4 ms·
> as little memory latency as possible Should be "as much memory bandwidth as possible". GPUs are designed to be (relatively) more insensitive to memory latenc
by rfoo 2y ago
> as little memory latency as possible
Should be "as much memory bandwidth as possible". GPUs are designed to be (relatively) more insensitive to memory latency than CPU.
- muyuu 2y agoyep that's true, although AI compute modules do get significant benefit from low latency cache as well