvariable variable=PATH]$(echo $PATH | sed -e 's;:$(Agent.BuildDirectory)/rocm/bin;;' -e 's;^/;;' -e 's;/$;;')"

View File

@@ -101,6 +101,7 @@ jobs:
dependencyList: ${{ parameters.rocmDependencies }}
dependencySource: staging
skipLibraryLinking: true
skipLlvmSymlink: true
gpuTarget: $(JOB_GPU_TARGET)
- script: df -h
displayName: System disk space after ROCm

View File

@@ -0,0 +1,34 @@
From 036307033d5c187b3123dae46477feacbd06d0ab Mon Sep 17 00:00:00 2001
From: Joseph Macaranas <Joseph.Macaranas@amd.com>
Date: Sun, 22 Sep 2024 23:03:48 -0400
Subject: [PATCH] Allow custom package name for CI builds of torchvision
---
setup.py | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/setup.py b/setup.py
index 4b0525d8e4..2c51ce04f2 100644
--- a/setup.py
+++ b/setup.py
@@ -42,7 +42,7 @@ CSRS_DIR = ROOT_DIR / "torchvision/csrc"
IS_ROCM = (torch.version.hip is not None) and (ROCM_HOME is not None)
BUILD_CUDA_SOURCES = (torch.cuda.is_available() and ((CUDA_HOME is not None) or IS_ROCM)) or FORCE_CUDA
-PACKAGE_NAME = "torchvision"
+PACKAGE_NAME = os.getenv("TORCHVISION_PACKAGE_NAME", "torchvision")
print("Torchvision build configuration:")
print(f"{FORCE_CUDA = }")
@@ -98,7 +98,7 @@ def get_requirements():
except DistributionNotFound:
return None
- pytorch_dep = "torch"
+ pytorch_dep = os.getenv("TORCH_PACKAGE_NAME", "torch")
if os.getenv("PYTORCH_VERSION"):
pytorch_dep += "==" + os.getenv("PYTORCH_VERSION")
--
2.44.0.windows.1

View File

@@ -0,0 +1,29 @@
variables:
- group: common
- template: /.azuredevops/variables-global.yml
parameters:
- name: checkoutRef
type: string
default: refs/tags/$(LATEST_RELEASE_TAG)
resources:
repositories:
- repository: pipelines_repo
type: github
endpoint: ROCm
name: ROCm/ROCm
- repository: release_repo
type: github
endpoint: ROCm
name: ROCm/omnitrace
ref: ${{ parameters.checkoutRef }}
trigger: none
pr: none
jobs:
- template: ${{ variables.CI_COMPONENT_PATH }}/omnitrace.yml
parameters:
checkoutRepo: release_repo
checkoutRef: ${{ parameters.checkoutRef }}

View File

@@ -0,0 +1,29 @@
variables:
- group: common
- template: /.azuredevops/variables-global.yml
parameters:
- name: checkoutRef
type: string
default: refs/tags/$(LATEST_RELEASE_TAG)
resources:
repositories:
- repository: pipelines_repo
type: github
endpoint: ROCm
name: ROCm/ROCm
- repository: release_repo
type: github
endpoint: ROCm
name: ROCm/rocprofiler-sdk
ref: ${{ parameters.checkoutRef }}
trigger: none
pr: none
jobs:
- template: ${{ variables.CI_COMPONENT_PATH }}/rocprofiler-sdk.yml
parameters:
checkoutRepo: release_repo
checkoutRef: ${{ parameters.checkoutRef }}

View File

@@ -29,6 +29,7 @@ parameters:
composable_kernel: develop
half: rocm
HIP: amd-staging
hip-tests: amd-staging
hipBLAS: develop
hipBLASLt: develop
hipBLAS-common: develop
@@ -44,6 +45,7 @@ parameters:
llvm-project: amd-staging
MIOpen: develop
MIVisionX: develop
omniperf: amd-staging
rccl: develop
rdc: develop
rocAL: develop
@@ -64,6 +66,7 @@ parameters:
rocPRIM: develop
rocprofiler-register: amd-mainline
rocprofiler: amd-staging
rocPyDecode: develop
ROCR-Runtime: amd-staging
rocRAND: develop
rocr_debug_agent: amd-staging
@@ -79,12 +82,14 @@ parameters:
- amdsmi
- aomp
- HIPIFY
- MIVisionX
- omniperf
- rccl
- rdc
- rocm-cmake
- rocm_smi_lib
- rocFFT
- MIVisionX
- rpp
# BELOW REQUIRED IF useDefaultBranch false
- name: branchName
type: string

View File

@@ -37,6 +37,7 @@ parameters:
composable_kernel: $(COMPOSABLE_KERNEL_PIPELINE_ID)
half: $(HALF_PIPELINE_ID)
HIP: $(HIP_PIPELINE_ID)
hip-tests: $(HIP_TESTS_PIPELINE_ID)
hipBLAS: $(HIPBLAS_PIPELINE_ID)
hipBLAS-common: $(HIPBLAS_COMMON_PIPELINE_ID)
hipBLASLt: $(HIPBLASLT_PIPELINE_ID)
@@ -73,6 +74,7 @@ parameters:
rocPRIM: $(ROCPRIM_PIPELINE_ID)
rocprofiler-register: $(ROCPROFILER_REGISTER_PIPELINE_ID)
rocprofiler: $(ROCPROFILER_PIPELINE_ID)
rocPyDecode: $(ROCPYDECODE_PIPELINE_ID)
ROCR-Runtime: $(ROCR_RUNTIME_PIPELINE_ID)
rocRAND: $(ROCRAND_PIPELINE_ID)
rocr_debug_agent: $(ROCR_DEBUG_AGENT_PIPELINE_ID)
@@ -94,6 +96,7 @@ parameters:
composable_kernel: $(COMPOSABLE_KERNEL_TAGGED_PIPELINE_ID)
half: $(HALF_TAGGED_PIPELINE_ID)
HIP: $(HIP_TAGGED_PIPELINE_ID)
hip-tests: $(HIP_TESTS_TAGGED_PIPELINE_ID)
hipBLAS: $(HIPBLAS_TAGGED_PIPELINE_ID)
hipBLAS-common: $(HIPBLAS_COMMON_TAGGED_PIPELINE_ID)
hipBLASLt: $(HIPBLASLT_TAGGED_PIPELINE_ID)
@@ -130,6 +133,7 @@ parameters:
rocPRIM: $(ROCPRIM_TAGGED_PIPELINE_ID)
rocprofiler-register: $(ROCPROFILER_REGISTER_TAGGED_PIPELINE_ID)
rocprofiler: $(ROCPROFILER_TAGGED_PIPELINE_ID)
rocPyDecode: $(ROCPYDECODE_TAGGED_PIPELINE_ID)
ROCR-Runtime: $(ROCR_RUNTIME_TAGGED_PIPELINE_ID)
rocRAND: $(ROCRAND_TAGGED_PIPELINE_ID)
rocr_debug_agent: $(ROCR_DEBUG_AGENT_TAGGED_PIPELINE_ID)
@@ -175,6 +179,7 @@ parameters:
- rocm-examples
- rocPRIM
- rocprofiler
- rocPyDecode
- rocRAND
- rocSOLVER
- rocSPARSE

View File

@@ -79,6 +79,10 @@ variables:
value: 93
- name: HIP_TAGGED_PIPELINE_ID
value: 31
- name: HIP_TESTS_PIPELINE_ID
value: 233
- name: HIP_TESTS_TAGGED_PIPELINE_ID
value: 220
- name: HIPBLAS_COMMON_PIPELINE_ID
value: 223
- name: HIPBLAS_COMMON_TAGGED_PIPELINE_ID
@@ -223,6 +227,10 @@ variables:
value: 22
- name: ROCM_EXAMPLES_GFX942_TEST_PIPELINE_ID
value: 204
- name: ROCM_EXAMPLES_PIPELINE_ID
value: 216
- name: ROCM_EXAMPLES_TAGGED_PIPELINE_ID
value: 245
- name: ROCM_SMI_LIB_PIPELINE_ID
value: 96
- name: ROCM_SMI_LIB_TAGGED_PIPELINE_ID
@@ -247,14 +255,18 @@ variables:
value: 20
- name: ROCPROFILER_GFX942_TEST_PIPELINE_ID
value: 190
- name: ROCPROFILER_PIPELINE_ID
value: 143
- name: ROCPROFILER_REGISTER_PIPELINE_ID
value: 1
- name: ROCPROFILER_REGISTER_TAGGED_PIPELINE_ID
value: 25
- name: ROCPROFILER_PIPELINE_ID
value: 143
- name: ROCPROFILER_TAGGED_PIPELINE_ID
value: 28
- name: ROCPYDECODE_PIPELINE_ID
value: 239
- name: ROCPYDECODE_TAGGED_PIPELINE_ID
value: 232
- name: ROCR_DEBUG_AGENT_PIPELINE_ID
value: 136
- name: ROCR_DEBUG_AGENT_TAGGED_PIPELINE_ID

View File

@@ -36,7 +36,6 @@ Bluefield
Bootloader
CCD
CDNA
CHTML
CIFAR
CLI
CLion
@@ -71,7 +70,6 @@ Concretized
Conda
ConnectX
CuPy
Dashboarding
DDR
DF
DGEMM
@@ -229,7 +227,6 @@ Mellanox's
Meta's
Miniconda
MirroredStrategy
Mixtral
Multicore
Multithreaded
MyEnvironment
@@ -275,8 +272,6 @@ OpenMPI
OpenSSL
OpenVX
OpenXLA
Oversubscription
PagedAttention
PCC
PCI
PCIe
@@ -298,7 +293,6 @@ PowerShell
PyPi
PyTorch
Qcycles
Qwen
RAII
RAS
RCCL
@@ -567,7 +561,6 @@ hipfort
hipify
hipsolver
hipsparse
hlist
hotspotting
hpc
hpp
@@ -591,7 +584,6 @@ intra
invariants
invocating
ipo
jax
kdb
kfd
latencies
@@ -612,7 +604,6 @@ migraphx
miopen
miopengemm
mivisionx
mjx
mkdir
mlirmiopen
mtypes

View File

View File

@@ -76,7 +76,7 @@ The Build time will reduce significantly if we limit the GPU Architecture/s agai
mkdir -p ~/WORKSPACE/ # Or any folder name other than WORKSPACE
cd ~/WORKSPACE/
export ROCM_VERSION=6.2.1
export ROCM_VERSION=6.2.2 # Or 6.2.0 or 6.2.1
~/bin/repo init -u http://github.com/ROCm/ROCm.git -b roc-6.2.x -m tools/rocm-build/rocm-${ROCM_VERSION}.xml
~/bin/repo sync

