A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 17, 04:15 UTC
[ROCm] include gfx908 as supported (#2792)
James Whedbee committed
Feb 19, 2024 at 19:58 UTC
264017a2bf030f060ebad91eb9be9b4e0033edb9
1 file changed
+1
-1
setup.py
+1
-1
@@ -24,7 +24,7 @@ MAIN_CUDA_VERSION = "12.1"
24
25
# Supported NVIDIA GPU architectures.
26
NVIDIA_SUPPORTED_ARCHS = {"7.0", "7.5", "8.0", "8.6", "8.9", "9.0"}
27
-ROCM_SUPPORTED_ARCHS = {"gfx90a", "gfx942", "gfx1100"}
27
+ROCM_SUPPORTED_ARCHS = {"gfx908", "gfx90a", "gfx942", "gfx1100"}
28
# SUPPORTED_ARCHS = NVIDIA_SUPPORTED_ARCHS.union(ROCM_SUPPORTED_ARCHS)
29
30