4 ms·
Do server applications ever optimize for ARM-equivalent of AVX512? Given that AWS heavily uses Graviton, I assume so? But I never hear of them - only AVX512 op
by aurareturn 2y ago
Do server applications ever optimize for ARM-equivalent of AVX512?
Given that AWS heavily uses Graviton, I assume so? But I never hear of them - only AVX512 optimizations.
- EdSchouten 2y agoARM64’s equivalent (NEON) only has 128 bit registers, so the gains of using SIMD are not as massive as on x86-64. Edit: turns out there’s nowadays something called SVE that provides wider registers: https://developer.arm.com/Architectures/Scalable%20Vector%20Extensions https://developer.arm.com/Architectures/Scalable%20Vector%20...