2 ms·

Not really on topic, but compare those numbers with the slowness of RAM and cache. A single shared variable that gets updated on one socket then another -- that

Not really on topic, but compare those numbers with the slowness of RAM and cache. A single shared variable that gets updated on one socket then another -- that can cost like 300 cycles, at least on Nehalem.

cbsmith 11y ago
In terms of throughput, the RAM is great. It's the latency thing. Which is why intelligent scheduling with embarrassingly parallel problems like this makes such a huge difference.