RYZEN, FX 9 & 8 seriers cpu, Passmark benchmarks, Legit?

[WamerD]

I've been using and watching Passmark benchmarks for years. Why is it that AMD cpu's have such poor memory latency and bandwidth benchmarks. Also 2d graphics benchmarks show low results. I've tested both AMD and Intel (Haswell) with the same memory, gpu's and ssd's (sata3 and m.2 PCIE) and the results are quite different. Any insight would be helpful.

 

[synphul]

It's likely due to how the cpu's handle memory management differently from one another. Passmark is only a general idea, more helpful are looking at benchmarks that pertain to real world applications you're planning to use. Gaming, photoshop, office programs, video encoding? Synthetic benchmarks aren't always the best to go by.

A synthetic benchmark can show a technical numerical difference based upon an arbitrary points based score between say single and dual channel ram. In real world apps there's little to no perceptible difference 90% of the time so it doesn't pay to get stressed over it. If you game and it doesn't add fps, is it helping you? If you encode video and your tasks don't realize a shorter time processing a video clip or movie, does it matter? Someone can show a huge difference on some imaginary scale of measure but unless the actual tasks you perform benefit then it's nothing more than theory.

I wouldn't get too hung up on passmark scores, instead if comparing ryzen, fx or intel then it would be more beneficial to compare which will work better for you under actual programs you're likely to use.