View File

@@ -1,6 +1,43 @@
# ROCm 6.2.1 release notes
# ROCm 6.2.2 release notes
The release notes provide a summary of notable changes since the previous ROCm release.
These release notes provide a summary of notable changes since the previous ROCm release.
```{note}
As ROCm 6.2.2 was released shortly after 6.2.1, the changes between these versions
are minimal. For a comprehensive overview of recent updates, the ROCm 6.2.1 release
notes are appended to the end of this document.
For detailed information about the changes in ROCm 6.2.1, refer to the appended
section: [ROCm 6.2.1 release notes](rocm-6-2-1-release-notes).
```
The [Compatibility matrix](https://rocm.docs.amd.com/en/docs-6.2.2/compatibility/compatibility-matrix.html)
provides the full list of supported hardware, operating systems, ecosystems, third-party components, and ROCm components
for each ROCm release.
Release notes for previous ROCm releases are available in earlier versions of the documentation.
See the [ROCm documentation release history](https://rocm.docs.amd.com/en/latest/release/versions.html).
## Release highlights
The following is a significant fix introduced in ROCm 6.2.2.
### Fixed Instinct MI300X error recovery failure
Improved the reliability of AMD Instinct MI300X accelerators in scenarios involving
uncorrectable errors. Previously, error recovery did not occur as expected,
potentially leaving the system in an undefined state. This fix ensures that error
recovery functions as expected, maintaining system stability.
See the [original issue](#instinct-mi300x-gpu-recovery-failure-on-uncorrectable-errors)
noted in the ROCm 6.2.1 release notes.
---
## ROCm 6.2.1 release notes
The ROCm 6.2.1 release notes document newly added ecosystem support, ROCm Offline Installer Creator updates,
and improvements to several ROCm libraries and tools.
- [Release highlights](release-highlights)
@@ -14,31 +51,25 @@ The release notes provide a summary of notable changes since the previous ROCm r
- [ROCm upcoming changes](rocm-upcoming-changes)
The [Compatibility matrix](https://rocm.docs.amd.com/en/docs-6.2.1/compatibility/compatibility-matrix.html)
provides the full list of supported hardware, operating systems, ecosystems, third-party components, and ROCm components for each ROCm release.
Release notes for previous ROCm releases are available in earlier versions of the documentation.
See the [ROCm documentation release history](https://rocm.docs.amd.com/en/latest/release/versions.html).
## Release highlights
### Release highlights
The following are notable new features and improvements in ROCm 6.2.1. For changes to individual components, see [Detailed component changes](#detailed-component-changes).
### rocAL major version change
#### rocAL major version change
The new version of rocAL introduces many new features, but does not modify any of the existing public API functions. However, the version number was incremented from 1.3 to 2.0.
Applications linked to version 1.3 must be recompiled to link against version 2.0.
See [the rocAL detailed changes](#rocal-2-0-0) for more information.
### New support for FBGEMM (Facebook General Matrix Multiplication)
#### New support for FBGEMM (Facebook General Matrix Multiplication)
As of ROCm 6.2.1, ROCm supports Facebook General Matrix Multiplication (FBGEMM) and the related FBGEMM_GPU library.
FBGEMM is a low-precision, high-performance CPU kernel library for convolution and matrix multiplication. It is used for server-side inference and as a back end for PyTorch quantized operators. FBGEMM_GPU includes a collection of PyTorch GPU operator libraries for training and inference. For more information, see the ROCm [Model acceleration libraries guide](https://rocm.docs.amd.com/en/docs-6.2.1/how-to/llm-fine-tuning-optimization/model-acceleration-libraries.html)
and [PyTorch's FBGEMM GitHub repository](https://github.com/pytorch/FBGEMM).
### ROCm Offline Installer Creator changes
#### ROCm Offline Installer Creator changes
The [ROCm Offline Installer Creator 6.2.1](https://rocm.docs.amd.com/projects/install-on-linux/en/docs-6.2.1/install/rocm-offline-installer.html) introduces several new features and improvements including:
@@ -47,27 +78,25 @@ The [ROCm Offline Installer Creator 6.2.1](https://rocm.docs.amd.com/projects/in
* Updated prerequisite repositories
* Fixed CTest issues
### ROCm documentation changes
#### ROCm documentation changes
The HIP documentation has been updated with several new topics aimed at improving usability
and providing more detailed information.
There have been no changes to supported hardware or operating systems from ROCm 6.2.0 to ROCm 6.2.1.
* The Programming Model Reference and Understanding the Programming Model topics in HIP have been consolidated into one topic,
[HIP programming model (conceptual)](https://rocm.docs.amd.com/projects/HIP/en/docs-6.2.1/understand/programming_model.html).
* The [HIP virtual memory management](https://rocm.docs.amd.com/projects/HIP/en/docs-6.2.1/how-to/virtual_memory.html) and [HIP virtual memory management API](https://rocm.docs.amd.com/projects/HIP/en/docs-6.2.1/reference/virtual_memory_reference.html) topics have been added.
```{note}
To contribute to ROCm documentation, see the [ROCm documentation contribution guidelines](https://rocm.docs.amd.com/en/latest/contribute/contributing.html).
The ROCm documentation, like all ROCm projects, is open source and available on GitHub. To contribute to ROCm documentation, see the [ROCm documentation contribution guidelines](https://rocm.docs.amd.com/en/latest/contribute/contributing.html).
```
## Operating system and hardware support changes
### Operating system and hardware support changes
ROCm 6.2.1 adds support for Ubuntu 24.04.1 (kernel: 6.8 [GA]).
See the [Compatibility matrix](https://rocm.docs.amd.com/en/docs-6.2.1/compatibility/compatibility-matrix.html) for the full list of supported operating systems and hardware architectures.
## ROCm components
### ROCm components
The following table lists the versions of ROCm components for ROCm 6.2.1, including any version changes from 6.2.0 to 6.2.1.
@@ -295,7 +324,7 @@ Click the component's updated version to go to a detailed list of its changes. C
</tr>
<tr>
<td><a href="https://rocm.docs.amd.com/projects/rdc/en/docs-6.2.1">ROCm Data Center Tool</a></td>
<td>0.3.0</td>
<td>1.0.0</td>
<td><a href="https://github.com/ROCm/rdc/releases/tag/rocm-6.2.1"><i
class="fab fa-github fa-lg"></i></a></td>
</tr>
@@ -317,7 +346,7 @@ Click the component's updated version to go to a detailed list of its changes. C
<th rowspan="6"></th>
<th rowspan="6">Performance</th>
<td><a href="https://rocm.docs.amd.com/projects/omniperf/en/docs-6.2.1">Omniperf</a></td>
<td>2.0.1&nbsp;&Rightarrow;&nbsp;<a href="#omniperf-2-0-1">2.0.1</a></td>
<td>2.0.1</td>
<td><a href="https://github.com/ROCm/omniperf/releases/tag/rocm-6.2.1"><i
class="fab fa-github fa-lg"></i></a></td>
</tr>
@@ -422,76 +451,55 @@ Click the component's updated version to go to a detailed list of its changes. C
</table>
</div>
## Detailed component changes
### Detailed component changes
The following sections describe key changes to ROCm components.
### **AMD SMI** (24.6.3)
#### **AMD SMI** (24.6.3)
#### Changes
##### Changes
* Added `amd-smi static --ras` on Guest VMs. Guest VMs can view enabled/disabled RAS features on Host cards.
#### Removals
##### Removals
* Removed `amd-smi metric --ecc` & `amd-smi metric --ecc-blocks` on Guest VMs. Guest VMs do not support getting current ECC counts from the Host cards.
#### Resolved issues
##### Resolved issues
* Fixed TypeError in `amd-smi process -G`.
* Updated CLI error strings to handle empty and invalid GPU/CPU inputs.
* Fixed Guest VM showing passthrough options.
* Fixed firmware formatting where leading 0s were missing.
### **HIP** (6.2.1)
#### **HIP** (6.2.1)
#### Resolved issues
##### Resolved issues
* Soft hang when using `AMD_SERIALIZE_KERNEL`
* Memory leak in `hipIpcCloseMemHandle`
### **HIPIFY** (18.0.0)
#### **HIPIFY** (18.0.0)
#### Changes
##### Changes
* Added CUDA 12.5.1 support.
* Added cuDNN 9.2.1 support.
* Added LLVM 18.1.8 support.
* Added `hipBLAS` 64-bit APIs support.
* Added Support for math constants `math_constants.h`.
* Added CUDA 12.5.1 support
* Added cuDNN 9.2.1 support
* Added LLVM 18.1.8 support
* Added `hipBLAS` 64-bit APIs support
* Added Support for math constants `math_constants.h`
### **Omniperf** (2.0.1)
#### **Omnitrace** (1.11.2)
#### Changes
* Enabled rocprofv1 for MI300 hardware.
* Added dependency checks on application launch.
* Updated Omniperf packaging.
* Rolled back Grafana version in Dockerfile for Angular plugin compatibility.
* Added GPU model distinction for MI300 systems.
* Refactored and updated documemtation.
#### Resolved issues
* Fixed an issue with analysis output.
* Fixed issues with profiling multi-process and multi-GPU applications.
#### Optimizations
* Reduced running time of Omniperf when profiling.
* Improved console logging.
### **Omnitrace** (1.11.2)
#### Known issues
##### Known issues
Perfetto can no longer open Omnitrace proto files. Loading Perfetto trace output `.proto` files in the latest version of `ui.perfetto.dev` can result in a dialog with the message, "Oops, something went wrong! Please file a bug." The information in the dialog will refer to an "Unknown field type." The workaround is to open the files with the previous version of the Perfetto UI found at [https://ui.perfetto.dev/v46.0-35b3d9845/#!/](https://ui.perfetto.dev/v46.0-35b3d9845/#!/).
See [issue #3767](https://github.com/ROCm/ROCm/issues/3767) on GitHub.
### **RCCL** (2.20.5)
#### **RCCL** (2.20.5)
#### Known issues
##### Known issues
On systems running Linux kernel 6.8.0, such as Ubuntu 24.04, Direct Memory Access (DMA) transfers between the GPU and NIC are disabled and impacts multi-node RCCL performance.
This issue was reproduced with RCCL 2.20.5 (ROCm 6.2.0 and 6.2.1) on systems with Broadcom Thor-2 NICs and affects other systems with RoCE networks using Linux 6.8.0 or newer.
@@ -501,9 +509,9 @@ This issue will be addressed in a future ROCm release.
See [issue #3772](https://github.com/ROCm/ROCm/issues/3772) on GitHub.
### **rocAL** (2.0.0)
#### **rocAL** (2.0.0)
#### Changes
##### Changes
* The new version of rocAL introduces many new features, but does not modify any of the existing public API functions.However, the version number was incremented from 1.3 to 2.0.
Applications linked to version 1.3 must be recompiled to link against version 2.0.
@@ -524,37 +532,37 @@ See [issue #3772](https://github.com/ROCm/ROCm/issues/3772) on GitHub.
* Image to tensor updates
* ROCm install - use case graphics removed
#### Known issues
##### Known issues
* Dependencies are not installed with the rocAL package installer. Dependencies must be installed with the prerequisite setup script provided. See the [rocAL README on GitHub](https://github.com/ROCm/rocAL/blob/docs/6.2.1/README.md#prerequisites-setup-script) for details.
### **rocBLAS** (4.2.1)
#### **rocBLAS** (4.2.1)
#### Removals
##### Removals
* Removed Device_Memory_Allocation.pdf link in documentation.
#### Resolved issues
##### Resolved issues
* Fixed error/warning message during `rocblas_set_stream()` call.
### **rocFFT** (1.0.29)
#### **rocFFT** (1.0.29)
#### Optimizations
##### Optimizations
* Implemented 1D kernels for factorizable sizes less than 1024.
### **ROCm SMI** (7.3.0)
#### **ROCm SMI** (7.3.0)
#### Optimizations
##### Optimizations
* Improved handling of UnicodeEncodeErrors with non UTF-8 locales. Non UTF-8 locales were causing crashes on UTF-8 special characters.
#### Resolved issues
##### Resolved issues
* Fixed an issue where the Compute Partition tests segfaulted when AMDGPU was loaded with optional parameters.
#### Known issues
##### Known issues
* When setting CPX as a partition mode, there is a DRM node limit of 64. This is a known limitation when multiple drivers are using the DRM nodes. The `ls /sys/class/drm` command can be used to see the number of DRM nodes, and the following steps can be used to remove unnecessary drivers:
@@ -562,18 +570,18 @@ See [issue #3772](https://github.com/ROCm/ROCm/issues/3772) on GitHub.
2. Remove any unnecessary drivers using `rmmod`. For example, to remove an AST driver, run `sudo rmmod ast`.
3. Reload AMDGPU using `modprobe`: `sudo modprobe amdgpu`.
### **rocPRIM** (3.2.1)
#### **rocPRIM** (3.2.1)
#### Optimizations
##### Optimizations
* Improved performance of `block_reduce_warp_reduce` when warp size equals block size.
## ROCm known issues
### ROCm known issues
ROCm known issues are tracked on [GitHub](https://github.com/ROCm/ROCm/labels/Verified%20Issue). Known issues related to
individual components are listed in the [Detailed component changes](detailed-component-changes) section.
### Instinct MI300X GPU recovery failure on uncorrectable errors
#### Instinct MI300X GPU recovery failure on uncorrectable errors
For the AMD Instinct MI300X accelerator, GPU recovery resets triggered by uncorrectable errors (UE) might not complete
successfully, which can result in the system being left in an undefined state. A system reboot is needed to recover from
@@ -583,14 +591,14 @@ This issue is under investigation and will be resolved in a future ROCm release.
See [issue #3766](https://github.com/ROCm/ROCm/issues/3766) on GitHub.
## ROCm upcoming changes
### ROCm upcoming changes
The following changes to the ROCm software stack are anticipated for future releases.
### rocm-llvm-alt
#### rocm-llvm-alt
The `rocm-llvm-alt` package will be removed in an upcoming release. Users relying on the functionality provided by the closed-source compiler should transition to the open-source compiler. Once the `rocm-llvm-alt` package is removed, any compilation requesting functionality provided by the closed-source compiler will result in a Clang warning: "*[AMD] proprietary optimization compiler has been removed*".
### rccl-rdma-sharp-plugins
#### rccl-rdma-sharp-plugins
The RCCL plugin package, `rccl-rdma-sharp-plugins`, will be removed in an upcoming ROCm release.

View File

@@ -1,7 +1,7 @@
<?xml version="1.0" encoding="UTF-8"?>
<manifest>
<remote name="rocm-org" fetch="https://github.com/ROCm/" />
<default revision="refs/tags/rocm-6.2.1"
<default revision="refs/tags/rocm-6.2.2"
remote="rocm-org"
sync-c="true"
sync-j="4" />

View File

@@ -91,8 +91,7 @@ additional licenses. Please review individual repositories for more information.
Open sourced ROCm components are released via public GitHub
repositories, packages on [https://repo.radeon.com](https://repo.radeon.com) and other distribution channels.
Proprietary products are only available on [https://repo.radeon.com](https://repo.radeon.com). Currently, only
one component of ROCm, `rocm-llvm-alt` is governed by a proprietary license.
Proprietary products are only available on [https://repo.radeon.com](https://repo.radeon.com).
Proprietary components are organized in a proprietary subdirectory in the package
repositories to distinguish from open sourced packages.

