Performance
Our standard set of benchmarks includes MaxxMem Preview, AIDA64 and SuperPi 32M. These tests clearly show differences in memory performance, as such they’re some of the best to use to compare memory kits.
AIDA64 Memory and Cache Benchmark
As we see on the graph above, XMP #1 profile is really well configured. If notfor the lower read transfer then it could be as fast as the G.Skill kit at much tighter main timings.
XMP #2 also shows good performance but could be better. It’s still good option for all who can’t use faster memory frequency.
MaxxMem v1.99
In Maxxmem we see slightly different situation. 10th Anniversary edition memory has best results in memory copy and latency but the more relaxed settings are showing their weaker side in read and write transfers. Overall it still looks good.
SuperPi 32M
Kingston 10th Anniversary memory has the best time in SuperPi 32M. Looks like tight secondary settings are helping here much more than the first five timings.
Overclocking
Similar to previous Kingston reviews, here we can expect to set tighter sub timings but not really much higher memory frequency.
Using the Ivy Bridge platform I wasn’t able to set much higher memory clock. DDR3-2600 was working but wasn’t really stable.
After moving to the latest Intel Haswell platform results were slightly better. On a Gigabyte Z87X-OC motherboard I was able to stabilize 1300MHz ( DDR3-2600 ) using quite relaxed main timings of 11-14-14 and 1.65V.
In this case optimal settings will be still 1200MHz ( DDR3-2400 ), so XMP #1 after some tweaking. Lower memory clock and tighter timings should result in better overall performance so I can say that SPD/XMP programmers did a good job setting this profile.
