LLVM Clang 22 Compiler Performance Largely Unchanged Over Clang 21 On AMD Zen 5

Written by Michael Larabel in Software on 25 February 2026 at 03:05 PM EST. Page 4 of 4. Add A Comment.

👁 OpenVINO benchmark with settings of Model: Machine Translation EN To DE FP16, Device: CPU, Benchmark Mode: Throughput. Clang 20.1.8 was the fastest.

👁 OpenVINO benchmark with settings of Model: Noise Suppression Poconet-Like FP16, Device: CPU, Benchmark Mode: Throughput. Clang 22.1.0 was the fastest.

👁 Llama.cpp benchmark with settings of Backend: CPU BLAS, Model: Qwen3-8B-Q8_0, Test: Prompt Processing 512. Clang 22.1.0 was the fastest.

👁 Llama.cpp benchmark with settings of Backend: CPU BLAS, Model: gpt-oss-20b-Q8_0, Test: Text Generation 128. Clang 22.1.0 was the fastest.

👁 Llama.cpp benchmark with settings of Backend: CPU BLAS, Model: gpt-oss-20b-Q8_0, Test: Prompt Processing 512. Clang 21.1.8 was the fastest.

👁 Llama.cpp benchmark with settings of Backend: CPU BLAS, Model: GLM-4.7-Flash-IQ4_XS, Test: Text Generation 128. Clang 22.1.0 was the fastest.

👁 Llama.cpp benchmark with settings of Backend: CPU BLAS, Model: MiniMax-M2.5-UD-TQ1_0, Test: Prompt Processing 512. Clang 22.1.0 was the fastest.

👁 Llama.cpp benchmark with settings of Backend: CPU BLAS, Model: Mistral-7B-Instruct-v0.3-Q8_0, Test: Prompt Processing 512. Clang 22.1.0 was the fastest.

👁 Llama.cpp benchmark with settings of Backend: CPU BLAS, Model: granite-3.0-3b-a800m-instruct-Q8_0, Test: Text Generation 128. Clang 22.1.0 was the fastest.

While no big performance gains were found out of using the new LLVM Clang 22.1 compiler release, at the same time no performance regressions were observed either. It's been smooth sailing with the LLVM Clang 22 benchmarking thus far on Intel and AMD hardware. I'll also be trying some RISC-V and AArch64 benchmarks on Clang 22 though there is less hardware around for that testing and what's here is slower than the latest Intel/AMD CPUs so will be some days before any assessments on those architectures.

Add A Comment

If you enjoyed this article consider joining Phoronix Premium to view this site ad-free, multi-page articles on a single page, and other benefits. PayPal or Stripe tips are also graciously accepted. Thanks for your support.

Page: 1 2 3 4

URL: https://www.phoronix.com/review/llvm-clang-22-znver5/4

⇱ LLVM Clang 22 Compiler Performance Largely Unchanged Over Clang 21 On AMD Zen 5 Review - Phoronix

LLVM Clang 22 Compiler Performance Largely Unchanged Over Clang 21 On AMD Zen 5