Performance
Performance has been tested on the Intel platform, including the i7-14700K processor, ASUS Z790 APEX motherboard, Colorful RTX4080 Advanced OC 16GB graphics card, Kingston Renegade 2TB NVMe SSD, and FSP 1350W 80+ Platinum PSU.
All results were performed on the KLEVV BOLT V 32GB DDR5-6800 memory kit. Our stable overclocking limit was 8200MT/s, but overclocking is not guaranteed. At this clock, the memory could run at respectable timings of CL38-48-48 at the 1.45/1.45V VDD/VDDQ. The voltage is suitable for daily usage in PC cases with good airflow.
We will start with the AIDA64 Memory and Cache benchmark, probably the best application for checking memory bandwidth and latency.
The results in the AIDA64 benchmark almost always look better at higher frequencies, so it is no wonder that our overclocked settings give higher bandwidth. It does not always translate into a performance gain in daily work but suggests that well-balanced frequency and timings provide the best results.
The latency test shows the difference between the XMP and overclocked settings is about 9ns. It’s not so much, and we can say that the XMP is pretty fast. Memory manufacturers don’t tweak timings too much to provide total stability and compatibility. Even though it could be better, it’s hard to find another 6800MT/s memory kit with tighter timings than KLEVV BOLT V.
AIDA64 tests are fully synthetic and usually do not present real-world performance. The following tests should give a better view of the daily performance.
As long as PCMark 10 Applications benchmarks show differences between various RAM settings, we can barely see it in 3DMark benchmarks.
The latest Cinebench 2024 reacts quite well to RAM performance. We can clearly say which memory will be faster in rendering, even though scores do not seem so different. We must remember that this benchmark takes only a few minutes, and rendering tasks usually take much longer.
Final Fantasy XV and Superposition benchmarks have all results close to the error margin again.
The most significant improvements are shown in games. At lower display resolutions like 1080p, the BOLT V at 6800MT/s should give us about 5-7% improvements over the standard 5600-6000 memory series and the next 5% if we overclock it. We shouldn’t expect as big improvements if we play at higher display resolutions. As long as 1440p is still scaling well, then at 4k, it will be insignificant.
KLEVV BOLT V provides high performance out of the box. We don’t have to overclock it to enjoy optimal results. However, since it uses Hynix A-die, it guarantees high overclocking potential, and users who wish, can push it much higher than it’s specified.
I will tell you more about overclocking on the next page of this review.
