ROCm 6.3.1 Release
Release notes
The release notes provide a summary of notable changes since the previous ROCm release.
If you’re using RadeonPRO or Radeon GPUs in a workstation setting with a display connected, continue to use ROCm 6.2.3. See the [Use ROCm on Radeon GPUs](https://rocm.docs.amd.com/projects/radeon/en/latest/index.html) documentation to verify compatibility and system requirements.
Release highlights
The following are notable new features and improvements in ROCm 6.3.1. For changes to individual components, see
Detailed component changes.Per queue resiliency for Instinct MI300 accelerators
The AMDGPU driver now includes enhanced resiliency for misbehaving applications on AMD Instinct MI300 accelerators. This helps isolate the impact of misbehaving applications, ensuring other workloads running on the same accelerator are unaffected.
ROCm Runfile Installer
ROCm 6.3.1 introduces the ROCm Runfile Installer, with initial support for Ubuntu 22.04. The ROCm Runfile Installer facilitates ROCm installation without using a native Linux package management system, with or without network or internet access. For more information, see the ROCm Runfile Installer documentation.
ROCm documentation updates
ROCm documentation continues to be updated to provide clearer and more comprehensive guidance for a wider variety of user needs and use cases.
Added documentation on training a model with ROCm Megatron-LM. AMD offers a Docker image for MI300X accelerators
containing essential components to get started, including ROCm libraries, PyTorch, and Megatron-LM utilities. See
Training a model using ROCm Megatron-LMto get started.The new ROCm Megatron-LM training Docker accompanies the ROCm vLLM inference
Docker
as a set of ready-to-use containerized solutions to get started with using ROCm
for AI.Updated the Instinct MI300X workload tuning
guide with more current optimization
strategies. The updated sections include guidance on vLLM optimization, PyTorch TunableOp, and hipBLASLt tuning.HIP graph-safe libraries operate safely in HIP execution graphs. HIP graphs are an alternative way of executing tasks on a GPU that can provide performance benefits over launching kernels using the standard method via streams. A topic that shows whether a ROCm library is graph-safe has been added.
The Device memory topic in the HIP memory management section has been updated.
The HIP documentation has expanded with new resources for developers:
Operating system and hardware support changes
ROCm 6.3.1 adds support for Debian 12 (kernel: 6.1). Debian is supported only on AMD Instinct accelerators. See the installation instructions at Debian native installation.
ROCm 6.3.1 enables support for AMD Instinct MI325X accelerator. For more information, see AMD Instinct
See the Compatibility
matrixfor more information about operating system and hardware compatibility.ROCm components
The following table lists the versions of ROCm components for ROCm 6.3.1, including any version
changes from 6.3.0 to 6.3.1. Click the component's updated version to go to a list of its changes.
Click {fab}githubto go to the component's source code on GitHub.Detailed component changes
The following sections describe key changes to ROCm components.
AMD SMI (24.7.1)
Changed
amd-smi monitordisplaysVCLOCKandDCLOCKinstead ofENC_CLOCKandDEC_CLOCK.Resolved issues
- Fixed
amd-smi monitor's reporting of encode and decode information.VCLOCKandDCLOCKare
now associated with bothENC_UTILandDEC_UTIL.See the full [AMD SMI changelog](https://github.com/ROCm/amdsmi/blob/6.3.x/CHANGELOG.md) for more details and examples.HIP (6.3.1)
Added
- An activeQueues set that tracks only the queues that have a command submitted to them, which allows fast iteration in
waitActiveStreams.Resolved issues
- A deadlock in a specific customer application by preventing hipLaunchKernel latency degradation with number of idle streams.
HIPIFY (18.0.0)
Added
- Support for:
- NVIDIA CUDA 12.6.2
- cuDNN 9.5.1
- LLVM 19.1.3
- Full
hipBLAS64-bit APIs- Full
rocBLAS64-bit APIsResolved issues
- Added missing support for device intrinsics and built-ins:
__all_sync,__any_sync,__ballot_sync,__activemask,__match_any_sync,__match_all_sync,__shfl_sync,__shfl_up_sync,__shfl_down_sync, and__shfl_xor_sync.MIVisionX (3.1.0)
Changed
- AMD Clang is now the default CXX and C compiler.
- The dependency on rocDecode has been removed and automatic rocDecode installation is now disabled in the setup script.
Resolved issues
- Canny failure on Instinct MI300 has been fixed.
