PRELIMINARY RESULT Benchmark Suite v1.0.15 - Results subject to change
Home › Hardware Acceleration Impact

OpenSSL Performance Benchmark

Hardware Acceleration Impact3 iterations per version
Last run: 2026-01-05 • Iterations: 1.1.1w (3x), 3.0.18 (3x), 3.1.8 (3x), 3.2.6 (3x), 3.3.5 (3x), 3.4.3 (3x), 3.5.4 (3x), 3.6.0 (3x)

⚡ Hardware Acceleration Impact on Cryptographic Performance

What This Page Shows

This benchmark measures the performance difference when hardware acceleration (AVX/AVX2 on x86, NEON/Crypto on ARM) is enabled vs disabled.

Testing Method: Each algorithm is benchmarked twice - once with hardware acceleration enabled (default), and once with it disabled using OpenSSL's capability masking (`OPENSSL_ia32cap` on x86, `OPENSSL_armcap` on ARM).

Why This Matters: Post-quantum algorithms like ML-KEM heavily rely on SIMD vectorization. Understanding the impact helps with hardware selection and performance planning.

Symmetric Cryptography Impact

AES-256-GCM and SHA256 throughput with and without hardware acceleration

ML-KEM-768 (Post-Quantum) Impact

Post-quantum key encapsulation performance with and without SIMD acceleration

Why ML-KEM Benefits Most from SIMD

ML-KEM (Kyber) is a lattice-based algorithm that involves:

  • Matrix-vector multiplications - parallelizable across SIMD lanes
  • Number Theoretic Transform (NTT) - butterfly operations map perfectly to AVX2
  • Polynomial arithmetic - coefficient operations are independent

This makes ML-KEM an ideal candidate for 50-100%+ performance gains with AVX2/NEON acceleration.

📈 Improvement Summary

View on GitHub
Open Source Benchmark
Found a problem? Have an improvement?
Fork the repository and submit a pull request!
Licensed under Apache 2.0 • Community-driven development • v1.0.15