|A member of the Funky Media Group|
|Review: Patriot Viper Xtreme DDR3-1600 Tri-Channel 6Gb Kit|
|Posted by Ed Smith|
|Thursday, 28 April 2011 23:01|
Page 7 of 8
Now for my favorite part, overclocking!
Because all of the benchmarks I've used also use the CPU I elected to keep the cpu at the exact same speed as for all the other tests, otherwise it wouldn't be a direct comparison.
This also means leaving the ram at the same speed, which seems reasonable to me anyway and you don't have to go much higher in ram speed before you start running into memory controller limitations on 1366 quad core cpus.
There are a few things I have done here, for starters I dropped the middle primary timing down one, this is an important timing for DDR3 ram and warrants attention.
The first timing, CAS, won't go any lower without adding more voltage to the ram, something I want to avoid as I can't really recommend violating specs and risking meltdown.
The third timing I dropped largely because I could, it makes only a very small difference. The same goes for "tRFC", I dropped it from 88 to 82 mostly because I could, and out of habit from benching. Plus every little bit helps.
Lastly I dropped the command rate from 2T to 1T, this made the largest difference by a good bit. As with all overclocking you may or may not be able to emulate this with the same kit, but the 2T to 1T drop you should be able to do, and I recommend trying.
Now on to the test results!
SiSoft Sandra Pro
By and large a decent gain from five minutes of playing with memory timings. Sciencemark didn't seem to care much, but the other benchmarks certainly did!