Increasing RAM speed requires additional power.
Increasing RAM speed requires additional power.
Hey there. You're asking about RAM overclocking, specifically how to push your memory beyond its standard 3600MHz XMP limit. After testing, you found that running the speed up to 3700MHz without errors was safe, but you wanted to see if going higher would help stability. You also tried 3800MHz and encountered a no-POST boot issue, which made you wonder about voltage adjustments. It seems you're considering increasing the voltage from the XMP profile, but you're unsure whether that would resolve the problem or if it points to a hardware limitation. You might want to check the recommended voltage for your RAM type and see if a higher value could improve performance without causing crashes.
I would have simply set the voltage to 1.5 or 1.7 right away and checked how quickly the RAM would start. I think the BIOS chip is inside the PC on your end, and 4x8 will block any chance of success, so I’d just try raising the imc voltage and I believe boosting the system agent helps too. Just run a test at 3600-4000 with a clock speed of 14-14-14 because if you can’t reach high speeds, timing might be the issue.
I've heard that 1.5 is the upper limit, and anything above isn't advisable, so I'm only comfortable with 1.4V max. But sure, I could try pushing it up to 1.4V and observe the results. And yes, having four RAM sticks doesn't really help with overclocking... Yes, I might raise the IMC voltage, but I won't do that since my CPU is likely one of the poorest 8700K ever made—it gets super hot. I had to customize the IHS and use liquid metal to keep temperatures down.
Great feedback on Stahlmann. At 1.35V my RAM reached a peak of 51°C once it hit full coverage during the HCI MemTest Pro. From a thermal standpoint, it seems safe to raise the voltage without worrying about overheating.
I didn't test temperatures directly, but I used my memory at 1.45V for roughly two years before switching to a different setup.
It seems you're asking if raising the RAM voltage might resolve the no-POST problem.
lower than 1.5v means poor gains, i've noticed buildzoid hit 1.97v probably due to a defective die from a team member. Based on my estimate, staying under 1.7v is safer since DDR4 deathzone is around 2v. Some kits seem to push to 1.65v at stock, but I only saw Tridentz Royal 5333 at 1.6v. At 1.4v it's really limited. Check buildzoid videos if you want to properly boost those RAM speeds. Buildzoid doesn't limit voltage much, so don't follow them if you're aiming for 1.7v+ or more. That would likely damage your RAM. Above 1.75v he mentioned on the extreme overclocking team that it will cause BDs on dies with more than 8GB, and voltage matters too—no voltage = no gain potential. Stick to something that doesn't require high voltage for good scaling, like 2.9v at best, but be careful since DDR2 is already in its deathzone range. I might try 2.9v later if needed, but right now it's probably safer around 1.4v.
Great details provided. No prior overclocking experience, preferring a reliable upgrade rather than expensive new parts. Planning to go with a 3700 MHz RAM kit at 16, 16, 16, 36 MHz clock speed and 1.35V power. Seems like a solid choice based on current benchmarks, and I appreciate the focus on efficient performance without breaking the bank.
Silently hesitates under poor output Its not that risky to crank it up to 1.5v usually a safe setting is around -0.4v, I’m okay with pushing it all the way to -0.3v from there but I prefer a bit more speed if possible. If you need a general safe range just check what the factory specs say for your RAM model like the Tridentz Royal 5333 which runs at 1.6v—it should be fine. Running it at the stock voltage keeps things stable, though raising the IMC voltage can boost performance but may stress the component.