Is My RAM Overclock Stable/Safe? Some Concerns.
Is My RAM Overclock Stable/Safe? Some Concerns.
I recently adjusted the RAM overclock settings in my system without altering any voltage parameters, which will remain consistent for all components moving forward. I have a few queries regarding this change.
My main concern is about the timing values. Presently, I'm using 4 sticks of 4GB Corsair Vengeance Blue CAS 9 1600Mhz RAM, overclocked to 2000Mhz with a CAS of 10. My XMP settings were 9-9-9-24, but after the overclock, they changed to 10-11-11-26. I'm unsure about the TRAS setting. I've read online that it should be roughly the sum of the other three timings. Is this advice accurate? Shouldn't I just push all timing parameters to their maximum without any adjustments?
Additionally, I completed two full passes in MemTest86 with no errors and ran three Hyper PI tests without encountering any warnings. There have been no signs of performance drops or issues when using the system. I'm curious whether these results confirm that my setup remains stable and secure. I've heard that incorrect timing can potentially cause OS problems, so I want to be confident in my configuration.
Memtest isn't really effective for stressing, it's more useful for checking if a faulty module exists rather than for causing stress. I'd run Prime95 for at least an hour on blended if it performs well, but if you're concerned, try a 32M Hyper Pi since you have it.
Isn't it just safe to adjust all timings as much as possible without errors? Yes, but your experience with sub timings like tRAS won't accurately reflect real-world performance. It's better to test using GeeKbench or similar tools to verify actual improvements compared to the stock version and specifically with tRAS settings.
It should be completely safe at the stock voltage, though I’m not sure how much above 1.5v is secure for DDR3—I won’t comment on that. For stability, running two passes of memtest seems appropriate; I haven’t used Hyper Pie before, so I usually stick with Super Pi 32 million, but if it behaves similarly to Super Pi, it would be a good test. Regarding TRAS, it doesn’t make sense for DDR4 to sum all three numbers together, as that would exceed the XMP profile limits I set. So I’d say no, it doesn’t need to match the first three values, although I’ve heard similar advice before. Generally, I focus on tightening primary timings and leaving secondary ones alone.
JamesSneed shares his experience with different testing methods. He finds MemTest useful for checking module health without causing stress, suggesting Prime95 or Hyper Pi for better stability. He emphasizes the importance of real-world performance tests using benchmarks like GeeKbench. James also mentions using Windows at 2133 MHz with relaxed timings, but Hyper Pi can still identify issues that MemTest misses. He tested at 2000 MHz in BIOs and ran Pi three times with 32M and 4 cores without problems. He mainly relies on games and computing for stability checks, noting Prime pushes his system further than expected.
The system should remain safe under the stock voltage, though I’m uncertain about the exact limit beyond 1.5v for DDR3—I won’t discuss that further. For stability, running two passes of memtest should suffice; I haven’t used Hyper Pi before, but Super Pi 32 million works well. I’ve also heard similar advice for TRAS, noting that adding all three numbers together doesn’t make sense and isn’t necessary. Generally, I focus on tightening primary timings as much as possible, leaving secondary timings alone unless needed. Regarding temperature, I haven’t changed the voltages, but I’m aware it’s often overlooked. It seems unnecessary to over-tighten unless you notice errors or instability. For better memory testing, especially with dual or quad cores, Hyper Pi is recommended—it quickly detects issues and may be more efficient than the current Super Pi version.