Synthetic System Benchmarks
We are going to start with synthetic system application benchmarks on this page. Gaming performance will be shown later on. DDR4 RAM vs DDR5 RAM is being compared on each graph on the same CPU, DDR5 is the orange bar and DDR4 is the blue bar.
We are introducing a new test into our suite, we are using 3DMark Professional’s CPU Profile test. This test specifically tests CPUs and reports an overall score, it tests various thread counts. We are going to report on the “Max Threads” and “1-Thread” results only. Right off the bat, using Max Threads on the CPU this test reveals no difference in performance between the DDR4 or DDR5 platforms. The motherboards are different, so this could account for tiny differences, but there is nothing blaring from this test to show that either platform has an advantage because of the RAM type.
In the single-thread testing, we also find absolutely no difference in performance between DDR4 or DDR5 in this benchmark.
Next up we have the latest version of Geekbench 5. This benchmark tests overall CPU performance and can show us a result in both multi-core and single-core performance. This is the first evidence that DDR5 does make a performance difference. Geekbench, unlike 3DMark, tests a variety of workloads, whereas 3DMark is simply a single workload type. Therefore in multiple different workload types, and in multi-core, we are seeing DDR5 perform 9% faster than DDR4. That’s no small amount, that kind of performance difference is significant, and noticeable.
In single-core though, we are back to there not being a difference, it swings a little more toward the DDR4 in fact, perhaps due to better timings? It’s not significant though.