View on GitHub ROCm_Logo

ROCm, a New Era in Open GPU Computing

Platform for GPU-Enabled HPC and Ultrascale Computing

Hardware to Play ROCm

ROCm Platform Supports Two Graphics Core Next (GCN) GPU Generations

Experimental support for GFX7 GPU's: Radeon R9 290 4 GB, Radeon R9 290X 8 GB, Radeon R9 390 8 GB, Radeon R9 390X 8 GB, FirePro W9100 (16GB), FirePro S9150 (16 GB) and FirePro S9170 (32 GB); strong half-rate double-precision performance

Production Support for GFX8: Radeon RX 480, Radeon RX 470, Radeon RX 460,Radeon R9 Nano, Radeon R9 Fury, Radeon R9 Fury X Radeon Pro WX7100, Radeon Pro WX5100,Radeon Pro WX4100, and FirePro S9300 x2; revolutionary HBM memories with 512 GB/s per ASIC. Currently the highest efficiency at 46 Gflops/watt. The R9 Nano delivers 8 Tflops and the S9300x2 delivers a solid 13.9 Tflops of single-precision performance to attack machine-learning, molecular-dynamics, energy and signal-processing problems.

CPU support: ROCm requires CPUs with PCIe Gen3 atomics capability, such as the Intel Xeon E5v3 and newer Xeon processors. It also supports consumer CPUs, Intel Core i3, Intel Core i5, Intel Core i7 CPUs (v3 (Haswell) and newer). We know you’re going to ask—yes, ROCm is optimized to pull all the rich capabilities out of the new AMD Zen-based processors.

More Information on ROCm-Supported Server-Capable Hardware

FirePro S9150
FirePro S9170
FirePro S9300 X2
Radeon R9 Nano