View File

@@ -1,10 +1,10 @@
ROCm Version,6.2.1,6.2.0, 6.1.5, 6.1.2, 6.1.1, 6.1.0, 6.0.2, 6.0.0
:ref:`Operating Systems & kernels <OS-kernel-versions>`,"Ubuntu 24.04.1, 24.04",Ubuntu 24.04,,,,,,
,"Ubuntu 22.04.5, 22.04.4","Ubuntu 22.04.5, 22.04.4","Ubuntu 22.04.5, 22.04.4, 22.04.3","Ubuntu 22.04.5, 22.04.4, 22.04.3","Ubuntu 22.04.4, 22.04.3","Ubuntu 22.04.4, 22.04.3","Ubuntu 22.04.4, 22.04.3, 22.04.2","Ubuntu 22.04.4, 22.04.3, 22.04.2"
,,,"Ubuntu 20.04.6, 20.04.5","Ubuntu 20.04.6, 20.04.5","Ubuntu 20.04.6, 20.04.5","Ubuntu 20.04.6, 20.04.5","Ubuntu 20.04.6, 20.04.5","Ubuntu 20.04.6, 20.04.5"
,"RHEL 9.4, 9.3","RHEL 9.4, 9.3","RHEL 9.4 [#red-hat94-past-60]_, 9.3, 9.2","RHEL 9.4 [#red-hat94-past-60]_, 9.3, 9.2","RHEL 9.4 [#red-hat94-past-60]_, 9.3, 9.2","RHEL 9.4 [#red-hat94-past-60]_, 9.3, 9.2","RHEL 9.3, 9.2","RHEL 9.3, 9.2"
,"RHEL 8.10, 8.9","RHEL 8.10, 8.9","RHEL 8.9, 8.8","RHEL 8.9, 8.8","RHEL 8.9, 8.8","RHEL 8.9, 8.8","RHEL 8.9, 8.8","RHEL 8.9, 8.8"
,"SLES 15 SP6, SP5","SLES 15 SP6, SP5","SLES 15 SP5, SP4","SLES 15 SP5, SP4","SLES 15 SP5, SP4","SLES 15 SP5, SP4","SLES 15 SP5, SP4","SLES 15 SP5, SP4"
ROCm Version,6.2.2,6.2.1,6.2.0, 6.1.2, 6.1.1, 6.1.0, 6.0.2, 6.0.0
:ref:`Operating systems & kernels <OS-kernel-versions>`,"Ubuntu 24.04.1, 24.04","Ubuntu 24.04.1, 24.04",Ubuntu 24.04,,,,,
,"Ubuntu 22.04.5, 22.04.4","Ubuntu 22.04.5, 22.04.4","Ubuntu 22.04.5, 22.04.4","Ubuntu 22.04.4, 22.04.3","Ubuntu 22.04.4, 22.04.3","Ubuntu 22.04.4, 22.04.3","Ubuntu 22.04.4, 22.04.3, 22.04.2","Ubuntu 22.04.4, 22.04.3, 22.04.2"
,,,,"Ubuntu 20.04.6, 20.04.5","Ubuntu 20.04.6, 20.04.5","Ubuntu 20.04.6, 20.04.5","Ubuntu 20.04.6, 20.04.5","Ubuntu 20.04.6, 20.04.5"
,"RHEL 9.4, 9.3","RHEL 9.4, 9.3","RHEL 9.4, 9.3","RHEL 9.4 [#red-hat94-past-60]_, 9.3, 9.2","RHEL 9.4 [#red-hat94-past-60]_, 9.3, 9.2","RHEL 9.4 [#red-hat94-past-60]_, 9.3, 9.2","RHEL 9.3, 9.2","RHEL 9.3, 9.2"
,"RHEL 8.10, 8.9","RHEL 8.10, 8.9","RHEL 8.10, 8.9","RHEL 8.9, 8.8","RHEL 8.9, 8.8","RHEL 8.9, 8.8","RHEL 8.9, 8.8","RHEL 8.9, 8.8"
,"SLES 15 SP6, SP5","SLES 15 SP6, SP5","SLES 15 SP6, SP5","SLES 15 SP5, SP4","SLES 15 SP5, SP4","SLES 15 SP5, SP4","SLES 15 SP5, SP4","SLES 15 SP5, SP4"
,,,,CentOS 7.9,CentOS 7.9,CentOS 7.9,CentOS 7.9,CentOS 7.9
,Oracle Linux 8.9 [#oracle89-past-60]_,Oracle Linux 8.9 [#oracle89-past-60]_,Oracle Linux 8.9 [#oracle89-past-60]_,Oracle Linux 8.9 [#oracle89-past-60]_,Oracle Linux 8.9 [#oracle89-past-60]_,,,
,.. _architecture-support-compatibility-matrix-past-60:,,,,,,,
@@ -16,102 +16,102 @@ ROCm Version,6.2.1,6.2.0, 6.1.5, 6.1.2, 6.1.1, 6.1.0, 6.0.2, 6.0.0
,.. _gpu-support-compatibility-matrix-past-60:,,,,,,,
:doc:`GPU / LLVM target <rocm-install-on-linux:reference/system-requirements>`,gfx1100,gfx1100,gfx1100,gfx1100,gfx1100,gfx1100,gfx1100,gfx1100
,gfx1030,gfx1030,gfx1030,gfx1030,gfx1030,gfx1030,gfx1030,gfx1030
,gfx942 [#mi300_621-past-60]_,gfx942 [#mi300_620-past-60]_, gfx942 [#mi300_612-past-60]_, gfx942 [#mi300_612-past-60]_, gfx942 [#mi300_611-past-60]_, gfx942 [#mi300_610-past-60]_, gfx942 [#mi300_602-past-60]_, gfx942 [#mi300_600-past-60]_
,gfx942 [#mi300_622-past-60]_,gfx942 [#mi300_621-past-60]_,gfx942 [#mi300_620-past-60]_, gfx942 [#mi300_612-past-60]_, gfx942 [#mi300_611-past-60]_, gfx942 [#mi300_610-past-60]_, gfx942 [#mi300_602-past-60]_, gfx942 [#mi300_600-past-60]_
,gfx90a,gfx90a,gfx90a,gfx90a,gfx90a,gfx90a,gfx90a,gfx90a
,gfx908,gfx908,gfx908,gfx908,gfx908,gfx908,gfx908,gfx908
,,,,,,,,
FRAMEWORK SUPPORT,.. _framework-support-compatibility-matrix-past-60:,,,,,,,
:doc:`PyTorch <rocm-install-on-linux:install/3rd-party/pytorch-install>`,"2.3, 2.2, 2.1, 2.0, 1.13","2.3, 2.2, 2.1, 2.0, 1.13","2.1, 2.0, 1.13","2.1, 2.0, 1.13","2.1, 2.0, 1.13","2.1, 2.0, 1.13","2.1, 2.0, 1.13","2.1, 2.0, 1.13"
:doc:`TensorFlow <rocm-install-on-linux:install/3rd-party/tensorflow-install>`,"2.16.1, 2.15.1, 2.14.1","2.16.1, 2.15.1, 2.14.1","2.15.0, 2.14.0, 2.13.1","2.15.0, 2.14.0, 2.13.1","2.15.0, 2.14.0, 2.13.1","2.15.0, 2.14.0, 2.13.1","2.14.0, 2.13.1, 2.12.1","2.14.0, 2.13.1, 2.12.1"
:doc:`PyTorch <rocm-install-on-linux:install/3rd-party/pytorch-install>`,"2.3, 2.2, 2.1, 2.0, 1.13","2.3, 2.2, 2.1, 2.0, 1.13","2.3, 2.2, 2.1, 2.0, 1.13","2.1, 2.0, 1.13","2.1, 2.0, 1.13","2.1, 2.0, 1.13","2.1, 2.0, 1.13","2.1, 2.0, 1.13"
:doc:`TensorFlow <rocm-install-on-linux:install/3rd-party/tensorflow-install>`,"2.16.1, 2.15.1, 2.14.1","2.16.1, 2.15.1, 2.14.1","2.16.1, 2.15.1, 2.14.1","2.15.0, 2.14.0, 2.13.1","2.15.0, 2.14.0, 2.13.1","2.15.0, 2.14.0, 2.13.1","2.14.0, 2.13.1, 2.12.1","2.14.0, 2.13.1, 2.12.1"
:doc:`JAX <rocm-install-on-linux:install/3rd-party/jax-install>`,0.4.26,0.4.26,0.4.26,0.4.26,0.4.26,0.4.26,0.4.26,0.4.26
`ONNX Runtime <https://onnxruntime.ai/docs/build/eps.html#amd-migraphx>`_,1.17.3,1.17.3,1.17.3,1.17.3,1.17.3,1.17.3,1.14.1,1.14.1
,,,,,,,,
THIRD PARTY COMMS,.. _thirdpartycomms-support-compatibility-matrix-past-60:,,,,,,,
`UCC <https://github.com/ROCm/ucc>`_,>=1.3.0,>=1.3.0,>=1.3.0,>=1.3.0,>=1.3.0,>=1.3.0,>=1.2.0,>=1.2.0
`UCX <https://github.com/ROCm/ucx>`_,>=1.15.0,>=1.15.0,>=1.14.1,>=1.14.1,>=1.14.1,>=1.14.1,>=1.14.1,>=1.14.1
`UCX <https://github.com/ROCm/ucx>`_,>=1.15.0,>=1.15.0,>=1.15.0,>=1.14.1,>=1.14.1,>=1.14.1,>=1.14.1,>=1.14.1
,,,,,,,,
THIRD PARTY ALGORITHM,.. _thirdpartyalgorithm-support-compatibility-matrix-past-60:,,,,,,,
Thrust,2.2.0,2.2.0,2.1.0,2.1.0,2.1.0,2.1.0,2.0.1,2.0.1
CUB,2.2.0,2.2.0,2.1.0,2.1.0,2.1.0,2.1.0,2.0.1,2.0.1
Thrust,2.2.0,2.2.0,2.2.0,2.1.0,2.1.0,2.1.0,2.0.1,2.0.1
CUB,2.2.0,2.2.0,2.2.0,2.1.0,2.1.0,2.1.0,2.0.1,2.0.1
,,,,,,,,
KFD & USER SPACE [#kfd_support-past-60]_,.. _kfd-userspace-support-compatibility-matrix-past-60:,,,,,,,
Tested user space versions,"6.1.x, 6.0.x","6.1.x, 6.0.x","6.2.x, 6.0.x, 5.7.x","6.2.x, 6.0.x, 5.7.x","6.2.x, 6.0.x, 5.7.x","6.2.x, 6.0.x, 5.7.x","6.2.x, 6.0.x, 5.7.x, 5.6.x","6.2.x, 6.0.x, 5.7.x, 5.6.x"
Tested user space versions,"6.1.x, 6.0.x","6.1.x, 6.0.x","6.1.x, 6.0.x","6.2.x, 6.0.x, 5.7.x","6.2.x, 6.0.x, 5.7.x","6.2.x, 6.0.x, 5.7.x","6.2.x, 6.0.x, 5.7.x, 5.6.x","6.2.x, 6.0.x, 5.7.x, 5.6.x"
,,,,,,,,
ML & COMPUTER VISION,.. _mllibs-support-compatibility-matrix-past-60:,,,,,,,
:doc:`Composable Kernel <composable_kernel:index>`,1.1.0,1.1.0,1.1.0,1.1.0,1.1.0,1.1.0,1.1.0,1.1.0
:doc:`MIGraphX <amdmigraphx:index>`,2.10.0,2.10.0,2.9.0,2.9.0,2.9.0,2.9.0,2.8.0,2.8.0
:doc:`MIOpen <miopen:index>`,3.2.0,3.2.0,3.1.0,3.1.0,3.1.0,3.1.0,3.0.0,3.0.0
:doc:`MIVisionX <mivisionx:index>`,3.0.0,3.0.0,2.5.0,2.5.0,2.5.0,2.5.0,2.5.0,2.5.0
:doc:`rocAL <rocal:index>`,2.0.0,1.0.0,1.0.0,1.0.0,1.0.0,1.0.0,1.0.0,1.0.0
:doc:`MIGraphX <amdmigraphx:index>`,2.10.0,2.10.0,2.10.0,2.9.0,2.9.0,2.9.0,2.8.0,2.8.0
:doc:`MIOpen <miopen:index>`,3.2.0,3.2.0,3.2.0,3.1.0,3.1.0,3.1.0,3.0.0,3.0.0
:doc:`MIVisionX <mivisionx:index>`,3.0.0,3.0.0,3.0.0,2.5.0,2.5.0,2.5.0,2.5.0,2.5.0
:doc:`rocAL <rocal:index>`,2.0.0,2.0.0,1.0.0,1.0.0,1.0.0,1.0.0,1.0.0,1.0.0
:doc:`rocDecode <rocdecode:index>`,0.6.0,0.6.0,0.6.0,0.6.0,0.5.0,0.5.0,N/A,N/A
:doc:`rocPyDecode <rocpydecode:index>`,0.1.0,0.1.0,N/A,N/A,N/A,N/A,N/A,N/A
:doc:`RPP <rpp:index>`,1.8.0,1.8.0,1.5.0,1.5.0,1.5.0,1.5.0,1.4.0,1.4.0
:doc:`rocPyDecode <rocpydecode:index>`,0.1.0,0.1.0,0.1.0,N/A,N/A,N/A,N/A,N/A
:doc:`RPP <rpp:index>`,1.8.0,1.8.0,1.8.0,1.5.0,1.5.0,1.5.0,1.4.0,1.4.0
,,,,,,,,
COMMUNICATION,.. _commlibs-support-compatibility-matrix-past-60:,,,,,,,
:doc:`RCCL <rccl:index>`,2.20.5,2.20.5,2.18.6,2.18.6,2.18.6,2.18.6,2.18.3,2.18.3
:doc:`RCCL <rccl:index>`,2.20.5,2.20.5,2.20.5,2.18.6,2.18.6,2.18.6,2.18.3,2.18.3
,,,,,,,,
MATH LIBS,.. _mathlibs-support-compatibility-matrix-past-60:,,,,,,,
`half <https://github.com/ROCm/half>`_ ,1.12.0,1.12.0,1.12.0,1.12.0,1.12.0,1.12.0,1.12.0,1.12.0
:doc:`hipBLAS <hipblas:index>`,2.2.0,2.2.0,2.1.0,2.1.0,2.1.0,2.1.0,2.0.0,2.0.0
:doc:`hipBLASLt <hipblaslt:index>`,0.8.0,0.8.0,0.7.0,0.7.0,0.7.0,0.7.0,0.6.0,0.6.0
:doc:`hipFFT <hipfft:index>`,1.0.15,1.0.14,1.0.14,1.0.14,1.0.14,1.0.14,1.0.13,1.0.13
:doc:`hipBLAS <hipblas:index>`,2.2.0,2.2.0,2.2.0,2.1.0,2.1.0,2.1.0,2.0.0,2.0.0
:doc:`hipBLASLt <hipblaslt:index>`,0.8.0,0.8.0,0.8.0,0.7.0,0.7.0,0.7.0,0.6.0,0.6.0
:doc:`hipFFT <hipfft:index>`,1.0.15,1.0.15,1.0.14,1.0.14,1.0.14,1.0.14,1.0.13,1.0.13
:doc:`hipFORT <hipfort:index>`,0.4.0,0.4.0,0.4.0,0.4.0,0.4.0,0.4.0,0.4.0,0.4.0
:doc:`hipRAND <hiprand:index>`,2.11.0,2.11.0,2.10.16,2.10.16,2.10.16,2.10.16,2.10.16,2.10.16
:doc:`hipSOLVER <hipsolver:index>`,2.2.0,2.2.0,2.1.1,2.1.1,2.1.1,2.1.0,2.0.0,2.0.0
:doc:`hipSPARSE <hipsparse:index>`,3.1.1,3.1.1,3.0.1,3.0.1,3.0.1,3.0.1,3.0.0,3.0.0
:doc:`hipSPARSELt <hipsparselt:index>`,0.2.1,0.2.1,0.2.0,0.2.0,0.1.0,0.1.0,0.1.0,0.1.0
:doc:`rocALUTION <rocalution:index>`,3.2.0,3.2.0,3.1.1,3.1.1,3.1.1,3.1.1,3.0.3,3.0.3
:doc:`rocBLAS <rocblas:index>`,4.2.1,4.2.0,4.1.2,4.1.2,4.1.0,4.1.0,4.0.0,4.0.0
:doc:`rocFFT <rocfft:index>`,1.0.29,1.0.28,1.0.27,1.0.27,1.0.27,1.0.26,1.0.25,1.0.23
:doc:`rocRAND <rocrand:index>`,3.1.0,3.1.0,3.0.1,3.0.1,3.0.1,3.0.1,3.0.0,2.10.17
:doc:`rocSOLVER <rocsolver:index>`,3.26.0,3.26.0,3.25.0,3.25.0,3.25.0,3.25.0,3.24.0,3.24.0
:doc:`rocSPARSE <rocsparse:index>`,3.2.0,3.2.0,3.1.2,3.1.2,3.1.2,3.1.2,3.0.2,3.0.2
:doc:`rocWMMA <rocwmma:index>`,1.5.0,1.5.0,1.4.0,1.4.0,1.4.0,1.4.0,1.3.0,1.3.0
:doc:`hipRAND <hiprand:index>`,2.11.0,2.11.0,2.11.0,2.10.16,2.10.16,2.10.16,2.10.16,2.10.16
:doc:`hipSOLVER <hipsolver:index>`,2.2.0,2.2.0,2.2.0,2.1.1,2.1.1,2.1.0,2.0.0,2.0.0
:doc:`hipSPARSE <hipsparse:index>`,3.1.1,3.1.1,3.1.1,3.0.1,3.0.1,3.0.1,3.0.0,3.0.0
:doc:`hipSPARSELt <hipsparselt:index>`,0.2.1,0.2.1,0.2.1,0.2.0,0.1.0,0.1.0,0.1.0,0.1.0
:doc:`rocALUTION <rocalution:index>`,3.2.0,3.2.0,3.2.0,3.1.1,3.1.1,3.1.1,3.0.3,3.0.3
:doc:`rocBLAS <rocblas:index>`,4.2.1,4.2.1,4.2.0,4.1.2,4.1.0,4.1.0,4.0.0,4.0.0
:doc:`rocFFT <rocfft:index>`,1.0.29,1.0.29,1.0.28,1.0.27,1.0.27,1.0.26,1.0.25,1.0.23
:doc:`rocRAND <rocrand:index>`,3.1.0,3.1.0,3.1.0,3.0.1,3.0.1,3.0.1,3.0.0,2.10.17
:doc:`rocSOLVER <rocsolver:index>`,3.26.0,3.26.0,3.26.0,3.25.0,3.25.0,3.25.0,3.24.0,3.24.0
:doc:`rocSPARSE <rocsparse:index>`,3.2.0,3.2.0,3.2.0,3.1.2,3.1.2,3.1.2,3.0.2,3.0.2
:doc:`rocWMMA <rocwmma:index>`,1.5.0,1.5.0,1.5.0,1.4.0,1.4.0,1.4.0,1.3.0,1.3.0
`Tensile <https://github.com/ROCm/Tensile>`_,4.40.0,4.40.0,4.40.0,4.40.0,4.40.0,4.40.0,4.39.0,4.39.0
,,,,,,,,
PRIMITIVES,.. _primitivelibs-support-compatibility-matrix-past-60:,,,,,,,
:doc:`hipCUB <hipcub:index>`,3.2.0,3.2.0,3.1.0,3.1.0,3.1.0,3.1.0,3.0.0,3.0.0
:doc:`hipTensor <hiptensor:index>`,1.3.0,1.3.0,1.2.0,1.2.0,1.2.0,1.2.0,1.1.0,1.1.0
:doc:`rocPRIM <rocprim:index>`,3.2.0,3.2.0,3.1.0,3.1.0,3.1.0,3.1.0,3.0.0,3.0.0
:doc:`rocThrust <rocthrust:index>`,3.1.0,3.0.1,3.0.1,3.0.1,3.0.1,3.0.1,3.0.0,3.0.0
:doc:`hipCUB <hipcub:index>`,3.2.0,3.2.0,3.2.0,3.1.0,3.1.0,3.1.0,3.0.0,3.0.0
:doc:`hipTensor <hiptensor:index>`,1.3.0,1.3.0,1.3.0,1.2.0,1.2.0,1.2.0,1.1.0,1.1.0
:doc:`rocPRIM <rocprim:index>`,3.2.0,3.2.0,3.2.0,3.1.0,3.1.0,3.1.0,3.0.0,3.0.0
:doc:`rocThrust <rocthrust:index>`,3.1.0,3.1.0,3.0.1,3.0.1,3.0.1,3.0.1,3.0.0,3.0.0
,,,,,,,,
SUPPORT LIBS,,,,,,,,
`hipother <https://github.com/ROCm/hipother>`_,6.2.41134,6.2.41133,6.1.40093,6.1.40093,6.1.40092,6.1.40091,6.1.32831,6.1.32830
`rocm-core <https://github.com/ROCm/rocm-core>`_,6.2.1,6.2.0,6.1.5,6.1.2,6.1.1,6.1.0,6.0.2,6.0.0
`ROCT-Thunk-Interface <https://github.com/ROCm/ROCT-Thunk-Interface>`_,20240607.4.05,20240607.1.4246,20240125.5.08,20240125.5.08,20240125.5.08,20240125.3.30,20231016.2.245,20231016.2.245
`hipother <https://github.com/ROCm/hipother>`_,6.2.41134,6.2.41134,6.2.41133,6.1.40093,6.1.40092,6.1.40091,6.1.32831,6.1.32830
`rocm-core <https://github.com/ROCm/rocm-core>`_,6.2.2,6.2.1,6.2.0,6.1.2,6.1.1,6.1.0,6.0.2,6.0.0
`ROCT-Thunk-Interface <https://github.com/ROCm/ROCT-Thunk-Interface>`_,20240607.4.05,20240607.4.05,20240607.1.4246,20240125.5.08,20240125.5.08,20240125.3.30,20231016.2.245,20231016.2.245
,,,,,,,,
SYSTEM MGMT TOOLS,.. _tools-support-compatibility-matrix-past-60:,,,,,,,
:doc:`AMD SMI <amdsmi:index>`,24.6.3,24.6.2,24.5.1,24.5.1,24.5.1,24.4.1,23.4.2,23.4.2
:doc:`ROCm Data Center Tool <rdc:index>`,0.3.0,0.3.0,0.3.0,0.3.0,0.3.0,0.3.0,0.3.0,0.3.0
:doc:`AMD SMI <amdsmi:index>`,24.6.3,24.6.3,24.6.2,24.5.1,24.5.1,24.4.1,23.4.2,23.4.2
:doc:`ROCm Data Center Tool <rdc:index>`,1.0.0,1.0.0,1.0.0,0.3.0,0.3.0,0.3.0,0.3.0,0.3.0
:doc:`rocminfo <rocminfo:index>`,1.0.0,1.0.0,1.0.0,1.0.0,1.0.0,1.0.0,1.0.0,1.0.0
:doc:`ROCm SMI <rocm_smi_lib:index>`,7.3.0,7.3.0,7.2.0,7.2.0,7.0.0,7.0.0,6.0.2,6.0.0
:doc:`ROCm Validation Suite <rocmvalidationsuite:index>`,rocm-6.2.1,rocm-6.2.0,rocm-6.1.5,rocm-6.1.2,rocm-6.1.1,rocm-6.1.0,rocm-6.0.2,rocm-6.0.0
:doc:`ROCm SMI <rocm_smi_lib:index>`,7.3.0,7.3.0,7.3.0,7.2.0,7.0.0,7.0.0,6.0.2,6.0.0
:doc:`ROCm Validation Suite <rocmvalidationsuite:index>`,rocm-6.2.2,rocm-6.2.1,rocm-6.2.0,rocm-6.1.2,rocm-6.1.1,rocm-6.1.0,rocm-6.0.2,rocm-6.0.0
,,,,,,,,
PERFORMANCE TOOLS,,,,,,,,
:doc:`Omniperf <omniperf:index>`,2.0.1,2.0.1,N/A,N/A,N/A,N/A,N/A,N/A
:doc:`Omnitrace <omnitrace:index>`,1.11.2,1.11.2,N/A,N/A,N/A,N/A,N/A,N/A
:doc:`Omniperf <omniperf:index>`,2.0.1,2.0.1,2.0.1,N/A,N/A,N/A,N/A,N/A
:doc:`Omnitrace <omnitrace:index>`,1.11.2,1.11.2,1.11.2,N/A,N/A,N/A,N/A,N/A
:doc:`ROCm Bandwidth Test <rocm_bandwidth_test:index>`,1.4.0,1.4.0,1.4.0,1.4.0,1.4.0,1.4.0,1.4.0,1.4.0
:doc:`ROCProfiler <rocprofiler:index>`,2.0.60201,2.0.60200,2.0.60105,2.0.60102,2.0.60101,2.0.60100,2.0.60002,2.0.60000
:doc:`ROCprofiler-SDK <rocprofiler-sdk:index>`,0.4.0,0.4.0,N/A,N/A,N/A,N/A,N/A,N/A
:doc:`ROCTracer <roctracer:index>`,4.1.60201,4.1.60200,4.1.60105,4.1.60102,4.1.60101,4.1.60100,4.1.60002,4.1.60000
:doc:`ROCProfiler <rocprofiler:index>`,2.0.60202,2.0.60201,2.0.60200,2.0.60102,2.0.60101,2.0.60100,2.0.60002,2.0.60000
:doc:`ROCprofiler-SDK <rocprofiler-sdk:index>`,0.4.0,0.4.0,0.4.0,N/A,N/A,N/A,N/A,N/A
:doc:`ROCTracer <roctracer:index>`,4.1.60202,4.1.60201,4.1.60200,4.1.60102,4.1.60101,4.1.60100,4.1.60002,4.1.60000
,,,,,,,,
DEVELOPMENT TOOLS,,,,,,,,
:doc:`HIPIFY <hipify:index>`,18.0.0.24355,18.0.0.24232,17.0.0.24193,17.0.0.24193,17.0.0.24154,17.0.0.24103,17.0.0.24012,17.0.0.23483
:doc:`ROCm CMake <rocmcmakebuildtools:index>`,0.13.0,0.13.0,0.12.0,0.12.0,0.12.0,0.12.0,0.11.0,0.11.0
:doc:`ROCdbgapi <rocdbgapi:index>`,0.76.0,0.76.0,0.71.0,0.71.0,0.71.0,0.71.0,0.71.0,0.71.0
:doc:`ROCm Debugger (ROCgdb) <rocgdb:index>`,14.2.0,14.2.0,14.1.0,14.1.0,14.1.0,14.1.0,13.2.0,13.2.0
`rocprofiler-register <https://github.com/ROCm/rocprofiler-register>`_,0.4.0,0.4.0,0.3.0,0.3.0,0.3.0,0.3.0,N/A,N/A
