Performance
Performance has been tested on the AMD Threadripper platform which contains the 1920X, 12-core processor, ASRock X399M Taichi motherboard, EVGA GTX1080Ti FE graphics card, and Crucial BX300 480GB SSD.
Because of the way how new AMD work, the Sport AT memory runs at DDR4-3000 CL18 as this is how motherboard sets it after enabling XMP profile. It doesn’t affect performance much. However, I understand that some readers are looking for additional results of how this memory can work on various platforms so in the comparison you can also find overclocking results. All results in the comparison are possible to achieve on our Sport AT DDR4-3000 memory kit but it’s not guaranteed you will be able to set them on your motherboards with other memory kits.
I will begin the tests with AIDA64 Memory and Cache benchmark which is probably the best software to check synthetic memory speed.
Even though Ballistix Sport AT has quite relaxed timings, memory bandwidth looks good. Results are of course also related to our test platform as AMD Ryzen and Threadripper is scaling well with memory frequency and less with memory latency. Large and fast processor’s cache is covering latency so end user can’t really see any delays.
Tests like PCMark 10 which base on popular applications is showing lower differences in performance between memory kit settings. The Sport AT DDR4-3000 is barely slower in this benchmark than the highest possible DDR4-3600.
The lower difference can be seen in rendering benchmarks like Cinebench R15. It’s because of what I mentioned earlier. Processor’s architecture is pretty much covering any delays while the memory is fast enough to deliver data for a constant high performance.
Graphics and gaming benchmarks like VRMark are showing a similar situation. Even though we can clearly say which settings are providing higher performance, the difference isn’t really significant. Ballistix Sport AT DDR4-3000 is fast even at XMP settings and we won’t need to overclock it to enjoy high performance.
I know that there are users who will say that DDR4-3000 at CL17 or CL18 is already slow but our tests are showing it’s more than enough for most users so it’s not always worth to believe marketing slogans and pay more for the memory, especially with current high prices.
The highest difference we can see in compression and decompression tests based on the 7-Zip benchmark. In this benchmark, we can see that higher memory frequency matters. In the same time, we can ask a question who uses compression and decompression all the time so faster memory would really matter?
Without any doubt, we can say there are faster memory kits on the market than the Ballistix Sport AT. We can find faster memory kits in the Ballistix offer. On the other hand, performance of the Ballistix Sport AT is more than enough for most users and for sure it meets gamers needs so I guess it’s better to spend saved money on a larger SSD or graphics card what would improve overall experience than to pay more for RAM which will show improvements mostly in synthetic benchmarks.
Overclocking
Disclaimer: Overclocking is never guaranteed, so the results may vary depending on certain conditions and various hardware configurations. 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.
We are used to seeing Ballistix memory kits at relaxed settings what is good because we are sure of the compatibility with various platforms and full stability. On the other hand, we always wish to see higher speed, regardless if the product is already overclocked or not.
Those who are interested in overclocking of the Ballistix Sport AT DDR4-3000 memory kit could already see results in our performance tests comparison. Below we are presenting some results out of the contest which is about maximum possible on our memory kit.
Because of used Micron IC, we are not able to set much tighter timings than what we already see on the product’s label. CL16 is maybe not so tight but already not bad setting. At CL16 we are able to set DDR4-3333 what is a quite good result.
Once we relax the timings then we can set the higher frequency. Maximum stable frequency seems 1800MHz (DDR4-3600). It’s also a limitation on the IMC side as everything above DDR4-3600 is already much harder to stabilize.
We can see how well memory bandwidth is scaling with memory frequency. Also, latency is improving, even though timings are even more relaxed. It’s one of the things about which people are forgetting while buying memory. Depends on the used platform, timings matter more or less. We can see that with each new platform, memory timings are playing a lesser role.
Maximum possible speed on our memory kit is DDR4-3733 but we had to set CL22 to make it work. We can still see higher bandwidth but for 24/7 work it can be already too much. In our case, this setting wasn’t fully stable but I guess if we spent some more time then it could work. We could only ask if it’s worth the time on all these tests when we won’t see any significant difference during daily work. For sure it’s worth to all who take a part in competitive benchmarking but those users will probably pick some other memory kits like Ballistix Elite which is designed for higher frequency and lower latency but in the same time costs more.