B2B Delay Testing

[Cecil]

Been trying to figure out how much difference B2B delay makes. Seems to help with multi threaded benches when lower, and helps with single threaded benches when higher.

4.2HT 1600Mhz 7-8-7-20-1t B2B 5

NuclearusMC - ALU 8469 - FPU 11920 - MT 33889 - Total 25294

Super Pi - 1m - 9.719sec

Wprime 1.55 - 32m - 6.106sec

Everest - Read 16965 - Write 15159 - Copy 22477 - Latency 45.6

4.2HT 1600Mhz 7-8-7-20-1t B2B 4

NuclearusMC - ALU 8450 - FPU 11952 - MT 38348 - Total 28190

Super Pi - 1m - 9.750sec

Wprime 1.55 - 32m - 5.987sec

Everest - Read 17976 - Write 15164 - Copy 21123 - Latency 45.9

4.2HT 1600Mhz 7-8-7-20-1t B2B 3

NuclearusMC - ALU 8444 - FPU 11920 - MT 38487 - 28277

Super Pi - 1m - 9.766sec

Wprime 1.55 - 32m - 5.972

Everest - Read 18265 - Write 15161 - Copy 22338 - Latency 46