Once the next generation (CannonLake) is out, AVX-512 will be available.
With that, the SIMD width will be doubled, and AVX-512 optimized code could possibly double in performance as well, as long as the memory can keep up.
With that, the SIMD width will be doubled, and AVX-512 optimized code could possibly double in performance as well, as long as the memory can keep up.
No comments yet.