AVX-512 Performance + Power Efficiency Shines With AMD Strix Halo

Written by Michael Larabel in Processors on 28 May 2025 at 12:30 PM EDT. Page 4 of 8. 22 Comments.

👁 Llama.cpp benchmark with settings of Backend: CPU BLAS, Model: Llama-3.1-Tulu-3-8B-Q8_0, Test: Prompt Processing 512. AVX-512 On was the fastest.

👁 Llama.cpp benchmark with settings of Backend: CPU BLAS, Model: granite-3.0-3b-a800m-instruct-Q8_0, Test: Prompt Processing 512. AVX-512 On was the fastest.

👁 Llama.cpp benchmark with settings of Backend: CPU BLAS, Model: granite-3.0-3b-a800m-instruct-Q8_0, Test: Prompt Processing 1024. AVX-512 On was the fastest.

👁 Llama.cpp benchmark with settings of Backend: CPU BLAS, Model: Mistral-7B-Instruct-v0.3-Q8_0, Test: Prompt Processing 512. AVX-512 On was the fastest.

Similarly, Llama.cpp also enjoyed very significant uplift thanks to AVX-512 on Strix Halo.

👁 LocalScore benchmark with settings of Model: Llama-3.2-1B-Instruct-Q4_K_M.gguf, Acceleration: CPU. AVX-512 On was the fastest.

👁 LocalScore benchmark with settings of Model: Meta-Llama-3.1-8B-Instruct-Q4_K_M.gguf, Acceleration: CPU. AVX-512 On was the fastest.

👁 LocalScore benchmark with settings of Model: Qwen2.5-14B-Instruct-Q4_K_M.gguf, Acceleration: CPU. AVX-512 On was the fastest.

Whether you are looking to leverage workloads on the Radeon 8060S graphics and/or the Zen 5 cores, AMD Strix Halo with AVX-512 is very strongly positioned for AI developers with no viable competition from Intel's current laptops. Lunar Lake as a reminder has no AVX-512, only eight cores without SMT, and comes well short of Strix Halo graphics performance with its Xe2 integrated graphics.

22 Comments - Next Page

Page: 1 2 3 4 5 6 7 8 Next Page

URL: https://www.phoronix.com/review/amd-strix-halo-avx512/4

⇱ AVX-512 Performance + Power Efficiency Shines With AMD Strix Halo - Phoronix

AVX-512 Performance + Power Efficiency Shines With AMD Strix Halo