Doubled memory capacity but still slower performance
Doubled memory capacity but still slower performance
I assembled a fresh PC a few months ago. The main parts include: 5950x 3.4GHZ Asus Rog Crosshair viii Formula G.SKILL TridentZ RGB 32GB (two 16GB modules), DDR4 4000 CL 16 MSI 2080 (from the previous build; luck with Newegg!). So far, I’ve been using DOCP and running RAM at 4000 MHz. At first, I thought 32GB would suffice, but since my work often requires heavy MS Access queries, I increased it to double the amount. I bought the identical setup again. After installing the new modules, the PC wouldn’t boot. I reset the CMOS and tried lowering the RAM speed. Eventually, after underclocking the RAM to 3666 MHz, it started working. Does this mean more RAM forces slower clock speeds? Is there a way to run all 64GB at full speed?
You're questioning whether the changes are noticeable. A quick check confirmed no significant difference. I was hoping for improved performance, but this upgrade mainly speeds up queries. The auto setting should help too. Asynchronous RAM sounds technical—what do you think?
Ryzen includes a communication layer named infinity fabric with a clockspeed (FCLK). For best results, this should align with the actual RAM speed (2000MHz here). But the starting FCLK on Zen 3 is only 1900MHz. It might be possible to raise it to 2000MHz, though that depends on the process. If it can't match exactly, it switches to asynchronous operation, which could cause a performance drop of 20% or more. The effect of "Auto" in this case isn't clear—it could mean running at 2000MHz or staying lower. This instability in FCLK might also impact RAM stability.