Originally Posted by Wutface
Ok. One thing at a time is a little more difficult than one thing at a time. I think it's because I had so many settings outside of (raise multi and voltage herp) in my initial picture that people assumed I shotgunned the settings and hoped for the best. That was and is still my best stable overclock. I didn't just pull all of those settings out of my ass and say look what I can do. I tried to figure things out based on previous knowledge and information I found here. I'm a little frustrated with myself for not figuring out something that is going to be completely obvious once I've finally managed 3.8 or 4.0GHz stable.
My question I guess would be this, since RAM is something I'm least knowledgeable about and I believe it may be the problem:
I have 2x 2G Corsair Vengeance RAM (4G total) that came together. They are rated at 9-9-9-24 @ 1.5V 1600frequency.
If I run a memtest with everything
else set to stock speeds, and they pass at 1600 9-9-9-24 @1.5V...
1) Would they still be stable at higher clock speeds assuming I solely used the multiplier to raise the clock speed.
No. In some systems, the RAM requires additional voltage to maintain a CPU overclock. Remember, the faster it goes, the more information travels between the CPU, Cache and RAM. One or more components might need additional voltage to maintain a strong, steady signal. Else it BSODs or causes file corruption.
2) What about if I raised the northbridge frequency?
Same as 1. By increasing the NB frequency, you're overclocking the L3 cache. Information can now pass out of it much faster, and as such, might require additional voltage to stabilize other components. Almost always, it's DRAM voltage.
3) If I drop the frequency down to 1333, and keep them at stock 9-9-9-24 @ 1.5V, should I find the tightest timings with everything else at stock, or wait until I get stable at my intended goal of 3.8 or 4.0Ghz?
Dropping the memory down to 1333 would be a good idea while attempting to overclock. This will almost assuredly rule out memory as an issue further down the line. You can increase it back to 1600 when you hit the wall on the CPU.
4) What would be the recommended order of operations to ensure my ram is running stably at higher clock speeds?
If it passes a Small or Large In-place FFT, but fails a blend or IBT, then it's RAM. That's the best way to isolate the component. If it fails Large In-place, but passes Small, it's IMC, usually.
My system will
boot into windows up until 4.3 by just raising the voltage. It will overclock to 3.6GHz by just raising the multiplier. I am hard pressed to believe that I would need to increase the voltage by 0.1V to get it stable at 3.7GHz. My temperatures are fine even when I solely raised the voltages to settings beyond what I was comfortable with.
Thank you for your time and consideration~
Some processor's just don't overclock well. It's like anything else, sometimes you get a good one and sometimes you don't. Unless you know someone with another motherboard or ram to swap components, it's going to be almost impossible to tell.