This repository contains the manifest file for ROCm™ releases, changelogs, and release information. The file default.xml contains information for all repositories and the associated commit used to build the current ROCm release.
The default.xml file uses the repo Manifest format.
cd docs
pip3 install -r .sphinx/requirements.txt
python3 -m sphinx -T -E -b html -d _build/doctrees -D language=en . _build/htmlROCm v5.4.3 is now released. For ROCm v5.4.3 documentation, refer to https://docs.amd.com.
ROCm v5.4.2 is now released. For ROCm v5.4.2 documentation, refer to https://docs.amd.com.
ROCm v5.4.1 is now released. For ROCm v5.4.1 documentation, refer to https://docs.amd.com.
ROCm v5.4 is now released. For ROCm v5.4 documentation, refer to https://docs.amd.com.
ROCm v5.3.3 is now released. For ROCm v5.3.3 documentation, refer to https://docs.amd.com.
ROCm v5.3.2 is now released. For ROCm v5.3.2 documentation, refer to https://docs.amd.com.
ROCm v5.3 is now released. For ROCm v5.3 documentation, refer to https://docs.amd.com.
The ROCm v5.2.3 patch release is now available. The details are listed below. Highlights of this release include enhancements in RCCL version compatibility and minor bug fixes in the HIP Runtime.
Additionally, ROCm releases will return to use of the ROCm repository for version-controlled release notes henceforth.
NOTE: This release of ROCm is validated with the AMDGPU release v22.20.1.
All users of the ROCm v5.2.1 release and below are encouraged to upgrade. Refer to https://docs.amd.com for documentation associated with this release.
Refer to the following article for information on the preview support for Ubuntu 20.04.5 HWE.
https://www.amd.com/en/support/kb/release-notes/rn-amdgpu-unified-linux-22-20
Support for Ubuntu 18.04 ends in this release. Future releases of ROCm will not provide prebuilt packages for Ubuntu 18.04.
- A bug was discovered in the HIP graph capture implementation in the ROCm v5.2.0 release. If the same kernel is called twice (with different argument values) in a graph capture, the implementation only kept the argument values for the second kernel call.
- A bug was introduced in the
hiprtcimplementation in the ROCm v5.2.0 release. This bug caused thehiprtcGetLoweredNamecall to fail for named expressions with whitespace in it.
Example: The named expression my_sqrt<complex<double>> passed but
my_sqrt<complex<double>> failed.
- Compatibility with NCCL 2.12.10
- Packages for test and benchmark executables on all supported OSes using CPack
- Adding custom signal handler - opt-in with
RCCL_ENABLE_SIGNALHANDLER=1- Additional details provided if Binary File Descriptor library (BFD) is pre-installed.
- Adding experimental support for using multiple ranks per device
- Requires using a new interface to create communicator
(
ncclCommInitRankMulti), refer to the interface documentation for details. - To avoid potential deadlocks, user might have to set an environment variables increasing the number of hardware queues. For example,
- Requires using a new interface to create communicator
(
export GPU_MAX_HW_QUEUES=16- Adding support for reusing ports in NET/IB channels
- Opt-in with
NCCL_IB_SOCK_CLIENT_PORT_REUSE=1andNCCL_IB_SOCK_SERVER_PORT_REUSE=1- When "
Call to bind failed: Address already in use" error happens in large-scale AlltoAll (for example, >=64 MI200 nodes), users are suggested to opt-in either one or both of the options to resolve the massive port usage issue - Avoid using
NCCL_IB_SOCK_SERVER_PORT_REUSEwhenNCCL_NCHANNELS_PER_NET_PEERis tuned >1
- When "
- Opt-in with
- Removed experimental clique-based kernels
No notable changes in this release for development tools, including the compiler, profiler, and debugger.
No notable changes in this release for deployment and management tools.
For release information for older ROCm™ releases, refer to CHANGELOG.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent