Advanced Micro Devices Inc (AMD)

[neye_eve]

chip guy - consistency question

If you could find magic zero latency memory most integer apps would improve in performance by only 40% or so. Going from 112 ns to 87 ns is a very small change in the bigger scheme of things and probably corresponds to a few percent higher improvement.

If, given your hypotheticals, a 112 to 0 "boost" in latency would give a 40% improvement, and going from 112 to 87 is a 22% improvement, then...

couldn't you apply a 40% performance gain on the 22% improvement, thus 0.4 * 0.22 = 9% improvement?

I realize that virtually nothing is linear, but a 10% gain is relatively large in the world of cpu's, and is double the 4% gain you hypothesized later:

I would be very surprised if it was more than 4% gain on
a 256 KB A64 and a 2% gain on a 1MB A64 on integer code.

TIA for any clarifications/explanations,
neye