AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Added ability to build for local gpu by env Variable (#204)

Browse Source 
* Added the ability to compile for Local gpu by environment variable

* adding gfx950 on default only on rocm 7.0 and above

* Updated docs

* removed xnack+ on specific gfx targets

---------

Co-authored-by: Yiltan Hassan Temucin <yiltan.temucin@amd.com>
This commit is contained in:
akolliasAMD
2025-08-11 12:35:50 -06:00
committed by GitHub
parent e95360961d
commit be630d9b93
4 changed files with 45 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/install.rst
+3 -1
View File
@@ -21,6 +21,8 @@ Requirements
* MI300X
* MI350X (Requires ROCm 7.0 or later)
* ROCm-aware Open MPI and UCX. For more information, see :ref:`install-dependencies`.
* Inter-node communication requires MPI, and is tested with Open MPI and CX7 Infiniband NICs.
@@ -34,7 +36,7 @@ rocSHMEM supports two network backends:
* The **RO (Reverse Offload)** backend enables communication between GPUs on different nodes through a NIC, using a host-based proxy to forward communication orders to and from the GPU. In this release, RO is the only inter-node communication backend and is built on an MPI-RMA compatibility layer.
You can activate IPC and RO backends in the same rocSHMEM build. In this case, IPC handles intra-node communication, while RO handles inter-node communication.
You can activate IPC and RO backends in the same rocSHMEM build. In this case, IPC handles intra-node communication, while RO handles inter-node communication.
.. note::