2 ms·
It could even help with low threadcount workloads since the L3 will presumably be able to be fewer cycles away from each core than a unified last level cache wo
by tcoppi 10y ago
It could even help with low threadcount workloads since the L3 will presumably be able to be fewer cycles away from each core than a unified last level cache would be.