Below average memory threaded score

[SilverShade]

Hi!
I have an Intel Core i7 10700F CPU with ADATA 2x8GB DDR4 RAM rated at 3000 MHz on a B460 motherboard which is capped at 2933 MHz and I'm getting these strange results in Performance Test and other benchmarks where everything looks decent except multi-core scores which are always below average and worse than my older PC which has worse specs.
I tried overclocking the RAM in BIOS and it does boost the results but multi-core scores still keep lagging behind (and still below average).
I'm trying to find out whether the issue lies in the fact that 3000 MHz RAM just doesn't work that well on 2933 MHz-capped motherboards or the issue is in the CPU or somewhere else.

I looked at the preferences and PerformanceTest does see all 16 threads. I have virtualization disabled. I have tried enabling XMP. Nothing in the system is overheating or throttling. But no matter what, my RAM just doesn't even get close to an average score in multi-core/thread tests.

Maybe there is a BIOS setting or some specific memory subtimings that affect multi core performance?

[David (PassMark)]

Memory benchmark results depends on the CPU, the RAM and BIOS settings.
The numbers you posted don't actually look that bad for a 10700F CPU. The base frequency of the 10700F isn't all that high (as it is only a 65W part) so it won't be as fast as some of the 100W+ parts.

[SilverShade]

Memory benchmark results depends on the CPU, the RAM and BIOS settings.
The numbers you posted don't actually look that bad for a 10700F CPU. The base frequency of the 10700F isn't all that high (as it is only a 65W part) so it won't be as fast as some of the 100W+ parts.

I see. My assumption that this score is subpar (while not terrible) came from the fact that my older PC (MSI B360 + i5-8600 + 2666MHz RAM) easily gets ~33-34k in the same test, while being worse at everything else. It is also a 65W processor, but its base clock is higher compared to that of 10700F. Could that be impacting the threaded memory score to such extent?

[David (PassMark)]

Old machine was actually the same on Memory Read, Memory Write and faster on threaded.
The Database test depends somewhat on single thread CPU performance. Read cached value depends more on CPU's memory cache.

So yes, it is a little strange that the DDR4-2666 was the same or better than DDR4-3000. Are you sure than XMP is on and Windows power plan is set to max performance?

[SilverShade]

Old machine was actually the same on Memory Read, Memory Write and faster on threaded.
The Database test depends somewhat on single thread CPU performance. Read cached value depends more on CPU's memory cache.

So yes, it is a little strange that the DDR4-2666 was the same or better than DDR4-3000. Are you sure than XMP is on and Windows power plan is set to max performance?

I did at some point switch the power plan to maximum performance but it didn't really affect the results. The results on the first screenshot were made with the built-in XMP activated, but since then I managed to tighten the timings a little and bumped the threaded performance to 32k. Still not great though. I wish I had an extra 2933 MHz kit to test, but the best I can do is to swap the RAM between two computers and see what it does.