Overclocking
For most people, 1866mhz ram speed is plenty. But for others they will want the maximum amount of performance from everything they have. For testing I kept the limit of 1.65v and went from there. No bclk was added so that benchmarks that are aided from faster processor speeds would not be biased in having a higher score.
CPU-Z Information
At first I tried the same timings but bumping the speed up to 2133mhz but that did not work well at all. Next I tried tightening up the timings. I tried 9-9-9 and that did not work at all, so I tried to cut down the three main timings by one each and that worked well. While you are not getting any higher frequency, you are however tightening up the timings which is always nice. There was not even any added voltage. The fact that you can change the command rate to one means that you are getting around a 20% increase in speed from just that one setting.
I was also able to tighten up the subtimings, in fact the tRFC timing is almost half of what it was with the XMP profile!
Superpi 32m
For some reason superpi 32m would not finish with any tightened timings. It would crash with any of the timings tightened up. This could be due to the combination of my processor possibly having a weak IMC (memory controller) and the chips on this particular set of ram. Others I have talked to have had no issues with this ram at all so I assume it is an isolated instance of test and memory incompatibility.
Wprime
Once again we see a bit of difference from tightened timings in wprime. While this is not a huge amount of difference, those few miliseconds can mean alot to competitive benchmarkers.
Maxxmem
Once again we see the lowered timings making a difference in the hwbot version of Maxxmem, particularly in the latency section.
Maxxmem2
Maxxmem2 improves as well, it just goes to show that even a slight tightening of timings can make a difference in testing.
