3 ms·
I wonder how this beast with AVX-512 compares to GPUs for AI model training and execution? I'm sure GPUs would still be more efficient for the specialize task
by api 3y ago
I wonder how this beast with AVX-512 compares to GPUs for AI model training and execution?
I'm sure GPUs would still be more efficient for the specialize task of heaps of matrix math, but given the insane price premium on high-end nVidia cards I wonder if this could be an alternative?
96 AVX512 cores is roughly 3072 parallel fp16 calculations per cycle.
- densh 3y agoNote: AMD implements AVX512 with execution over two cycles (aka double pumping), not in one cycle like Intel.
- bryanlarsen 3y agoIntel's latest chips don't implement AVX-512.
- pleoxy 3y agoThis lacks context. Avx-512 is disabled (fused off) in the big cores of Intel big.little. In the server and workstation chips that only have big cores, it is functional. It's right here in black and white. https://www.intel.com/content/www/us/en/products/sku/233483/intel-xeon-w93495x-processor-105m-cache-1-90-ghz/specifications.html https://www.intel.com/content/www/us/en/products/sku/233483/...