RAM Overclocking
RAM Overclocking
A 4 stick RAM setup will put a strain on the CPU's memory controller, making it tougher to maintain a stable OC compared to 2 sticks. If even one stick struggles at that speed, the system won't function properly regardless of the others. You might want to use 1usmus's DRAM Calculator for Ryzen to check if the settings are reliable. Overall, memory latency is unlikely to be the main issue, so performance gains probably won't be significant.
Running four sticks is fine. The real issue arises when you're aiming for the fastest clocks. Corsair packs their chips tightly, so you probably won<|pad|> to fall short.
When testing overclocking scenarios, compare stock or standard XMP profiles against your custom builds. In my case, boosting from 2933MHz to 3600MHz mainly affects latency rather than raw speed. The biggest impact comes from reduced delays, not just clock rates. I haven’t seen significant performance jumps, but results can vary. Regression matters—don’t assume improvements without verification. (It’s true, Corsair RAM stays close to rated speeds.) It might be worth checking with a more experienced source, maybe someone claiming expertise. (Probably needs more voltage too.) Generally, higher frequencies tend to increase latency, though the effect depends on measurement methods. My CPU is performing well—98th percentile, barely noticeable difference between sticks.)
I maintained the CPU at 4600MHz, with all sub timings matching what I can observe and adjust. I avoid using XMP since its profile settings are too broad. Instead, I apply the same profile on my 5600X to achieve speeds of 4000MHz or higher. Essentially, I only modified the fclk and mclk rates. I didn’t want the CPU to increase too much because there’s a significant difference between 2800MHz and 5150MHz, which could affect the outcomes. My 5600X: And one more from my 5900X
That matches what you expect (the person in the video also disabled CPU boost, likely for other reasons). I notice your latency improved but you maintained most of the timing settings. That’s worth mentioning—it’s interesting since RAM speed usually doesn’t affect latency much, but it can still play a role. When I adjusted my RAM speed, I had to change the timing settings, and I didn’t observe significant changes in performance tests. Probably because latency tends to increase when timing is relaxed, which aligns with what you saw in the benchmark. For games, a 1-2 FPS difference might be noticeable. I’m expecting latency to rise if timing is loosened, as seen in the benchmark test. It seems I might need higher voltage to get stable 3600MHz operation, but overall I’m satisfied the device started up fine. My motherboard still seems unstable, so I left it as is at 1.36V. Note: My Ub tests improved with faster RAM, which is good news.
I discovered the method for determining real-world latency (1 divided by GhzFrequency multiplied by CasLatency equals Actual Latency). This clarified how to assess various RAM types and overclocking possibilities.