How much effort should I invest in adding more RAM? - Beginner
How much effort should I invest in adding more RAM? - Beginner
I see myself as a beginner on RAM overclocking, but I'm confident here *poof*.
I have two RAM modules with comparable timings:
2x 18 - 3200
2x 32 - 2933
The issue is the laptop's BIOS is restricted. I discovered a GitHub tool that appears to alter the BIOS to believe faster RAM is present, tricking both the laptop and Windows.
I tested Cinebench Multi; the overclocked 64GB scored slightly better (just 40 points). I'm currently stress testing in CB for stability.
Is there any point I should risk by exceeding 3200 or beyond? Should I push further?
It's suitable for some older and moderately newer games plus video transcoding (hence 64GB). It features an AMD 6800m GPU (about a 6700 desktop) at 1080 w/300hz. If I can improve performance further and enough people believe it's worth it, I'll try it. However, I'm cautious about RAM and its actual impact compared to DDR4. Perhaps I'm overthinking in terms of CPU, but 300mhz is a noticeable improvement over XMP speeds that were already 2933. Still, I'm not a RAM expert. I was satisfied at the 3200 level but saw that 3200 might be the minimum for Ryzen (not sure if that's accurate). I'm not aiming for high scores, just real-world noticeable improvements. If we're talking about 1-2 frames, it doesn't seem worthwhile to me. The tool functioned, but I'm unsure how it affects the BIOS (even after reading the code) and I'm also worried about repeated changes.
There isn't a significant real-world boost from increasing RAM speed. A single example from a site examining DDR4-3200, DDR4-3600, and DDR4-3733 shows limited results. The article discusses AMD Ryzen 3000 systems and how memory timings and frequency affect performance in games. It notes that the changes AMD introduced in their third generation lineup are worth noting. The discussion also touches on how memory speed influences gaming outcomes, with some evidence suggesting minimal gains.
It seems like the situation is likely just a minor issue, so I’ll stick with what I have. I tested the laptop under light stress and didn’t encounter any major problems. I won’t bother with further tweaks or obsessing over it.
I avoid Low-Power RAM because I'm not worried about battery life and I don't want any performance issues.
Ram speeds vary depending on the application. Certain software such as AutoCAD can take advantage of higher-speed RAM, while games like Microsoft Flight Sim also benefit. However, for most users the difference is negligible. Ryzen processors gain from faster RAM, but beyond a point it surpasses the F-clock, shifting from a 1:1 to a 2:1 ratio and negatively impacting performance. Generally, performance improvements come from matching speed with appropriate timings. There isn’t a clear distinction between, for example, 3200/14 and 3600/16, as the faster speed is often offset by looser timings, and vice versa. Therefore, a device with tighter timings than another may match performance while offering the advantage of double capacity, except for specific RAM-related programs.
Are there indicators for CPU usage or similar metrics in Windows?
It's more about the knowledge aspect. If you could achieve 3200MHz, it would indicate something was done correctly. With Zen+, the numbers were 3466MHz, Zen2 reached 3733MHz, and it wasn't clear what Zen3 was. That explains why speeds around 3200MHz and 3600MHz were widely used, while 3433/3800MHz weren't as common, even if the clock frequency could support them. At those speeds, performance dropped significantly and the benefits only became noticeable at 4200MHz.
The f/m/u clocks relate to how quickly data moves between the I/O die and core dies. This is why the fclock typically matches half the ram speed—it reflects the round-trip time. For example, a 3600MHz RAM on a Zen2 would have an fclock of 1800, meaning one way at a 1:1 ratio. At 3800MHz, an effective fclock of 950 (or 1900 at a 2:1 ratio) might apply on a Zen2 CPU.
I'm not entirely sure, but I think some adjustments were made for Zen3, so it might not be relevant. Still, using higher voltages and heat through RAM for minor gains doesn't really help unless you're testing on a bench.