:doc:`HIPIFY <hipify:index>`,18.0.0.24355,18.0.0.24355,18.0.0.24232,17.0.0.24193,17.0.0.24154,17.0.0.24103,17.0.0.24012,17.0.0.23483
:doc:`ROCm CMake <rocmcmakebuildtools:index>`,0.13.0,0.13.0,0.13.0,0.12.0,0.12.0,0.12.0,0.11.0,0.11.0
:doc:`ROCdbgapi <rocdbgapi:index>`,0.76.0,0.76.0,0.76.0,0.71.0,0.71.0,0.71.0,0.71.0,0.71.0
:doc:`ROCm Debugger (ROCgdb) <rocgdb:index>`,14.2.0,14.2.0,14.2.0,14.1.0,14.1.0,14.1.0,13.2.0,13.2.0
`rocprofiler-register <https://github.com/ROCm/rocprofiler-register>`_,0.4.0,0.4.0,0.4.0,0.3.0,0.3.0,0.3.0,N/A,N/A
:doc:`ROCr Debug Agent <rocr_debug_agent:index>`,2.0.3,2.0.3,2.0.3,2.0.3,2.0.3,2.0.3,2.0.3,2.0.3
,,,,,,,,
COMPILERS,.. _compilers-support-compatibility-matrix-past-60:,,,,,,,
`clang-ocl <https://github.com/ROCm/clang-ocl>`_,N/A,N/A,0.5.0,0.5.0,0.5.0,0.5.0,0.5.0,0.5.0
:doc:`hipCC <hipcc:index>`,1.1.1,1.1.1,1.0.0,1.0.0,1.0.0,1.0.0,1.0.0,1.0.0
`Flang <https://github.com/ROCm/flang>`_,18.0.0.24355,18.0.0.24232,17.0.0.24193,17.0.0.24193,17.0.0.24154,17.0.0.24103,17.0.0.24012,17.0.0.23483
:doc:`llvm-project <llvm-project:index>`,18.0.0.24355,18.0.0.24232,17.0.0.24193,17.0.0.24193,17.0.0.24154,17.0.0.24103,17.0.0.24012,17.0.0.23483
`OpenMP <https://github.com/ROCm/llvm-project/tree/amd-staging/openmp>`_,18.0.0.24355,18.0.0.24232,17.0.0.24193,17.0.0.24193,17.0.0.24154,17.0.0.24103,17.0.0.24012,17.0.0.23483
`clang-ocl <https://github.com/ROCm/clang-ocl>`_,N/A,N/A,N/A,0.5.0,0.5.0,0.5.0,0.5.0,0.5.0
:doc:`hipCC <hipcc:index>`,1.1.1,1.1.1,1.1.1,1.0.0,1.0.0,1.0.0,1.0.0,1.0.0
`Flang <https://github.com/ROCm/flang>`_,18.0.0.24355,18.0.0.24355,18.0.0.24232,17.0.0.24193,17.0.0.24154,17.0.0.24103,17.0.0.24012,17.0.0.23483
:doc:`llvm-project <llvm-project:index>`,18.0.0.24355,18.0.0.24355,18.0.0.24232,17.0.0.24193,17.0.0.24154,17.0.0.24103,17.0.0.24012,17.0.0.23483
`OpenMP <https://github.com/ROCm/llvm-project/tree/amd-staging/openmp>`_,18.0.0.24355,18.0.0.24355,18.0.0.24232,17.0.0.24193,17.0.0.24154,17.0.0.24103,17.0.0.24012,17.0.0.23483
,,,,,,,,
RUNTIMES,.. _runtime-support-compatibility-matrix-past-60:,,,,,,,
:doc:`AMD CLR <hip:understand/amd_clr>`,6.2.41134,6.2.41133,6.1.40093,6.1.40093,6.1.40092,6.1.40091,6.1.32831,6.1.32830
:doc:`HIP <hip:index>`,6.2.41134,6.2.41133,6.1.40093,6.1.40093,6.1.40092,6.1.40091,6.1.32831,6.1.32830
:doc:`AMD CLR <hip:understand/amd_clr>`,6.2.41134,6.2.41134,6.2.41133,6.1.40093,6.1.40092,6.1.40091,6.1.32831,6.1.32830
:doc:`HIP <hip:index>`,6.2.41134,6.2.41134,6.2.41133,6.1.40093,6.1.40092,6.1.40091,6.1.32831,6.1.32830
`OpenCL Runtime <https://github.com/ROCm/clr/tree/develop/opencl>`_,2.0.0,2.0.0,2.0.0,2.0.0,2.0.0,2.0.0,2.0.0,2.0.0
:doc:`ROCR-Runtime <rocr-runtime:index>`,1.14.0,1.13.0,1.13.0,1.13.0,1.13.0,1.13.0,1.12.0,1.12.0
:doc:`ROCR-Runtime <rocr-runtime:index>`,1.14.0,1.14.0,1.13.0,1.13.0,1.13.0,1.13.0,1.12.0,1.12.0
1 ROCm Version 6.2.2 6.2.1 6.1.5 6.2.0 6.1.2 6.1.1 6.1.0 6.0.2 6.0.0
2 :ref:`Operating Systems & kernels <OS-kernel-versions>` :ref:`Operating systems & kernels <OS-kernel-versions>` Ubuntu 24.04.1, 24.04 Ubuntu 24.04.1, 24.04 Ubuntu 24.04
3 Ubuntu 22.04.5, 22.04.4 Ubuntu 22.04.5, 22.04.4 Ubuntu 22.04.5, 22.04.4, 22.04.3 Ubuntu 22.04.5, 22.04.4 Ubuntu 22.04.5, 22.04.4, 22.04.3 Ubuntu 22.04.4, 22.04.3 Ubuntu 22.04.4, 22.04.3 Ubuntu 22.04.4, 22.04.3 Ubuntu 22.04.4, 22.04.3, 22.04.2 Ubuntu 22.04.4, 22.04.3, 22.04.2
4 Ubuntu 20.04.6, 20.04.5 Ubuntu 20.04.6, 20.04.5 Ubuntu 20.04.6, 20.04.5 Ubuntu 20.04.6, 20.04.5 Ubuntu 20.04.6, 20.04.5 Ubuntu 20.04.6, 20.04.5
5 RHEL 9.4, 9.3 RHEL 9.4, 9.3 RHEL 9.4 [#red-hat94-past-60]_, 9.3, 9.2 RHEL 9.4, 9.3 RHEL 9.4 [#red-hat94-past-60]_, 9.3, 9.2 RHEL 9.4 [#red-hat94-past-60]_, 9.3, 9.2 RHEL 9.4 [#red-hat94-past-60]_, 9.3, 9.2 RHEL 9.3, 9.2 RHEL 9.3, 9.2
6 RHEL 8.10, 8.9 RHEL 8.10, 8.9 RHEL 8.9, 8.8 RHEL 8.10, 8.9 RHEL 8.9, 8.8 RHEL 8.9, 8.8 RHEL 8.9, 8.8 RHEL 8.9, 8.8 RHEL 8.9, 8.8
7 SLES 15 SP6, SP5 SLES 15 SP6, SP5 SLES 15 SP5, SP4 SLES 15 SP6, SP5 SLES 15 SP5, SP4 SLES 15 SP5, SP4 SLES 15 SP5, SP4 SLES 15 SP5, SP4 SLES 15 SP5, SP4
8 CentOS 7.9 CentOS 7.9 CentOS 7.9 CentOS 7.9 CentOS 7.9
9 Oracle Linux 8.9 [#oracle89-past-60]_ Oracle Linux 8.9 [#oracle89-past-60]_ Oracle Linux 8.9 [#oracle89-past-60]_ Oracle Linux 8.9 [#oracle89-past-60]_ Oracle Linux 8.9 [#oracle89-past-60]_ Oracle Linux 8.9 [#oracle89-past-60]_
10 .. _architecture-support-compatibility-matrix-past-60: .. _architecture-support-compatibility-matrix-past-60:
16 .. _gpu-support-compatibility-matrix-past-60: .. _gpu-support-compatibility-matrix-past-60:
17 :doc:`GPU / LLVM target <rocm-install-on-linux:reference/system-requirements>` gfx1100 gfx1100 gfx1100 gfx1100 gfx1100 gfx1100 gfx1100 gfx1100 gfx1100
18 gfx1030 gfx1030 gfx1030 gfx1030 gfx1030 gfx1030 gfx1030 gfx1030 gfx1030
19 gfx942 [#mi300_622-past-60]_ gfx942 [#mi300_621-past-60]_ gfx942 [#mi300_612-past-60]_ gfx942 [#mi300_620-past-60]_ gfx942 [#mi300_612-past-60]_ gfx942 [#mi300_611-past-60]_ gfx942 [#mi300_610-past-60]_ gfx942 [#mi300_602-past-60]_ gfx942 [#mi300_600-past-60]_
20 gfx90a gfx90a gfx90a gfx90a gfx90a gfx90a gfx90a gfx90a gfx90a
21 gfx908 gfx908 gfx908 gfx908 gfx908 gfx908 gfx908 gfx908 gfx908
22
23 FRAMEWORK SUPPORT .. _framework-support-compatibility-matrix-past-60: .. _framework-support-compatibility-matrix-past-60:
24 :doc:`PyTorch <rocm-install-on-linux:install/3rd-party/pytorch-install>` 2.3, 2.2, 2.1, 2.0, 1.13 2.3, 2.2, 2.1, 2.0, 1.13 2.1, 2.0, 1.13 2.3, 2.2, 2.1, 2.0, 1.13 2.1, 2.0, 1.13 2.1, 2.0, 1.13 2.1, 2.0, 1.13 2.1, 2.0, 1.13 2.1, 2.0, 1.13
25 :doc:`TensorFlow <rocm-install-on-linux:install/3rd-party/tensorflow-install>` 2.16.1, 2.15.1, 2.14.1 2.16.1, 2.15.1, 2.14.1 2.15.0, 2.14.0, 2.13.1 2.16.1, 2.15.1, 2.14.1 2.15.0, 2.14.0, 2.13.1 2.15.0, 2.14.0, 2.13.1 2.15.0, 2.14.0, 2.13.1 2.14.0, 2.13.1, 2.12.1 2.14.0, 2.13.1, 2.12.1
26 :doc:`JAX <rocm-install-on-linux:install/3rd-party/jax-install>` 0.4.26 0.4.26 0.4.26 0.4.26 0.4.26 0.4.26 0.4.26 0.4.26 0.4.26
27 `ONNX Runtime <https://onnxruntime.ai/docs/build/eps.html#amd-migraphx>`_ 1.17.3 1.17.3 1.17.3 1.17.3 1.17.3 1.17.3 1.17.3 1.14.1 1.14.1
28
29 THIRD PARTY COMMS .. _thirdpartycomms-support-compatibility-matrix-past-60: .. _thirdpartycomms-support-compatibility-matrix-past-60:
30 `UCC <https://github.com/ROCm/ucc>`_ >=1.3.0 >=1.3.0 >=1.3.0 >=1.3.0 >=1.3.0 >=1.3.0 >=1.3.0 >=1.2.0 >=1.2.0
31 `UCX <https://github.com/ROCm/ucx>`_ >=1.15.0 >=1.15.0 >=1.14.1 >=1.15.0 >=1.14.1 >=1.14.1 >=1.14.1 >=1.14.1 >=1.14.1
32
33 THIRD PARTY ALGORITHM .. _thirdpartyalgorithm-support-compatibility-matrix-past-60: .. _thirdpartyalgorithm-support-compatibility-matrix-past-60:
34 Thrust 2.2.0 2.2.0 2.1.0 2.2.0 2.1.0 2.1.0 2.1.0 2.0.1 2.0.1
35 CUB 2.2.0 2.2.0 2.1.0 2.2.0 2.1.0 2.1.0 2.1.0 2.0.1 2.0.1
36
37 KFD & USER SPACE [#kfd_support-past-60]_ .. _kfd-userspace-support-compatibility-matrix-past-60: .. _kfd-userspace-support-compatibility-matrix-past-60:
38 Tested user space versions 6.1.x, 6.0.x 6.1.x, 6.0.x 6.2.x, 6.0.x, 5.7.x 6.1.x, 6.0.x 6.2.x, 6.0.x, 5.7.x 6.2.x, 6.0.x, 5.7.x 6.2.x, 6.0.x, 5.7.x 6.2.x, 6.0.x, 5.7.x, 5.6.x 6.2.x, 6.0.x, 5.7.x, 5.6.x
39
40 ML & COMPUTER VISION .. _mllibs-support-compatibility-matrix-past-60: .. _mllibs-support-compatibility-matrix-past-60:
41 :doc:`Composable Kernel <composable_kernel:index>` 1.1.0 1.1.0 1.1.0 1.1.0 1.1.0 1.1.0 1.1.0 1.1.0 1.1.0
42 :doc:`MIGraphX <amdmigraphx:index>` 2.10.0 2.10.0 2.9.0 2.10.0 2.9.0 2.9.0 2.9.0 2.8.0 2.8.0
43 :doc:`MIOpen <miopen:index>` 3.2.0 3.2.0 3.1.0 3.2.0 3.1.0 3.1.0 3.1.0 3.0.0 3.0.0
44 :doc:`MIVisionX <mivisionx:index>` 3.0.0 3.0.0 2.5.0 3.0.0 2.5.0 2.5.0 2.5.0 2.5.0 2.5.0
45 :doc:`rocAL <rocal:index>` 2.0.0 2.0.0 1.0.0 1.0.0 1.0.0 1.0.0 1.0.0 1.0.0 1.0.0
46 :doc:`rocDecode <rocdecode:index>` 0.6.0 0.6.0 0.6.0 0.6.0 0.6.0 0.5.0 0.5.0 N/A N/A
47 :doc:`rocPyDecode <rocpydecode:index>` 0.1.0 0.1.0 N/A 0.1.0 N/A N/A N/A N/A N/A
48 :doc:`RPP <rpp:index>` 1.8.0 1.8.0 1.5.0 1.8.0 1.5.0 1.5.0 1.5.0 1.4.0 1.4.0
49
50 COMMUNICATION .. _commlibs-support-compatibility-matrix-past-60: .. _commlibs-support-compatibility-matrix-past-60:
51 :doc:`RCCL <rccl:index>` 2.20.5 2.20.5 2.18.6 2.20.5 2.18.6 2.18.6 2.18.6 2.18.3 2.18.3
52
53 MATH LIBS .. _mathlibs-support-compatibility-matrix-past-60: .. _mathlibs-support-compatibility-matrix-past-60:
54 `half <https://github.com/ROCm/half>`_ 1.12.0 1.12.0 1.12.0 1.12.0 1.12.0 1.12.0 1.12.0 1.12.0 1.12.0
55 :doc:`hipBLAS <hipblas:index>` 2.2.0 2.2.0 2.1.0 2.2.0 2.1.0 2.1.0 2.1.0 2.0.0 2.0.0
56 :doc:`hipBLASLt <hipblaslt:index>` 0.8.0 0.8.0 0.7.0 0.8.0 0.7.0 0.7.0 0.7.0 0.6.0 0.6.0
57 :doc:`hipFFT <hipfft:index>` 1.0.15 1.0.15 1.0.14 1.0.14 1.0.14 1.0.14 1.0.14 1.0.13 1.0.13
58 :doc:`hipFORT <hipfort:index>` 0.4.0 0.4.0 0.4.0 0.4.0 0.4.0 0.4.0 0.4.0 0.4.0 0.4.0
59 :doc:`hipRAND <hiprand:index>` 2.11.0 2.11.0 2.10.16 2.11.0 2.10.16 2.10.16 2.10.16 2.10.16 2.10.16
60 :doc:`hipSOLVER <hipsolver:index>` 2.2.0 2.2.0 2.1.1 2.2.0 2.1.1 2.1.1 2.1.0 2.0.0 2.0.0
61 :doc:`hipSPARSE <hipsparse:index>` 3.1.1 3.1.1 3.0.1 3.1.1 3.0.1 3.0.1 3.0.1 3.0.0 3.0.0
62 :doc:`hipSPARSELt <hipsparselt:index>` 0.2.1 0.2.1 0.2.0 0.2.1 0.2.0 0.1.0 0.1.0 0.1.0 0.1.0
63 :doc:`rocALUTION <rocalution:index>` 3.2.0 3.2.0 3.1.1 3.2.0 3.1.1 3.1.1 3.1.1 3.0.3 3.0.3
64 :doc:`rocBLAS <rocblas:index>` 4.2.1 4.2.1 4.1.2 4.2.0 4.1.2 4.1.0 4.1.0 4.0.0 4.0.0
65 :doc:`rocFFT <rocfft:index>` 1.0.29 1.0.29 1.0.27 1.0.28 1.0.27 1.0.27 1.0.26 1.0.25 1.0.23
66 :doc:`rocRAND <rocrand:index>` 3.1.0 3.1.0 3.0.1 3.1.0 3.0.1 3.0.1 3.0.1 3.0.0 2.10.17
67 :doc:`rocSOLVER <rocsolver:index>` 3.26.0 3.26.0 3.25.0 3.26.0 3.25.0 3.25.0 3.25.0 3.24.0 3.24.0
68 :doc:`rocSPARSE <rocsparse:index>` 3.2.0 3.2.0 3.1.2 3.2.0 3.1.2 3.1.2 3.1.2 3.0.2 3.0.2
69 :doc:`rocWMMA <rocwmma:index>` 1.5.0 1.5.0 1.4.0 1.5.0 1.4.0 1.4.0 1.4.0 1.3.0 1.3.0
70 `Tensile <https://github.com/ROCm/Tensile>`_ 4.40.0 4.40.0 4.40.0 4.40.0 4.40.0 4.40.0 4.40.0 4.39.0 4.39.0
71
72 PRIMITIVES .. _primitivelibs-support-compatibility-matrix-past-60: .. _primitivelibs-support-compatibility-matrix-past-60:
73 :doc:`hipCUB <hipcub:index>` 3.2.0 3.2.0 3.1.0 3.2.0 3.1.0 3.1.0 3.1.0 3.0.0 3.0.0
74 :doc:`hipTensor <hiptensor:index>` 1.3.0 1.3.0 1.2.0 1.3.0 1.2.0 1.2.0 1.2.0 1.1.0 1.1.0
75 :doc:`rocPRIM <rocprim:index>` 3.2.0 3.2.0 3.1.0 3.2.0 3.1.0 3.1.0 3.1.0 3.0.0 3.0.0
76 :doc:`rocThrust <rocthrust:index>` 3.1.0 3.1.0 3.0.1 3.0.1 3.0.1 3.0.1 3.0.1 3.0.0 3.0.0
77
78 SUPPORT LIBS
79 `hipother <https://github.com/ROCm/hipother>`_ 6.2.41134 6.2.41134 6.1.40093 6.2.41133 6.1.40093 6.1.40092 6.1.40091 6.1.32831 6.1.32830
80 `rocm-core <https://github.com/ROCm/rocm-core>`_ 6.2.2 6.2.1 6.1.5 6.2.0 6.1.2 6.1.1 6.1.0 6.0.2 6.0.0
81 `ROCT-Thunk-Interface <https://github.com/ROCm/ROCT-Thunk-Interface>`_ 20240607.4.05 20240607.4.05 20240125.5.08 20240607.1.4246 20240125.5.08 20240125.5.08 20240125.3.30 20231016.2.245 20231016.2.245
82
83 SYSTEM MGMT TOOLS .. _tools-support-compatibility-matrix-past-60: .. _tools-support-compatibility-matrix-past-60:
84 :doc:`AMD SMI <amdsmi:index>` 24.6.3 24.6.3 24.5.1 24.6.2 24.5.1 24.5.1 24.4.1 23.4.2 23.4.2
85 :doc:`ROCm Data Center Tool <rdc:index>` 1.0.0 0.3.0 1.0.0 0.3.0 0.3.0 1.0.0 0.3.0 0.3.0 0.3.0 0.3.0 0.3.0
86 :doc:`rocminfo <rocminfo:index>` 1.0.0 1.0.0 1.0.0 1.0.0 1.0.0 1.0.0 1.0.0 1.0.0 1.0.0
87 :doc:`ROCm SMI <rocm_smi_lib:index>` 7.3.0 7.3.0 7.2.0 7.3.0 7.2.0 7.0.0 7.0.0 6.0.2 6.0.0
88 :doc:`ROCm Validation Suite <rocmvalidationsuite:index>` rocm-6.2.2 rocm-6.2.1 rocm-6.1.5 rocm-6.2.0 rocm-6.1.2 rocm-6.1.1 rocm-6.1.0 rocm-6.0.2 rocm-6.0.0
89
90 PERFORMANCE TOOLS
91 :doc:`Omniperf <omniperf:index>` 2.0.1 2.0.1 N/A 2.0.1 N/A N/A N/A N/A N/A
92 :doc:`Omnitrace <omnitrace:index>` 1.11.2 1.11.2 N/A 1.11.2 N/A N/A N/A N/A N/A
93 :doc:`ROCm Bandwidth Test <rocm_bandwidth_test:index>` 1.4.0 1.4.0 1.4.0 1.4.0 1.4.0 1.4.0 1.4.0 1.4.0 1.4.0
94 :doc:`ROCProfiler <rocprofiler:index>` 2.0.60202 2.0.60201 2.0.60105 2.0.60200 2.0.60102 2.0.60101 2.0.60100 2.0.60002 2.0.60000
95 :doc:`ROCprofiler-SDK <rocprofiler-sdk:index>` 0.4.0 0.4.0 N/A 0.4.0 N/A N/A N/A N/A N/A
96 :doc:`ROCTracer <roctracer:index>` 4.1.60202 4.1.60201 4.1.60105 4.1.60200 4.1.60102 4.1.60101 4.1.60100 4.1.60002 4.1.60000
97
98 DEVELOPMENT TOOLS
99 :doc:`HIPIFY <hipify:index>` 18.0.0.24355 18.0.0.24355 17.0.0.24193 18.0.0.24232 17.0.0.24193 17.0.0.24154 17.0.0.24103 17.0.0.24012 17.0.0.23483
100 :doc:`ROCm CMake <rocmcmakebuildtools:index>` 0.13.0 0.13.0 0.12.0 0.13.0 0.12.0 0.12.0 0.12.0 0.11.0 0.11.0
101 :doc:`ROCdbgapi <rocdbgapi:index>` 0.76.0 0.76.0 0.71.0 0.76.0 0.71.0 0.71.0 0.71.0 0.71.0 0.71.0
102 :doc:`ROCm Debugger (ROCgdb) <rocgdb:index>` 14.2.0 14.2.0 14.1.0 14.2.0 14.1.0 14.1.0 14.1.0 13.2.0 13.2.0
103 `rocprofiler-register <https://github.com/ROCm/rocprofiler-register>`_ 0.4.0 0.4.0 0.3.0 0.4.0 0.3.0 0.3.0 0.3.0 N/A N/A
104 :doc:`ROCr Debug Agent <rocr_debug_agent:index>` 2.0.3 2.0.3 2.0.3 2.0.3 2.0.3 2.0.3 2.0.3 2.0.3 2.0.3
105
106 COMPILERS .. _compilers-support-compatibility-matrix-past-60: .. _compilers-support-compatibility-matrix-past-60:
107 `clang-ocl <https://github.com/ROCm/clang-ocl>`_ N/A N/A 0.5.0 N/A 0.5.0 0.5.0 0.5.0 0.5.0 0.5.0
108 :doc:`hipCC <hipcc:index>` 1.1.1 1.1.1 1.0.0 1.1.1 1.0.0 1.0.0 1.0.0 1.0.0 1.0.0
109 `Flang <https://github.com/ROCm/flang>`_ 18.0.0.24355 18.0.0.24355 17.0.0.24193 18.0.0.24232 17.0.0.24193 17.0.0.24154 17.0.0.24103 17.0.0.24012 17.0.0.23483
110 :doc:`llvm-project <llvm-project:index>` 18.0.0.24355 18.0.0.24355 17.0.0.24193 18.0.0.24232 17.0.0.24193 17.0.0.24154 17.0.0.24103 17.0.0.24012 17.0.0.23483
111 `OpenMP <https://github.com/ROCm/llvm-project/tree/amd-staging/openmp>`_ 18.0.0.24355 18.0.0.24355 17.0.0.24193 18.0.0.24232 17.0.0.24193 17.0.0.24154 17.0.0.24103 17.0.0.24012 17.0.0.23483
112
113 RUNTIMES .. _runtime-support-compatibility-matrix-past-60: .. _runtime-support-compatibility-matrix-past-60:
114 :doc:`AMD CLR <hip:understand/amd_clr>` 6.2.41134 6.2.41134 6.1.40093 6.2.41133 6.1.40093 6.1.40092 6.1.40091 6.1.32831 6.1.32830
115 :doc:`HIP <hip:index>` 6.2.41134 6.2.41134 6.1.40093 6.2.41133 6.1.40093 6.1.40092 6.1.40091 6.1.32831 6.1.32830
116 `OpenCL Runtime <https://github.com/ROCm/clr/tree/develop/opencl>`_ 2.0.0 2.0.0 2.0.0 2.0.0 2.0.0 2.0.0 2.0.0 2.0.0 2.0.0
117 :doc:`ROCR-Runtime <rocr-runtime:index>` 1.14.0 1.14.0 1.13.0 1.13.0 1.13.0 1.13.0 1.13.0 1.12.0 1.12.0

