3 ms·
Apparently in specific workloads the move from AVX -> AVX2 was very substantial, though I can't seem to find the article that showed this.
by supergauntlet 12y ago
Apparently in specific workloads the move from AVX -> AVX2 was very substantial, though I can't seem to find the article that showed this.
- vardump 12y agoThe transition is more like from SSE2 -> AVX2, in integer workloads. AVX2 adds 256-bit wide integer instructions and gather-instruction. Gather does multiple loads in one instruction. AVX was just about floating point.