- Ubuntu 24.04 CTest failures have been fixed.
Known issues
- CentOS, Red Hat, and SLES requires the manual installation of
OpenCVandFFMPEG.- Hardware decode requires that ROCm is installed with
--usecase=graphics.Upcoming changes
- Optimized audio augmentations support for VX_RPP.
RCCL (2.21.5)
Changed
- Enhanced the user documentation.
Resolved Issues
- Corrected some user help strings in
install.sh.ROCm Compute Profiler (3.0.0)
Resolved issues
- Fixed a minor issue for users upgrading to ROCm 6.3 from 6.2 post-rename from
omniperf.
See ROCm Compute Profiler and ROCm Systems Profiler post-upgrade issues.ROCm Systems Profiler (0.1.0)
Added
- Improvements to support OMPT target offload.
Resolved issues
Fixed an issue with generated Perfetto files. See issue #3767 for more information.
Fixed an issue with merging multiple
.protofiles.Fixed an issue causing GPU resource data to be missing from traces of Instinct MI300A systems.
Fixed a minor issue for users upgrading to ROCm 6.3 from 6.2 post-rename from
omnitrace.
See ROCm Compute Profiler and ROCm Systems Profiler post-upgrade issues.ROCm known issues
ROCm known issues are noted on {fab}
githubGitHub. For known
issues related to individual components, review the Detailed component changes.PCI Express Qualification Tool failure on Debian 12
The PCI Express Qualification Tool (PEQT) module present in the ROCm Validation Suite (RVS) might fail due to the segmentation issue in Debian 12 (bookworm). This will result in failure to determine the characteristics of the PCIe interconnect between the host platform and the GPU like support for Gen 3 atomic completers, DMA transfer statistics, link speed, and link width. The standard PCIe command
lspcican be used as an alternative to view the characteristics of the PCIe bus interconnect with the GPU. This issue is under investigation and will be addressed in a future release. See GitHub issue #4175.ROCm resolved issues
The following are previously known issues resolved in this release. For resolved issues related to
individual components, review the Detailed component changes.Instinct MI300 series: backward weights convolution performance issue
Fixed a performance issue affecting certain tensor shapes during backward weights convolution when using FP16 or FP32 data types on Instinct MI300 series accelerators. See GitHub issue #4080.
ROCm Compute Profiler and ROCm Systems Profiler post-upgrade issues
Packaging metadata for ROCm Compute Profiler (
rocprofiler-compute) and ROCm Systems Profiler
(rocprofiler-systems) has been updated to handle the renaming from Omniperf and Omnitrace,
respectively. This fixes minor issues when upgrading from ROCm 6.2 to 6.3. For more information, see the GitHub issues
#4082 and
#4083.Stale file due to OpenCL ICD loader deprecation
When upgrading from ROCm 6.2.x to ROCm 6.3.0, the issue of removal of the
rocm-icd-loaderpackage
leaving a stale file in the oldrocm-6.2.xdirectory has been resolved. The stale files left during
the upgrade from ROCm 6.2.x to ROCm 6.3.0 will be removed when upgrading to ROCm 6.3.1. For more
information, see GitHub issue #4084.ROCm upcoming changes
The following changes to the ROCm software stack are anticipated for future releases.
AMDGPU wavefront size compiler macro deprecation
The
__AMDGCN_WAVEFRONT_SIZE__macro will be deprecated in an upcoming
release. It is recommended to remove any use of this macro. For more information, see AMDGPU
support.HIPCC Perl scripts deprecation
The HIPCC Perl scripts (
hipcc.plandhipconfig.pl) will be removed in an upcoming release.
AMD has released ROCm 6.3.1, featuring enhancements to its operating system and hardware support, versioning of components, identification of known and resolved issues, as well as forthcoming changes. Significant new features comprise improved resiliency for Instinct MI300 accelerators, the launch of the ROCm Runfile Installer, and revisions to the ROCm documentation. The latest ROCm Megatron-LM training Docker is now available alongside the ROCm vLLM inference Docker. Additionally, the Instinct MI300X workload tuning guide has been revised to include updated optimization strategies. The HIP graph-safe libraries function securely within HIP execution graphs, and the device memory topic in the HIP memory management section has been revised. ROCm 6.3.1 introduces support for Debian 12 (kernel 6.1) and the AMD Instinct MI325X accelerator.