View File

@@ -17,10 +17,10 @@ You can also refer to the :ref:`past versions of ROCm compatibility matrix<past-
.. container:: format-big-table
.. csv-table::
:header: "ROCm Version", "6.2.1", "6.2.0", "6.1.0"
:header: "ROCm Version", "6.2.2", "6.2.1", "6.1.0"
:stub-columns: 1
:ref:`Operating Systems & kernels <OS-kernel-versions>`,"Ubuntu 24.04.1, 24.04",Ubuntu 24.04,
:ref:`Operating systems & kernels <OS-kernel-versions>`,"Ubuntu 24.04.1, 24.04","Ubuntu 24.04.1, 24.04",Ubuntu 24.04
,"Ubuntu 22.04.5, 22.04.4","Ubuntu 22.04.5, 22.04.4","Ubuntu 22.04.4, 22.04.3"
,,,"Ubuntu 20.04.6, 20.04.5"
,"RHEL 9.4, 9.3","RHEL 9.4, 9.3","RHEL 9.4 [#red-hat94]_, 9.3, 9.2"
@@ -37,7 +37,7 @@ You can also refer to the :ref:`past versions of ROCm compatibility matrix<past-
,.. _gpu-support-compatibility-matrix:,,
:doc:`GPU / LLVM target <rocm-install-on-linux:reference/system-requirements>`,gfx1100,gfx1100,gfx1100
,gfx1030,gfx1030,gfx1030
,gfx942 [#mi300_621]_,gfx942 [#mi300_620]_, gfx942 [#mi300_610]_
,gfx942 [#mi300_622]_,gfx942 [#mi300_621]_, gfx942 [#mi300_610]_
,gfx90a,gfx90a,gfx90a
,gfx908,gfx908,gfx908
,,,
@@ -63,7 +63,7 @@ You can also refer to the :ref:`past versions of ROCm compatibility matrix<past-
:doc:`MIGraphX <amdmigraphx:index>`,2.10.0,2.10.0,2.9.0
:doc:`MIOpen <miopen:index>`,3.2.0,3.2.0,3.1.0
:doc:`MIVisionX <mivisionx:index>`,3.0.0,3.0.0,2.5.0
:doc:`rocAL <rocal:index>`,2.0.0,1.0.0,1.0.0
:doc:`rocAL <rocal:index>`,2.0.0,2.0.0,1.0.0
:doc:`rocDecode <rocdecode:index>`,0.6.0,0.6.0,0.5.0
:doc:`rocPyDecode <rocpydecode:index>`,0.1.0,0.1.0,N/A
:doc:`RPP <rpp:index>`,1.8.0,1.8.0,1.5.0
@@ -75,15 +75,15 @@ You can also refer to the :ref:`past versions of ROCm compatibility matrix<past-
`half <https://github.com/ROCm/half>`_ ,1.12.0,1.12.0,1.12.0
:doc:`hipBLAS <hipblas:index>`,2.2.0,2.2.0,2.1.0
:doc:`hipBLASLt <hipblaslt:index>`,0.8.0,0.8.0,0.7.0
:doc:`hipFFT <hipfft:index>`,1.0.15,1.0.14,1.0.14
:doc:`hipFFT <hipfft:index>`,1.0.15,1.0.15,1.0.14
:doc:`hipFORT <hipfort:index>`,0.4.0,0.4.0,0.4.0
:doc:`hipRAND <hiprand:index>`,2.11.0,2.11.0,2.10.16
:doc:`hipSOLVER <hipsolver:index>`,2.2.0,2.2.0,2.1.0
:doc:`hipSPARSE <hipsparse:index>`,3.1.1,3.1.1,3.0.1
:doc:`hipSPARSELt <hipsparselt:index>`,0.2.1,0.2.1,0.1.0
:doc:`rocALUTION <rocalution:index>`,3.2.0,3.2.0,3.1.1
:doc:`rocBLAS <rocblas:index>`,4.2.1,4.2.0,4.1.0
:doc:`rocFFT <rocfft:index>`,1.0.29,1.0.28,1.0.26
:doc:`rocBLAS <rocblas:index>`,4.2.1,4.2.1,4.1.0
:doc:`rocFFT <rocfft:index>`,1.0.29,1.0.29,1.0.26
:doc:`rocRAND <rocrand:index>`,3.1.0,3.1.0,3.0.1
:doc:`rocSOLVER <rocsolver:index>`,3.26.0,3.26.0,3.25.0
:doc:`rocSPARSE <rocsparse:index>`,3.2.0,3.2.0,3.1.2
@@ -94,30 +94,30 @@ You can also refer to the :ref:`past versions of ROCm compatibility matrix<past-
:doc:`hipCUB <hipcub:index>`,3.2.0,3.2.0,3.1.0
:doc:`hipTensor <hiptensor:index>`,1.3.0,1.3.0,1.2.0
:doc:`rocPRIM <rocprim:index>`,3.2.0,3.2.0,3.1.0
:doc:`rocThrust <rocthrust:index>`,3.1.0,3.0.1,3.0.1
:doc:`rocThrust <rocthrust:index>`,3.1.0,3.1.0,3.0.1
,,,
SUPPORT LIBS,,,
`hipother <https://github.com/ROCm/hipother>`_,6.2.41134,6.2.41133,6.1.40091
`rocm-core <https://github.com/ROCm/rocm-core>`_,6.2.1,6.2.0,6.1.0
`ROCT-Thunk-Interface <https://github.com/ROCm/ROCT-Thunk-Interface>`_,20240607.4.05,20240607.1.4246,20240125.3.30
`hipother <https://github.com/ROCm/hipother>`_,6.2.41134,6.2.41134,6.1.40091
`rocm-core <https://github.com/ROCm/rocm-core>`_,6.2.2,6.2.1,6.1.0
`ROCT-Thunk-Interface <https://github.com/ROCm/ROCT-Thunk-Interface>`_,20240607.4.05,20240607.4.05,20240125.3.30
,,,
SYSTEM MGMT TOOLS,.. _tools-support-compatibility-matrix:,,
:doc:`AMD SMI <amdsmi:index>`,24.6.3,24.6.2,24.4.1
:doc:`ROCm Data Center Tool <rdc:index>`,0.3.0,0.3.0,0.3.0
:doc:`AMD SMI <amdsmi:index>`,24.6.3,24.6.3,24.4.1
:doc:`ROCm Data Center Tool <rdc:index>`,1.0.0,1.0.0,0.3.0
:doc:`rocminfo <rocminfo:index>`,1.0.0,1.0.0,1.0.0
:doc:`ROCm SMI <rocm_smi_lib:index>`,7.3.0,7.3.0,7.0.0
:doc:`ROCm Validation Suite <rocmvalidationsuite:index>`,rocm-6.2.1,rocm-6.2.0,rocm-6.1.0
:doc:`ROCm Validation Suite <rocmvalidationsuite:index>`,rocm-6.2.2,rocm-6.2.1,rocm-6.1.0
,,,
PERFORMANCE TOOLS,,,
:doc:`Omniperf <omniperf:index>`,2.0.1,2.0.1,N/A
:doc:`Omnitrace <omnitrace:index>`,1.11.2,1.11.2,N/A
:doc:`ROCm Bandwidth Test <rocm_bandwidth_test:index>`,1.4.0,1.4.0,1.4.0
:doc:`ROCProfiler <rocprofiler:index>`,2.0.60201,2.0.60200,2.0.60100
:doc:`ROCProfiler <rocprofiler:index>`,2.0.60202,2.0.60201,2.0.60100
:doc:`ROCprofiler-SDK <rocprofiler-sdk:index>`,0.4.0,0.4.0,N/A
:doc:`ROCTracer <roctracer:index>`,4.1.60201,4.1.60200,4.1.60100
:doc:`ROCTracer <roctracer:index>`,4.1.60202,4.1.60201,4.1.60100
,,,
DEVELOPMENT TOOLS,,,
:doc:`HIPIFY <hipify:index>`,18.0.0.24355,18.0.0.24232,17.0.0.24103
:doc:`HIPIFY <hipify:index>`,18.0.0.24355,18.0.0.24355,17.0.0.24103
:doc:`ROCm CMake <rocmcmakebuildtools:index>`,0.13.0,0.13.0,0.12.0
:doc:`ROCdbgapi <rocdbgapi:index>`,0.76.0,0.76.0,0.71.0
:doc:`ROCm Debugger (ROCgdb) <rocgdb:index>`,14.2.0,14.2.0,14.1.0
@@ -127,25 +127,25 @@ You can also refer to the :ref:`past versions of ROCm compatibility matrix<past-
COMPILERS,.. _compilers-support-compatibility-matrix:,,
`clang-ocl <https://github.com/ROCm/clang-ocl>`_,N/A,N/A,0.5.0
:doc:`hipCC <hipcc:index>`,1.1.1,1.1.1,1.0.0
`Flang <https://github.com/ROCm/flang>`_,18.0.0.24355,18.0.0.24232,17.0.0.24103
:doc:`llvm-project <llvm-project:index>`,18.0.0.24355,18.0.0.24232,17.0.0.24103
`OpenMP <https://github.com/ROCm/llvm-project/tree/amd-staging/openmp>`_,18.0.0.24355,18.0.0.24232,17.0.0.24103
`Flang <https://github.com/ROCm/flang>`_,18.0.0.24355,18.0.0.24355,17.0.0.24103
:doc:`llvm-project <llvm-project:index>`,18.0.0.24355,18.0.0.24355,17.0.0.24103
`OpenMP <https://github.com/ROCm/llvm-project/tree/amd-staging/openmp>`_,18.0.0.24355,18.0.0.24355,17.0.0.24103
,,,
RUNTIMES,.. _runtime-support-compatibility-matrix:,,
:doc:`AMD CLR <hip:understand/amd_clr>`,6.2.41134,6.2.41133,6.1.40091
:doc:`HIP <hip:index>`,6.2.41134,6.2.41133,6.1.40091
:doc:`AMD CLR <hip:understand/amd_clr>`,6.2.41134,6.2.41134,6.1.40091
:doc:`HIP <hip:index>`,6.2.41134,6.2.41134,6.1.40091
`OpenCL Runtime <https://github.com/ROCm/clr/tree/develop/opencl>`_,2.0.0,2.0.0,2.0.0
:doc:`ROCR-Runtime <rocr-runtime:index>`,1.14.0,1.13.0,1.13.0
:doc:`ROCR-Runtime <rocr-runtime:index>`,1.14.0,1.14.0,1.13.0
.. rubric:: Footnotes
.. [#red-hat94] **For ROCm 6.1** - RHEL 9.4 is supported only on AMD Instinct MI300A.
.. [#red-hat94] RHEL 9.4 is supported only on AMD Instinct MI300A.
.. [#oracle89] Oracle Linux is supported only on AMD Instinct MI300X.
.. [#mi300_622] **For ROCm 6.2.2** - MI300X (gfx942) is supported on listed operating systems *except* Ubuntu 22.04.5 [6.8 HWE] and Ubuntu 22.04.4 [6.5 HWE].
.. [#mi300_621] **For ROCm 6.2.1** - MI300X (gfx942) is supported on listed operating systems *except* Ubuntu 22.04.5 [6.8 HWE] and Ubuntu 22.04.4 [6.5 HWE].
.. [#mi300_620] **For ROCm 6.2.0** - MI300X (gfx942) is supported on listed operating systems *except* Ubuntu 22.04.5 [6.8 HWE] and Ubuntu 22.04.4 [6.5 HWE].
.. [#mi300_610] **For ROCm 6.1.0** - MI300A (gfx942) is supported on Ubuntu 22.04.4, RHEL 9.4, RHEL 9.3, RHEL 8.9, and SLES 15 SP5. MI300X (gfx942) is only supported on Ubuntu 22.04.4.
.. [#kfd_support] ROCm provides forward and backward compatibility between the Kernel Fusion Driver (KFD) and its user space software for +/- 2 releases. The tested user space versions on this page were accurate as of the time of initial ROCm release. For the most up-to-date information, see the latest version of this information at `User and kernel-space support matrix <https://rocm.docs.amd.com/projects/install-on-linux/en/latest/reference/user-kernel-space-compat-matrix.html>`_.
.. [#kfd_support] ROCm provides forward and backward compatibility between the Kernel Fusion Driver (KFD) and its user space software for +/- 2 releases. These are the compatibility combinations that are currently supported.
.. _OS-kernel-versions:
@@ -207,7 +207,6 @@ Expand for full historical view of:
.. csv-table::
:file: compatibility-matrix-historical-6.0.csv
:widths: 20,10,10,10,10,10,10,10,10
:header-rows: 1
:stub-columns: 1
@@ -215,6 +214,7 @@ Expand for full historical view of:
.. [#red-hat94-past-60] RHEL 9.4 is supported only on AMD Instinct MI300A.
.. [#oracle89-past-60] Oracle Linux is supported only on AMD Instinct MI300X.
.. [#mi300_622-past-60] **For ROCm 6.2.2** - MI300X (gfx942) is supported on listed operating systems *except* Ubuntu 22.04.5 [6.8 HWE] and Ubuntu 22.04.4 [6.5 HWE].
.. [#mi300_621-past-60] **For ROCm 6.2.1** - MI300X (gfx942) is supported on listed operating systems *except* Ubuntu 22.04.5 [6.8 HWE] and Ubuntu 22.04.4 [6.5 HWE].
.. [#mi300_620-past-60] **For ROCm 6.2.0** - MI300X (gfx942) is supported on listed operating systems *except* Ubuntu 22.04.5 [6.8 HWE] and Ubuntu 22.04.4 [6.5 HWE].
.. [#mi300_612-past-60] **For ROCm 6.1.2** - MI300A (gfx942) is supported on Ubuntu 22.04.4, RHEL 9.4, RHEL 9.3, RHEL 8.9, and SLES 15 SP5. MI300X (gfx942) is only supported on Ubuntu 22.04.4 and Oracle Linux.
@@ -222,4 +222,4 @@ Expand for full historical view of:
.. [#mi300_610-past-60] **For ROCm 6.1.0** - MI300A (gfx942) is supported on Ubuntu 22.04.4, RHEL 9.4, RHEL 9.3, RHEL 8.9, and SLES 15 SP5. MI300X (gfx942) is only supported on Ubuntu 22.04.4.
.. [#mi300_602-past-60] **For ROCm 6.0.2** - MI300A (gfx942) is supported on Ubuntu 22.04.3, RHEL 8.9, and SLES 15 SP5. MI300X (gfx942) is only supported on Ubuntu 22.04.3.
.. [#mi300_600-past-60] **For ROCm 6.0.0** - MI300A (gfx942) is supported on Ubuntu 22.04.3, RHEL 8.9, and SLES 15 SP5. MI300X (gfx942) is only supported on Ubuntu 22.04.3.
.. [#kfd_support-past-60] ROCm provides forward and backward compatibility between the Kernel Fusion Driver (KFD) and its user space software for +/- 2 releases. The tested user space versions on this page were accurate as of the time of initial ROCm release. For the most up-to-date information, see the latest version of this information at `User and kernel-space support matrix <https://rocm.docs.amd.com/projects/install-on-linux/en/latest/reference/user-kernel-space-compat-matrix.html>`_.
.. [#kfd_support-past-60] ROCm provides forward and backward compatibility between the Kernel Fusion Driver (KFD) and its user space software for +/- 2 releases. These are the compatibility combinations that are currently supported.

