Finally, we went searching for the highest clocks without regard for the timings. The best speed we got was 1920MHz at 9-10-9-24. CAS 10 did not give us better clockspeeds, and the system simply would not POST at CAS 11.
We also tried increasing memory voltage but this did not have any effect on the overclocking results. Unfortunately we did not manage to make the memory stable above the 2000MHz mark.
A summary of our overclocking results is as follows
- 8-9-8-24, 1600Mhz (Tightest timings @ Default frequency) (CPU speed: 160×25=4000Mhz)
- 8-9-8-24, 1764Mhz (Tightest timings + Higher frequency) (CPU speed: 176×23=4048MHz)
- 9-10-9-24, 1920Mhz (Highest frequency at any timings) (CPU speed: 160×25=4000MHz)
We can see from the results below that going for tighter timings in favour of higher frequencies gives the best performance increase. Loosening the timings to achieve higher speeds was actually detrimental to performance in some cases. Clearly, one has to find the optimal balance between timings and clockspeeds.
The results of the Kingston HyperX T1 were decent but not spectacular. In this price region, there are many other enthusiast memory kits offering similar performance.
In the whole scheme of things, though, we should note once again that memory overclocking only increases performance marginally. For most users, the price premium for high grade memory modules could be better spent on upgrades elsewhere.
|Decent maximum frequency||Not very tight timings|
|Large heatsink may be incompatible with large CPU heatsinks|