AVX-512 Performance With 256-bit vs. 512-bit Data Path For AMD EPYC 9005 CPUs

Written by Michael Larabel in Processors on 11 October 2024 at 10:40 AM EDT. Page 6 of 7. 21 Comments.

👁 libxsmm benchmark with settings of M N K: 128. AVX-512 512b DP was the fastest.

👁 ONNX Runtime benchmark with settings of Model: fcn-resnet101-11, Device: CPU, Executor: Standard. AVX-512 512b DP was the fastest.

👁 ONNX Runtime benchmark with settings of Model: bertsquad-12, Device: CPU, Executor: Standard. AVX-512 512b DP was the fastest.

👁 ONNX Runtime benchmark with settings of Model: GPT-2, Device: CPU, Executor: Standard. AVX-512 512b DP was the fastest.

👁 Numpy Benchmark benchmark with settings of . AVX-512 512b DP was the fastest.

👁 SVT-AV1 benchmark with settings of Encoder Mode: Preset 13, Input: Bosphorus 4K. AVX-512 512b DP was the fastest.

👁 SVT-AV1 benchmark with settings of Encoder Mode: Preset 8, Input: Bosphorus 4K. AVX-512 512b DP was the fastest.

👁 SVT-AV1 benchmark with settings of Encoder Mode: Preset 5, Input: Bosphorus 4K. AVX-512 512b DP was the fastest.

👁 SVT-AV1 benchmark with settings of Encoder Mode: Preset 3, Input: Bosphorus 4K. AVX-512 512b DP was the fastest.

By now you likely have a clear picture to the benefits of the 512-bit data path found with the Zen 5 cores on the AMD EPYC 9005 series. The 256-bit data path with Zen 4 was nice and still shows significant gains for EPYC 9005 even if running in the reduced FP256 mode, but a nice evolution in performance now with the native FP512 mode.

21 Comments - Next Page

Page: 1 2 3 4 5 6 7 Next Page

URL: https://www.phoronix.com/review/amd-epyc-9755-avx512/6

⇱ AVX-512 Performance With 256-bit vs. 512-bit Data Path For AMD EPYC 9005 CPUs Review - Phoronix

AVX-512 Performance With 256-bit vs. 512-bit Data Path For AMD EPYC 9005 CPUs