View File

@@ -30,16 +30,15 @@ if os.environ.get("READTHEDOCS", "") == "True":
project = "ROCm Documentation"
author = "Advanced Micro Devices, Inc."
copyright = "Copyright (c) 2024 Advanced Micro Devices, Inc. All rights reserved."
version = "6.2.1"
release = "6.2.1"
version = "6.2.2"
release = "6.2.2"
setting_all_article_info = True
all_article_info_os = ["linux", "windows"]
all_article_info_author = ""
# pages with specific settings
article_pages = [
{"file": "about/release-notes", "os": ["linux", "windows"], "date": "2024-09-20"},
{"file": "about/changelog", "os": ["linux", "windows"], "date": "2024-09-20"},
{"file": "about/release-notes", "os": ["linux", "windows"], "date": "2024-09-27"},
{"file": "how-to/deep-learning-rocm", "os": ["linux"]},
{"file": "how-to/rocm-for-ai/index", "os": ["linux"]},
{"file": "how-to/rocm-for-ai/install", "os": ["linux"]},
@@ -82,7 +81,6 @@ article_pages = [
"file": "how-to/llm-fine-tuning-optimization/profiling-and-debugging",
"os": ["linux"],
},
{"file": "how-to/performance-validation/mi300x/vllm-benchmark", "os": ["linux"]},
{"file": "how-to/system-optimization/index", "os": ["linux"]},
{"file": "how-to/system-optimization/mi300x", "os": ["linux"]},
{"file": "how-to/system-optimization/mi200", "os": ["linux"]},

View File

@@ -16,7 +16,7 @@ This section discusses how to implement `vLLM <https://docs.vllm.ai/en/latest>`_
vLLM inference
==============
vLLM is renowned for its PagedAttention algorithm that can reduce memory consumption and increase throughput thanks to
vLLM is renowned for its paged attention algorithm that can reduce memory consumption and increase throughput thanks to
its paging scheme. Instead of allocating GPU high-bandwidth memory (HBM) for the maximum output token lengths of the
models, the paged attention of vLLM allocates GPU HBM dynamically for its actual decoding lengths. This paged attention
is also effective when multiple requests share the same key and value contents for a large value of beam search or
@@ -139,7 +139,9 @@ Refer to :ref:`mi300x-vllm-optimization` for performance optimization tips.
ROCm provides a prebuilt optimized Docker image for validating the performance of LLM inference with vLLM
on the MI300X accelerator. The Docker image includes ROCm, vLLM, PyTorch, and tuning files in the CSV
format. For more information, see :doc:`/how-to/performance-validation/mi300x/vllm-benchmark`.
format. For more information, see the guide to
`LLM inference performance validation with vLLM on the AMD Instinct™ MI300X accelerator <https://github.com/ROCm/MAD/blob/develop/benchmark/vllm/README.md>`_
on the ROCm GitHub repository.
.. _fine-tuning-llms-tgi:

View File

@@ -1,407 +0,0 @@
.. meta::
:description: Learn how to validate LLM inference performance on MI300X accelerators using AMD MAD and the unified
ROCm Docker image.
:keywords: model, MAD, automation, dashboarding, validate
***********************************************************
LLM inference performance validation on AMD Instinct MI300X
***********************************************************
.. _vllm-benchmark-unified-docker:
The `ROCm vLLM Docker <https://hub.docker.com/r/rocm/vllm/tags>`_ image offers
a prebuilt, optimized environment designed for validating large language model
(LLM) inference performance on the AMD Instinct™ MI300X accelerator. This
ROCm vLLM Docker image integrates vLLM and PyTorch tailored specifically for the
MI300X accelerator and includes the following components:
* `ROCm 6.2.1 <https://github.com/ROCm/ROCm>`_
* `vLLM 0.6.4 <https://docs.vllm.ai/en/latest>`_
* `PyTorch 2.5.0 <https://github.com/pytorch/pytorch>`_
* Tuning files (in CSV format)
With this Docker image, you can quickly validate the expected inference
performance numbers on the MI300X accelerator. This topic also provides tips on
optimizing performance with popular AI models.
.. hlist::
:columns: 6
* Llama 3.1 8B
* Llama 3.1 70B
* Llama 3.1 405B
* Llama 2 7B
* Llama 2 70B
* Mixtral 8x7B
* Mixtral 8x22B
* Mixtral 7B
* Qwen2 7B
* Qwen2 72B
* JAIS 13B
* JAIS 30B
.. _vllm-benchmark-vllm:
.. note::
vLLM is a toolkit and library for LLM inference and serving. AMD implements
high-performance custom kernels and modules in vLLM to enhance performance.
See :ref:`fine-tuning-llms-vllm` and :ref:`mi300x-vllm-optimization` for
more information.
Getting started
===============
Use the following procedures to reproduce the benchmark results on an
MI300X accelerator with the prebuilt vLLM Docker image.
.. _vllm-benchmark-get-started:
1. Disable NUMA auto-balancing.
To optimize performance, disable automatic NUMA balancing. Otherwise, the GPU
might hang until the periodic balancing is finalized. For more information,
see :ref:`AMD Instinct MI300X system optimization <mi300x-disable-numa>`.
.. code-block:: shell
# disable automatic NUMA balancing
sh -c 'echo 0 > /proc/sys/kernel/numa_balancing'
# check if NUMA balancing is disabled (returns 0 if disabled)
cat /proc/sys/kernel/numa_balancing
0
2. Download the :ref:`ROCm vLLM Docker image <vllm-benchmark-unified-docker>`.
Use the following command to pull the Docker image from Docker Hub.
.. code-block:: shell
docker pull rocm/vllm:rocm6.2_mi300_ubuntu20.04_py3.9_vllm_0.6.4
Once setup is complete, you can choose between two options to reproduce the
benchmark results:
- :ref:`MAD-integrated benchmarking <vllm-benchmark-mad>`
- :ref:`Standalone benchmarking <vllm-benchmark-standalone>`
.. _vllm-benchmark-mad:
MAD-integrated benchmarking
===========================
Clone the ROCm Model Automation and Dashboarding (`<https://github.com/ROCm/MAD>`__) repository to a local
directory and install the required packages on the host machine.
.. code-block:: shell
git clone https://github.com/ROCm/MAD
cd MAD
pip install -r requirements.txt
Use this command to run a performance benchmark test of the Llama 3.1 8B model
on one GPU with ``float16`` data type in the host machine.
.. code-block:: shell
export MAD_SECRETS_HFTOKEN="your personal Hugging Face token to access gated models"
python3 tools/run_models.py --tags pyt_vllm_llama-3.1-8b --keep-model-dir --live-output --timeout 28800
ROCm MAD launches a Docker container with the name
``container_ci-pyt_vllm_llama-3.1-8b``. The latency and throughput reports of the
model are collected in the following path: ``~/MAD/reports_float16/``.
Although the following models are preconfigured to collect latency and
throughput performance data, you can also change the benchmarking parameters.
Refer to the :ref:`Standalone benchmarking <vllm-benchmark-standalone>` section.
Available models
----------------
.. hlist::
:columns: 3
* ``pyt_vllm_llama-3.1-8b``
* ``pyt_vllm_llama-3.1-70b``
* ``pyt_vllm_llama-3.1-405b``
* ``pyt_vllm_llama-2-7b``
* ``pyt_vllm_llama-2-70b``
* ``pyt_vllm_mixtral-8x7b``
* ``pyt_vllm_mixtral-8x22b``
* ``pyt_vllm_mistral-7b``
* ``pyt_vllm_qwen2-7b``
* ``pyt_vllm_qwen2-72b``
* ``pyt_vllm_jais-13b``
* ``pyt_vllm_jais-30b``
* ``pyt_vllm_llama-3.1-8b_fp8``
* ``pyt_vllm_llama-3.1-70b_fp8``
* ``pyt_vllm_llama-3.1-405b_fp8``
* ``pyt_vllm_mixtral-8x7b_fp8``
* ``pyt_vllm_mixtral-8x22b_fp8``
.. _vllm-benchmark-standalone:
Standalone benchmarking
=======================
You can run the vLLM benchmark tool independently by starting the
:ref:`Docker container <vllm-benchmark-get-started>` as shown in the following
snippet.
.. code-block::
docker pull rocm/vllm:rocm6.2_mi300_ubuntu20.04_py3.9_vllm_0.6.4
docker run -it --device=/dev/kfd --device=/dev/dri --group-add video --shm-size 128G --security-opt seccomp=unconfined --security-opt apparmor=unconfined --cap-add=SYS_PTRACE -v $(pwd):/workspace --env HUGGINGFACE_HUB_CACHE=/workspace --name vllm_v0.6.4 rocm/vllm:rocm6.2_mi300_ubuntu20.04_py3.9_vllm_0.6.4
In the Docker container, clone the ROCm MAD repository and navigate to the
benchmark scripts directory at ``~/MAD/scripts/vllm``.
.. code-block::
git clone https://github.com/ROCm/MAD
cd MAD/scripts/vllm
Command
-------
To start the benchmark, use the following command with the appropriate options.
See :ref:`Options <vllm-benchmark-standalone-options>` for the list of
options and their descriptions.
.. code-block:: shell
./vllm_benchmark_report.sh -s $test_option -m $model_repo -g $num_gpu -d $datatype
See the :ref:`examples <vllm-benchmark-run-benchmark>` for more information.
.. note::
The input sequence length, output sequence length, and tensor parallel (TP) are
already configured. You don't need to specify them with this script.
.. note::
If you encounter the following error, pass your access-authorized Hugging
Face token to the gated models.
.. code-block:: shell
OSError: You are trying to access a gated repo.
# pass your HF_TOKEN
export HF_TOKEN=$your_personal_hf_token
.. _vllm-benchmark-standalone-options:
Options
-------
.. list-table::
:header-rows: 1
:align: center
* - Name
- Options
- Description
* - ``$test_option``
- latency
- Measure decoding token latency
* -
- throughput
- Measure token generation throughput
* -
- all
- Measure both throughput and latency
* - ``$model_repo``
- ``meta-llama/Meta-Llama-3.1-8B-Instruct``
- Llama 3.1 8B
* - (``float16``)
- ``meta-llama/Meta-Llama-3.1-70B-Instruct``
- Llama 3.1 70B
* -
- ``meta-llama/Meta-Llama-3.1-405B-Instruct``
- Llama 3.1 405B
* -
- ``meta-llama/Llama-2-7b-chat-hf``
- Llama 2 7B
* -
- ``meta-llama/Llama-2-70b-chat-hf``
- Llama 2 70B
* -
- ``mistralai/Mixtral-8x7B-Instruct-v0.1``
- Mixtral 8x7B
* -
- ``mistralai/Mixtral-8x22B-Instruct-v0.1``
- Mixtral 8x22B
* -
- ``mistralai/Mistral-7B-Instruct-v0.3``
- Mixtral 7B
* -
- ``Qwen/Qwen2-7B-Instruct``
- Qwen2 7B
* -
- ``Qwen/Qwen2-72B-Instruct``
- Qwen2 72B
* -
- ``core42/jais-13b-chat``
- JAIS 13B
* -
- ``core42/jais-30b-chat-v3``
- JAIS 30B
* - ``$model_repo``
- ``amd/Meta-Llama-3.1-8B-Instruct-FP8-KV``
- Llama 3.1 8B
* - (``float8``)
- ``amd/Meta-Llama-3.1-70B-Instruct-FP8-KV``
- Llama 3.1 70B
* -
- ``amd/Meta-Llama-3.1-405B-Instruct-FP8-KV``
- Llama 3.1 405B
* -
- ``amd/Mixtral-8x7B-Instruct-v0.1-FP8-KV``
- Mixtral 8x7B
* -
- ``amd/Mixtral-8x22B-Instruct-v0.1-FP8-KV``
- Mixtral 8x22B
* - ``$num_gpu``
- 1 or 8
- Number of GPUs
* - ``$datatype``
- ``float16`` or ``float8``
- Data type
.. _vllm-benchmark-run-benchmark:
Running the benchmark on the MI300X accelerator
-----------------------------------------------
Here are some examples of running the benchmark with various options.
See :ref:`Options <vllm-benchmark-standalone-options>` for the list of
options and their descriptions.
Example 1: latency benchmark
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Use this command to benchmark the latency of the Llama 3.1 8B model on one GPU with the ``float16`` and ``float8`` data types.
.. code-block::
./vllm_benchmark_report.sh -s latency -m meta-llama/Meta-Llama-3.1-8B-Instruct -g 1 -d float16
./vllm_benchmark_report.sh -s latency -m amd/Meta-Llama-3.1-8B-Instruct-FP8-KV -g 1 -d float8
Find the latency reports at:
- ``./reports_float16/summary/Meta-Llama-3.1-8B-Instruct_latency_report.csv``
- ``./reports_float8/summary/Meta-Llama-3.1-8B-Instruct-FP8-KV_latency_report.csv``
Example 2: throughput benchmark
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Use this command to benchmark the throughput of the Llama 3.1 8B model on one GPU with the ``float16`` and ``float8`` data types.
.. code-block:: shell
./vllm_benchmark_report.sh -s throughput -m meta-llama/Meta-Llama-3.1-8B-Instruct -g 1 -d float16
./vllm_benchmark_report.sh -s throughput -m amd/Meta-Llama-3.1-8B-Instruct-FP8-KV -g 1 -d float8
Find the throughput reports at:
- ``./reports_float16/summary/Meta-Llama-3.1-8B-Instruct_throughput_report.csv``
- ``./reports_float8/summary/Meta-Llama-3.1-8B-Instruct-FP8-KV_throughput_report.csv``
.. raw:: html
<style>
mjx-container[jax="CHTML"][display="true"] {
text-align: left;
margin: 0;
}
</style>
.. note::
Throughput is calculated as:
- .. math:: throughput\_tot = requests \times (\mathsf{\text{input lengths}} + \mathsf{\text{output lengths}}) / elapsed\_time
- .. math:: throughput\_gen = requests \times \mathsf{\text{output lengths}} / elapsed\_time
Further reading
===============
- For application performance optimization strategies for HPC and AI workloads,
including inference with vLLM, see :doc:`/how-to/tuning-guides/mi300x/workload`.
- To learn more about the options for latency and throughput benchmark scripts,
see `<https://github.com/ROCm/vllm/tree/main/benchmarks>`_.
- To learn more about system settings and management practices to configure your system for
MI300X accelerators, see :doc:`/how-to/system-optimization/mi300x`.
- To learn how to run LLM models from Hugging Face or your own model, see
:doc:`Using ROCm for AI </how-to/rocm-for-ai/index>`.
- To learn how to optimize inference on LLMs, see
:doc:`Fine-tuning LLMs and inference optimization </how-to/llm-fine-tuning-optimization/index>`.
- For a list of other ready-made Docker images for ROCm, see the
:doc:`Docker image support matrix <rocm-install-on-linux:reference/docker-image-support-matrix>`.
- To compare with the previous version of the ROCm vLLM Docker image for performance validation, refer to
`LLM inference performance validation on AMD Instinct MI300X (ROCm 6.2.0) <https://rocm.docs.amd.com/en/docs-6.2.0/how-to/performance-validation/mi300x/vllm-benchmark.html>`_.

