Performance
All of the tests were performed on an ASUS Maximus VII Gene motherboard equipped with an Intel i5 4690K CPU overclocked to 4.2GHz so it will not bottleneck the memory bandwidth. It is also easy to achieve this frequency and many enthusiasts are actually setting unlocked Haswell CPUs higher. I think that the presented below results will be easy to compare for all our readers especially that most of the used benchmarking software is free.
AIDA64 Cache and Memory Benchmark
AIDA64 is one of my favorite benchmarks. It is great to compare memory kits and thanks to FinalWire, we always have the latest version.
On the graphs above we can see that the higher clock is giving us higher bandwidth. The memory kits in the comparison are all based on 4GB modules so it is easier to see how they are scaling. The ADATA XPG V3 DDR3-2933 kit is winning in this stage but the results are not much higher than the DDR3-2800 memory. A higher processor clock would probably show the higher write bandwidth of this memory but the most important speed in daily usage seems to be memory read.
MaxxMem Preview v1.99
AIDA64 is showing us the multithreaded performance while MaxxMem Preview is using only a single thread. Our processor limited to one thread is clearly reaching a maximum write bandwidth of about 24GB/s. The memory read and copy are not limited by this factor and we can see that the higher clock of ADATA memory is again showing some advantage. Again, not much higher than the DDR3-2800 memory but it still counts.
HyperPi 32M
HyperPi 32M loves fast memory so it is clear that the ADATA DDR3-2933 kit has the best results in our comparison. Looking at the memory clock we could actually expect better results. However it is hard to find any faster DDR3-2933 memory kit on our market as all are using similar memory chips. In a direct comparison the most important thing seems to be the XMP profile programming and ADATA made it good. The final result is not bad however tightening the sub-timings can for sure improve it.
Cinebench R15
Cinebench is a rendering benchmark and usually the differences between memory kits are low. Also here we see that the difference between the fastest and slowest memory kit is only 10 points. This does not change the fact that the ADATA XPG V3 memory achieved the best result in this test.
Futuremark PCMark 8
PCMark 8 is testing the general performance of our computer. The memory tests are affecting the final score but as we could see in previous review scores are not raising much above the DDR3-2400 mark. The ADATA memory is again the best but other DDR3-2400 or higher memory kits are not much slower. In daily work we would not see any difference.
The ADATA XPG V3 DDR3-2933 memory performance is good but not as good as we could expect looking at the high base frequency. In many tests we can see higher performance than the other, slower clocked memory kits but in daily work it really does not matter. Overclockers will probably see it otherwise but more average users should stick to much cheaper memory like DDR3-2133 or 2400.
Overclocking
Overclocking is never guaranteed so the presented results may vary from results on other memory kits. I am not recommending overclocking if you do not know what are you doing. High voltages may damage hardware and it will not be covered by warranty.
Overclocking was performed on two motherboards. At the beginning I was trying to set maximum a clock on ASUS Maximus VII Gene but I could not pass the DDR3-3200 clock. It was not satisfying as I know that the Hynix MFR can make some more speed. The Maximus VII Gene gave me a nice 1500MHz clock (DDR3-3000) at the declared XMP settings of CL12-14-14-36 and 1.65V. It is not bad as many DDR3-2933 need higher voltage to reach these settings.
Since I could not make more on the ASUS motherboard, I moved on and checked overclocking on the Gigabyte Z97X-SOC Force. I was right that the ADATA XPG V3 DDR3-2933 can make a higher clock and I have reached the 1667MHz (DDR3-3333) mark which passed couple of tests like HyperPi 32M without bigger issues.
The frequency looks great but as we can see in the AIDA64 screenshots, the higher clock does not go in pair with high performance. These results are actually as high as DDR3-1866/2133 memory can offer. Higher memory clock means also more relaxed timings and huge drop in general performance. However I can assure you that it is not ADATA’s fault as most other memory brands perform similar at DDR3-3000+.