View File

@@ -46,7 +46,9 @@ Validating vLLM performance
ROCm provides a prebuilt optimized Docker image for validating the performance of LLM inference with vLLM
on the MI300X accelerator. The Docker image includes ROCm, vLLM, PyTorch, and tuning files in the CSV
format. For more information, see :doc:`/how-to/performance-validation/mi300x/vllm-benchmark`.
format. For more information, see the guide to
`LLM inference performance validation with vLLM on the AMD Instinct™ MI300X accelerator <https://github.com/ROCm/MAD/blob/develop/benchmark/vllm/README.md>`_
on the ROCm GitHub repository.
.. _rocm-for-ai-serve-hugging-face-tgi:

View File

@@ -342,8 +342,8 @@ If SMT is enabled by setting "CCD/Core/Thread Enablement > SMT Control" to
(logical) cores of the system:
* In the server BIOS, set IOMMU to "Enabled".
* When configuring the Grub boot loader, add the following argument for the
Linux kernel: `iommu=pt`
* When configuring the Grub boot loader, add the following arguments for the
Linux kernel: `amd_iommu=on iommu=pt`
* Update Grub to use the modified configuration:
```shell
@@ -355,7 +355,7 @@ If SMT is enabled by setting "CCD/Core/Thread Enablement > SMT Control" to
```none
[...]
[ 0.000000] Kernel command line: [...] iommu=pt
[ 0.000000] Kernel command line: [...] amd_iommu=on iommu=pt
[...]
```

View File

@@ -327,8 +327,8 @@ If SMT is enabled by setting "CCD/Core/Thread Enablement > SMT Control" to
(logical) cores of the system:
* In the server BIOS, set IOMMU to "Enabled".
* When configuring the Grub boot loader, add the following argument for the
Linux kernel: `iommu=pt`
* When configuring the Grub boot loader, add the following arguments for the
Linux kernel: `amd_iommu=on iommu=pt`
* Update Grub to use the modified configuration:
```shell
@@ -340,7 +340,7 @@ If SMT is enabled by setting "CCD/Core/Thread Enablement > SMT Control" to
```none
[...]
[ 0.000000] Kernel command line: [...] iommu=pt
[ 0.000000] Kernel command line: [...] amd_iommu=on iommu=pt
[...]
```

View File

@@ -299,7 +299,7 @@ For a system that has AMD host CPUs add this to ``GRUB_CMDLINE_LINUX``:
.. code-block:: text
iommu=pt
amd_iommu=on iommu=pt
Otherwise, if the system has Intel host CPUs add this instead to
``GRUB_CMDLINE_LINUX``:
@@ -500,7 +500,7 @@ If SMT is enabled by setting ``CCD/Core/Thread Enablement > SMT Control`` to
#. In the server BIOS, set IOMMU to ``Enabled``.
#. When configuring the GRUB boot loader, add the following argument for the Linux kernel: ``iommu=pt``.
#. When configuring the GRUB boot loader, add the following arguments for the Linux kernel: ``amd_iommu=on iommu=pt``.
#. Update GRUB.
@@ -515,7 +515,7 @@ If SMT is enabled by setting ``CCD/Core/Thread Enablement > SMT Control`` to
.. code-block:: shell
[...]
[ 0.000000] Kernel command line: [...] iommu=pt
[ 0.000000] Kernel command line: [...] amd_iommu=on iommu=pt
[...]
Once the system is properly configured, ROCm software can be

View File

@@ -111,7 +111,7 @@ sudo virsh net-start default /*to enable Virtual network by default
Enable input-output memory management unit (IOMMU) in GRUB settings by adding the following line to `/etc/default/grub`:
```none
GRUB_CMDLINE_LINUX_DEFAULT="quiet splash" for AMD CPU
GRUB_CMDLINE_LINUX_DEFAULT="quiet splash amd_iommu=on" for AMD CPU
```
Update grub and reboot

View File

@@ -8,8 +8,6 @@ accelerators. They include detailed instructions on system settings and
application tuning suggestions to help you fully leverage the capabilities of
these accelerators, thereby achieving optimal performance.
* :doc:`/how-to/performance-validation/mi300x/vllm-benchmark`
* :doc:`/how-to/tuning-guides/mi300x/system`
* :doc:`/how-to/tuning-guides/mi300x/workload`

View File

@@ -152,7 +152,9 @@ address any new bottlenecks that may emerge.
ROCm provides a prebuilt optimized Docker image that has everything required to implement
the tips in this section. It includes ROCm, vLLM, PyTorch, and tuning files in the CSV
format. For more information, see :doc:`/how-to/performance-validation/mi300x/vllm-benchmark`.
format. For more information, see the guide to
`LLM inference performance validation with vLLM on the AMD Instinct™ MI300X accelerator <https://github.com/ROCm/MAD/blob/develop/benchmark/vllm/README.md>`_
on the ROCm GitHub repository.
.. _mi300x-profiling-tools:
@@ -376,10 +378,11 @@ Refer to `vLLM documentation <https://docs.vllm.ai/en/latest/models/performance.
for additional performance tips. :ref:`fine-tuning-llms-vllm` describes vLLM
usage with ROCm.
ROCm provides a prebuilt optimized Docker image for validating the performance
of LLM inference with vLLM on the MI300X accelerator. The Docker image includes
ROCm, vLLM, PyTorch, and tuning files in the CSV format. For more information,
see :doc:`/how-to/performance-validation/mi300x/vllm-benchmark`.
ROCm provides a prebuilt optimized Docker image for validating the performance of LLM inference with vLLM
on the MI300X accelerator. The Docker image includes ROCm, vLLM, PyTorch, and tuning files in the CSV
format. For more information, see the guide to
`LLM inference performance validation with vLLM on the AMD Instinct™ MI300X accelerator <https://github.com/ROCm/MAD/blob/develop/benchmark/vllm/README.md>`_
on the ROCm GitHub repository.
Maximize throughput
-------------------

View File

@@ -45,8 +45,7 @@ ROCm documentation is organized into the following categories:
* [Using ROCm for HPC](./how-to/rocm-for-hpc/index.rst)
* [Fine-tuning LLMs and inference optimization](./how-to/llm-fine-tuning-optimization/index.rst)
* [System optimization](./how-to/system-optimization/index.rst)
* [AMD Instinct MI300X performance validation and tuning](./how-to/tuning-guides/mi300x/index.rst)
* [GPU cluster networking](https://dcgpu.docs.amd.com/projects/gpu-cluster-networking/en/latest/index.html)
* [AMD Instinct MI300X tuning guides](./how-to/tuning-guides/mi300x/index.rst)
* [System debugging](./how-to/system-debugging.md)
* [Using MPI](./how-to/gpu-enabled-mpi.rst)
* [Using advanced compiler features](./conceptual/compiler-topics.md)

View File

@@ -8,6 +8,7 @@
| Version | Release date |
| ------- | ------------ |
| [6.2.2](https://rocm.docs.amd.com/en/docs-6.2.2/) | September 27, 2024 |
| [6.2.1](https://rocm.docs.amd.com/en/docs-6.2.1/) | September 20, 2024 |
| [6.2.0](https://rocm.docs.amd.com/en/docs-6.2.0/) | August 2, 2024 |
| [6.1.2](https://rocm.docs.amd.com/en/docs-6.1.2/) | June 4, 2024 |

View File

@@ -70,11 +70,9 @@ subtrees:
- file: how-to/system-optimization/w6000-v620.md
title: AMD RDNA 2
- file: how-to/tuning-guides/mi300x/index.rst
title: AMD MI300X performance validation and tuning
title: AMD MI300X tuning guides
subtrees:
- entries:
- file: how-to/performance-validation/mi300x/vllm-benchmark.rst
title: Performance validation
- file: how-to/tuning-guides/mi300x/system.rst
title: System tuning
- file: how-to/tuning-guides/mi300x/workload.rst

View File

@@ -1,7 +1,7 @@
<?xml version="1.0" encoding="UTF-8"?>
<manifest>
<remote name="rocm-org" fetch="https://github.com/ROCm/" />
<default revision="refs/tags/rocm-6.1.1"
<default revision="refs/tags/rocm-6.2.2"
remote="rocm-org"
sync-c="true"
sync-j="4" />

View File

@@ -184,5 +184,5 @@ clinfo, and HelloWord.cl and cause a system crash.
* IRQ remapping does not support X2APIC mode
* NMI error
Workaround: To avoid the system crash, add `iommu=pt` as the kernel bootparam, as
Workaround: To avoid the system crash, add `amd_iommu=on iommu=pt` as the kernel bootparam, as
indicated in the warning message.

View File

@@ -1,3 +1,28 @@
The release notes provide a comprehensive summary of changes since the previous ROCm release.
- [Release highlights](release-highlights)
- [Operating system and hardware support changes](operating-system-and-hardware-support-changes)
- [ROCm components versioning](rocm-components)
- [Detailed component changes](detailed-component-changes)
- [ROCm known issues](rocm-known-issues)
- [ROCm upcoming changes](rocm-upcoming-changes)
The [Compatibility matrix](https://rocm.docs.amd.com/en/docs-6.2.0/compatibility/compatibility-matrix.html)
provides an overview of operating system, hardware, ecosystem, and ROCm component support across ROCm releases.
Release notes for previous ROCm releases are available in earlier versions of the documentation.
See the [ROCm documentation release history](https://rocm.docs.amd.com/en/latest/release/versions.html).
## Release highlights
This section introduces notable new features and improvements in ROCm 6.2. See the
[Detailed component changes](#detailed-component-changes) for individual component changes.
### New components
ROCm 6.2.0 introduces the following new components to the ROCm software stack.

View File

@@ -1,6 +1,30 @@
The release notes provide a summary of notable changes since the previous ROCm release.
- [Release highlights](release-highlights)
- [Operating system and hardware support changes](operating-system-and-hardware-support-changes)
- [ROCm components versioning](rocm-components)
- [Detailed component changes](detailed-component-changes)
- [ROCm known issues](rocm-known-issues)
- [ROCm upcoming changes](rocm-upcoming-changes)
The [Compatibility matrix](https://rocm.docs.amd.com/en/docs-6.2.1/compatibility/compatibility-matrix.html)
provides the full list of supported hardware, operating systems, ecosystems, third-party components, and ROCm components for each ROCm release.
Release notes for previous ROCm releases are available in earlier versions of the documentation.
See the [ROCm documentation release history](https://rocm.docs.amd.com/en/latest/release/versions.html).
## Release highlights
The following are notable new features and improvements in ROCm 6.2.1. For changes to individual components, see [Detailed component changes](#detailed-component-changes).
### rocAL major version change
The new version of rocAL introduces many new features, but does not modify any of the existing public API functions.However, the version number was incremented from 1.3 to 2.0.
The new version of rocAL introduces many new features, but does not modify any of the existing public API functions. However, the version number was incremented from 1.3 to 2.0.
Applications linked to version 1.3 must be recompiled to link against version 2.0.
See [the rocAL detailed changes](#rocal-2-0-0) for more information.
@@ -9,12 +33,12 @@ See [the rocAL detailed changes](#rocal-2-0-0) for more information.
As of ROCm 6.2.1, ROCm supports Facebook General Matrix Multiplication (FBGEMM) and the related FBGEMM_GPU library.
FBGEMM is a low-precision, high-performance CPU kernel library for convolution and matrix multiplication. It is used for server-side inference and as a back end for PyTorch quantized operators. FBGEMM_GPU includes a collection of PyTorch GPU operator libraries for training and inference. For more information, see the ROCm [Model acceleration libraries guide](https://rocm.docs.amd.com/en/6.2.1/how-to/llm-fine-tuning-optimization/model-acceleration-libraries.html)
FBGEMM is a low-precision, high-performance CPU kernel library for convolution and matrix multiplication. It is used for server-side inference and as a back end for PyTorch quantized operators. FBGEMM_GPU includes a collection of PyTorch GPU operator libraries for training and inference. For more information, see the ROCm [Model acceleration libraries guide](https://rocm.docs.amd.com/en/docs-6.2.1/how-to/llm-fine-tuning-optimization/model-acceleration-libraries.html)
and [PyTorch's FBGEMM GitHub repository](https://github.com/pytorch/FBGEMM).
### ROCm Offline Installer Creator changes
The [ROCm Offline Installer Creator 6.2.1](https://rocm.docs.amd.com/projects/install-on-linux/en/6.2.1/install/rocm-offline-installer.html) introduces several new features and improvements including:
The [ROCm Offline Installer Creator 6.2.1](https://rocm.docs.amd.com/projects/install-on-linux/en/docs-6.2.1/install/rocm-offline-installer.html) introduces several new features and improvements including:
* Logging support for create and install logs
* More stringent checks for Linux versions and distributions
@@ -26,8 +50,8 @@ The [ROCm Offline Installer Creator 6.2.1](https://rocm.docs.amd.com/projects/in
There have been no changes to supported hardware or operating systems from ROCm 6.2.0 to ROCm 6.2.1.
* The Programming Model Reference and Understanding the Programming Model topics in HIP have been consolidated into one topic,
[HIP programming model (conceptual)](https://rocm.docs.amd.com/projects/HIP/en/6.2.1/understand/programming_model.html).
* The [HIP virtual memory management](https://rocm.docs.amd.com/projects/HIP/en/6.2.1/how-to/virtual_memory.html) and [HIP virtual memory management API](https://rocm.docs.amd.com/projects/HIP/en/6.2.1/reference/virtual_memory_reference.html) topics have been added.
[HIP programming model (conceptual)](https://rocm.docs.amd.com/projects/HIP/en/docs-6.2.1/understand/programming_model.html).
* The [HIP virtual memory management](https://rocm.docs.amd.com/projects/HIP/en/docs-6.2.1/how-to/virtual_memory.html) and [HIP virtual memory management API](https://rocm.docs.amd.com/projects/HIP/en/docs-6.2.1/reference/virtual_memory_reference.html) topics have been added.
```{note}
The ROCm documentation, like all ROCm projects, is open source and available on GitHub. To contribute to ROCm documentation, see the [ROCm documentation contribution guidelines](https://rocm.docs.amd.com/en/latest/contribute/contributing.html).

View File

@@ -0,0 +1,31 @@
These release notes provide a summary of notable changes since the previous ROCm release.
```{note}
As ROCm 6.2.2 was released shortly after 6.2.1, the changes between these versions
are minimal. For a comprehensive overview of recent updates, the ROCm 6.2.1 release
notes are appended to the end of this document.
For detailed information about the changes in ROCm 6.2.1, refer to the appended
section: [ROCm 6.2.1 release notes](rocm-6-2-1-release-notes).
```
The [Compatibility matrix](https://rocm.docs.amd.com/en/docs-6.2.2/compatibility/compatibility-matrix.html)
provides the full list of supported hardware, operating systems, ecosystems, third-party components, and ROCm components
for each ROCm release.
Release notes for previous ROCm releases are available in earlier versions of the documentation.
See the [ROCm documentation release history](https://rocm.docs.amd.com/en/latest/release/versions.html).
## Release highlights
The following is a significant fix introduced in ROCm 6.2.2.
### Fixed Instinct MI300X error recovery failure
Improved the reliability of AMD Instinct MI300X accelerators in scenarios involving
uncorrectable errors. Previously, error recovery did not occur as expected,
potentially leaving the system in an undefined state. This fix ensures that error
recovery functions as expected, maintaining system stability.
See the [original issue](#instinct-mi300x-gpu-recovery-failure-on-uncorrectable-errors)
noted in the ROCm 6.2.1 release notes.

View File

@@ -0,0 +1,86 @@
## ROCm known issues
ROCm known issues are noted on {fab}`github` [GitHub](https://github.com/ROCm/ROCm/labels/Verified%20Issue). For known
issues related to individual components, review the [Detailed component changes](detailed-component-changes).
### Default processor affinity behavior for helper threads
Processor affinity is a critical setting to ensure that ROCm helper threads run on the correct cores. By default, ROCm
helper threads are spawned on all available cores, ignoring the parent threads processor affinity. This can lead to
threads competing for available cores, which may result in suboptimal performance. This behavior occurs by default if
the environment variable `HSA_OVERRIDE_CPU_AFFINITY_DEBUG` is not set or is set to `1`. If
`HSA_OVERRIDE_CPU_AFFINITY_DEBUG` is set to `0`, the ROCr runtime uses the parent process's core affinity mask when
creating helper threads. The parents affinity mask should then be set to account for the presence of additional threads
by ensuring the affinity mask contains enough cores. Depending on the affinity settings of the software environment,
batch system, launch commands like `numactl`/`taskset`, or explicit mask manipulation by the application itself, changing
the setting may be advantageous to performance.
To ensure the parent's core affinity mask is honored by the ROCm helper threads, set the
`HSA_OVERRIDE_CPU_AFFINITY_DEBUG` environment variable as follows:
```{code} shell
export HSA_OVERRIDE_CPU_AFFINITY_DEBUG=0
```
To ensure ROCm helper threads run on all available cores, set the `HSA_OVERRIDE_CPU_AFFINITY_DEBUG` environment variable
as follows:
``` shell
export HSA_OVERRIDE_CPU_AFFINITY_DEBUG=1
```
Or the default:
``` shell
unset HSA_OVERRIDE_CPU_AFFINITY_DEBUG
```
If unsure of the default processor affinity settings for your environment, run the following command from the shell:
``` shell
bash -c "echo taskset -p \$\$"
```
See [issue #3493](https://github.com/ROCm/ROCm/issues/3493) on GitHub.
### Display issues on servers with Instinct MI300-series accelerators when loading AMDGPU driver
AMD Instinct MI300-series accelerators and third-party GPUs such as the Matrox G200 have an issue impacting video
output. The issue was reproduced on a Dell server model PowerEdge XE9680. Servers from other vendors utilizing Matrox
G200 cards may be impacted as well. This issue was found with ROCm 6.2.0 but is present in older ROCm versions.
The AMDGPU driver shipped with ROCm interferes with the operation of the display card video output. On Dell systems,
this includes both the local video output and remote access via iDRAC. The display appears blank (black) after loading
the `amdgpu` driver modules. Video output impacts both terminal access when running in `runlevel 3` and GUI access when
running in `runlevel 5`. Server functionality can still be accessed via SSH or other remote connection methods.
See [issue #3494](https://github.com/ROCm/ROCm/issues/3494) on GitHub.
### KFDTest failure on Instinct MI300X with Oracle Linux 8.9
The `KFDEvictTest.QueueTest` is failing on the MI300X platform during KFD (Kernel Fusion Driver) tests, causing the full
suite to not execute properly. This issue is suspected to be hardware-related.
See [issue #3495](https://github.com/ROCm/ROCm/issues/3495) on GitHub.
### Bandwidth limitation in gang and non-gang modes on Instinct MI300A
Expected target peak non-gang performance (~60GB/s) and target peak gang performance (~90GB/s) are not achieved. Both gang
and non-gang performance are observed to be limited at 45GB/s.
This issue will be addressed in a future ROCm release.
See [issue #3496](https://github.com/ROCm/ROCm/issues/3496) on GitHub.
### rocm-llvm-alt
ROCm provides an optional package -- `rocm-llvm-alt` -- that provides a closed-source compiler for
users interested in additional closed-source CPU optimizations. This feature is not functional in
the ROCm 6.2.0 release. Users who attempt to invoke the closed-source compiler will experience an
LLVM consumer-producer mismatch and the compilation will fail. There is no workaround that allows
use of the closed-source compiler. It is recommended to compile using the default open-source
compiler, which generates high-quality AMD CPU and AMD GPU code.
See [issue #3492](https://github.com/ROCm/ROCm/issues/3492) on GitHub.

View File

@@ -1,3 +1,8 @@
## ROCm known issues
ROCm known issues are tracked on [GitHub](https://github.com/ROCm/ROCm/labels/Verified%20Issue). Known issues related to
individual components are listed in the [Detailed component changes](detailed-component-changes) section.
### Instinct MI300X GPU recovery failure on uncorrectable errors
For the AMD Instinct MI300X accelerator, GPU recovery resets triggered by uncorrectable errors (UE) might not complete

View File

@@ -1,85 +1,3 @@
### Default processor affinity behavior for helper threads
Processor affinity is a critical setting to ensure that ROCm helper threads run on the correct cores. By default, ROCm
helper threads are spawned on all available cores, ignoring the parent threads processor affinity. This can lead to
threads competing for available cores, which may result in suboptimal performance. This behavior occurs by default if
the environment variable `HSA_OVERRIDE_CPU_AFFINITY_DEBUG` is not set or is set to `1`. If
`HSA_OVERRIDE_CPU_AFFINITY_DEBUG` is set to `0`, the ROCr runtime uses the parent process's core affinity mask when
creating helper threads. The parents affinity mask should then be set to account for the presence of additional threads
by ensuring the affinity mask contains enough cores. Depending on the affinity settings of the software environment,
batch system, launch commands like `numactl`/`taskset`, or explicit mask manipulation by the application itself, changing
the setting may be advantageous to performance.
To ensure the parent's core affinity mask is honored by the ROCm helper threads, set the
`HSA_OVERRIDE_CPU_AFFINITY_DEBUG` environment variable as follows:
```{code} shell
export HSA_OVERRIDE_CPU_AFFINITY_DEBUG=0
```
To ensure ROCm helper threads run on all available cores, set the `HSA_OVERRIDE_CPU_AFFINITY_DEBUG` environment variable
as follows:
``` shell
export HSA_OVERRIDE_CPU_AFFINITY_DEBUG=1
```
Or the default:
``` shell
unset HSA_OVERRIDE_CPU_AFFINITY_DEBUG
```
If unsure of the default processor affinity settings for your environment, run the following command from the shell:
``` shell
bash -c "echo taskset -p \$\$"
```
See [issue #3493](https://github.com/ROCm/ROCm/issues/3493) on GitHub.
### Display issues on servers with Instinct MI300-series accelerators when loading AMDGPU driver
AMD Instinct MI300-series accelerators and third-party GPUs such as the Matrox G200 have an issue impacting video
output. The issue was reproduced on a Dell server model PowerEdge XE9680. Servers from other vendors utilizing Matrox
G200 cards may be impacted as well. This issue was found with ROCm 6.2.0 but is present in older ROCm versions.
The AMDGPU driver shipped with ROCm interferes with the operation of the display card video output. On Dell systems,
this includes both the local video output and remote access via iDRAC. The display appears blank (black) after loading
the `amdgpu` driver modules. Video output impacts both terminal access when running in `runlevel 3` and GUI access when
running in `runlevel 5`. Server functionality can still be accessed via SSH or other remote connection methods.
See [issue #3494](https://github.com/ROCm/ROCm/issues/3494) on GitHub.
### KFDTest failure on Instinct MI300X with Oracle Linux 8.9
The `KFDEvictTest.QueueTest` is failing on the MI300X platform during KFD (Kernel Fusion Driver) tests, causing the full
suite to not execute properly. This issue is suspected to be hardware-related.
See [issue #3495](https://github.com/ROCm/ROCm/issues/3495) on GitHub.
### Bandwidth limitation in gang and non-gang modes on Instinct MI300A
Expected target peak non-gang performance (~60GB/s) and target peak gang performance (~90GB/s) are not achieved. Both gang
and non-gang performance are observed to be limited at 45GB/s.
This issue will be addressed in a future ROCm release.
See [issue #3496](https://github.com/ROCm/ROCm/issues/3496) on GitHub.
### rocm-llvm-alt
ROCm provides an optional package -- `rocm-llvm-alt` -- that provides a closed-source compiler for
users interested in additional closed-source CPU optimizations. This feature is not functional in
the ROCm 6.2.0 release. Users who attempt to invoke the closed-source compiler will experience an
LLVM consumer-producer mismatch and the compilation will fail. There is no workaround that allows
use of the closed-source compiler. It is recommended to compile using the default open-source
compiler, which generates high-quality AMD CPU and AMD GPU code.
See [issue #3492](https://github.com/ROCm/ROCm/issues/3492) on GitHub.
## ROCm upcoming changes
The section notes upcoming changes to the ROCm software stack. For upcoming changes related to individual components, review

View File

@@ -1,3 +1,7 @@
## ROCm upcoming changes
The following changes to the ROCm software stack are anticipated for future releases.
### rocm-llvm-alt
The `rocm-llvm-alt` package will be removed in an upcoming release. Users relying on the functionality provided by the closed-source compiler should transition to the open-source compiler. Once the `rocm-llvm-alt` package is removed, any compilation requesting functionality provided by the closed-source compiler will result in a Clang warning: "*[AMD] proprietary optimization compiler has been removed*".

View File

@@ -36,6 +36,7 @@ build_mivisionx() {
"$COMPONENT_SRC"
cmake --build "$BUILD_DIR" -- -j${PROC}
cmake --build "$BUILD_DIR" -- install
cpack -G ${PKGTYPE^^}
rm -rf _CPack_Packages/ && find -name '*.o' -delete

View File

@@ -20,9 +20,27 @@ build_rocal() {
set_address_sanitizer_on
fi
mkdir -p $BUILD_DIR && cd $BUILD_DIR
# python3 ${COMPONENT_SRC}/rocAL-setup.py
pushd /tmp
# PyBind11
git clone -b v2.11.1 https://github.com/pybind/pybind11
cd pybind11 && mkdir build && cd build
cmake -DDOWNLOAD_CATCH=ON -DDOWNLOAD_EIGEN=ON ../
make -j$(nproc) && sudo make install
cd ../..
# Turbo JPEG
git clone -b 3.0.2 https://github.com/libjpeg-turbo/libjpeg-turbo.git
cd libjpeg-turbo && mkdir build && cd build
cmake -DCMAKE_INSTALL_PREFIX=/usr -DCMAKE_BUILD_TYPE=RELEASE -DENABLE_STATIC=FALSE -DCMAKE_INSTALL_DEFAULT_LIBDIR=lib -DWITH_JPEG8=TRUE ..
make -j$(nproc) && sudo make install
cd ../..
# RapidJSON
git clone https://github.com/Tencent/rapidjson.git
cd rapidjson && mkdir build && cd build
cmake .. && make -j$(nproc) && sudo make install
popd
python3 ${COMPONENT_SRC}/rocAL-setup.py
mkdir -p $BUILD_DIR && cd $BUILD_DIR
cmake -DAMDRPP_PATH=$ROCM_PATH ${COMPONENT_SRC}
make -j${PROC}

View File

@@ -0,0 +1,74 @@
<?xml version="1.0" encoding="UTF-8"?>
<manifest>
<remote name="rocm-org" fetch="https://github.com/ROCm/" />
<default revision="refs/tags/rocm-6.2.1"
remote="rocm-org"
sync-c="true"
sync-j="4" />
<!--list of projects for ROCm-->
<project name="ROCm" revision="roc-6.2.x" />
<project name="ROCK-Kernel-Driver" />
<project name="ROCR-Runtime" />
<project name="ROCT-Thunk-Interface" />
<project name="amdsmi" />
<project name="omniperf" />
<project name="omnitrace" />
<project name="rdc" />
<project name="rocm_bandwidth_test" />
<project name="rocm_smi_lib" />
<project name="rocm-core" />
<project name="rocminfo" />
<project name="rocprofiler" />
<project name="rocprofiler-register" />
<project name="rocprofiler-sdk" />
<project name="roctracer" />
<!--HIP Projects-->
<project name="HIP" />
<project name="hip-tests" />
<project name="HIP-Examples" />
<project name="HIPIFY" />
<project name="clr" />
<project name="hipother" />
<!-- The following projects are all associated with the AMDGPU LLVM compiler -->
<project name="half" />
<project name="llvm-project" />
<!-- gdb projects -->
<project name="ROCdbgapi" />
<project name="ROCgdb" />
<project name="rocr_debug_agent" />
<!-- ROCm Libraries -->
<project groups="mathlibs" name="AMDMIGraphX" />
<project groups="mathlibs" name="MIOpen" />
<project groups="mathlibs" name="MIVisionX" />
<project groups="mathlibs" name="ROCmValidationSuite" />
<project groups="mathlibs" name="Tensile" />
<project groups="mathlibs" name="composable_kernel" />
<project groups="mathlibs" name="hipBLAS" />
<project groups="mathlibs" name="hipBLASLt" />
<project groups="mathlibs" name="hipCUB" />
<project groups="mathlibs" name="hipFFT" />
<project groups="mathlibs" name="hipRAND" />
<project groups="mathlibs" name="hipSOLVER" />
<project groups="mathlibs" name="hipSPARSE" />
<project groups="mathlibs" name="hipSPARSELt" />
<project groups="mathlibs" name="hipTensor" />
<project groups="mathlibs" name="hipfort" />
<project groups="mathlibs" name="rccl" />
<project groups="mathlibs" name="rocAL" />
<project groups="mathlibs" name="rocALUTION" />
<project groups="mathlibs" name="rocBLAS" />
<project groups="mathlibs" name="rocDecode" />
<project groups="mathlibs" name="rocFFT" />
<project groups="mathlibs" name="rocPRIM" />
<project groups="mathlibs" name="rocRAND" />
<project groups="mathlibs" name="rocSOLVER" />
<project groups="mathlibs" name="rocSPARSE" />
<project groups="mathlibs" name="rocThrust" />
<project groups="mathlibs" name="rocWMMA" />
<project groups="mathlibs" name="rocm-cmake" />
<project groups="mathlibs" name="rpp" />
<!-- Projects for OpenMP-Extras -->
<project name="aomp" path="openmp-extras/aomp" />
<project name="aomp-extras" path="openmp-extras/aomp-extras" />
<project name="flang" path="openmp-extras/flang" />
</manifest>