Originally Posted by ozzy983
That's exactly what it is. If you test each component and isolate it, seeing what it's maximum is, then juicing it up a bit with the HTT will give you the highest possible overclock. That's not to say, if you are running at 250 HTT that somehow you'll get more out of your CPU, but multi might only give you 4.3ghz, and the chip might flake out at 4.360.
The system does seem a tad bit snazzier when I have it set to a high HTT; placebo or not, I'm not 100% sure. The benchmarks seem to indicate it as well, but the difference is so minute that it's not worth the additional headaches.
My system is weird too, because if I drop the ram multi to 4x (800) and jack the FSB to say 250, it doesn't work nearly as well. It's also rated to run 2.3V @ 1066 @ 5-5-5-15-2. I'm not sure where they got their "ratings" from but I run it @ 1066 @ 5-5-5-15-2 @ 1.9V, lol. Anything higher than that and it gets funny, which I think is a problem with ddr2 imc to begin with.
I wish it were that simple. I test every component...but every time I change something, 'stable' maxes change. For instance, I can run say 4GHz at 1.45v and 200MHz x 20. But lets say I go to 250 core clock so I can OC my RAM and change the multiplier to match to 4GHz, now it's not stable at all, even at 1.5v. Or I change nothing but the NB to 3000MHz and it's fine. I change the CPU up and NB is no longer stable. I wish it were as easy as finding the max of each component individually and throwing them all together...but they all seem to effect each other in a multitude of ways. So far I know I can run 3GHz NB stable by itself, but once I start changing other things, that changes. And with RAM, too, some frequencies are stable with some settings and not with others, and not necissarily what you'd think, i.e. higher or lower. Weird combinations...
On a side note I had a nightmare tonight when testing for max RAM clocks by changing the core clock. I started at 260 core/2080Mhz and worked my way all the way down to 220/1760. At this point, as to be expected, I had some really weird stuff going on when the system would hiccup from unstable RAM speeds, like it would POST and boot but restart once it got to the welcome screen, etc. One of the weird things: Windows 7 gives the option to do a 'startup repair' if it detects the system didn't boot properly, and it goes into a recovery console. Sometimes, when I choose to "Start Windows Normally", it shows the boot animation for recovery console but boots into Win 7 normally.
Well, eventually it just restarted at welcome screen every single time..no matter what my settings were. Safe Mode worked, but nothing else. CMOS clear, etc etc, you name it. System would not boot from testing RAM speeds...nightmare. Thought I needed new RAM or corrupted the OS somehow. Eventually I got back into Windows by using F8 and choosing "Last know good configuration", but I got an 'unknown device could not be installed' error, and could not connect to the internet.
After some investigation, I found that critical Windows services were not starting and I could not start them manually, nor could Windows (access denied). I ended up in the registry and all that ****...HKLM/CurrentControlSet/Services and changed permissions to allow all for all everything...cuz I was desperate. I finally figured out what happened: Windows booted using it's "Selective Startup" rules from either the Recovery Console or Safe Mode in the NORMAL operating system environment, and rewrote access privileges and service startup rules for everything in the OS onto the normal boot routine when it started up one of those times. Ended up in msconfig and all I had to do was uncheck "Selective Startup" and recheck normal startup. I want my 3 hours back
tl;dr of the above: crashing your computer over and over again testing components can eff you over big time. Felt very tech savvy figuring out this problem but wasn't worth the time spent
Also, for those of you that have made it this far, I have a question: it's about vBoost (I think I've heard it called vDrop as opposed to vDroop). I did 4GHz @ 200 x 20, and I ended up putting my vCore all the way up to about 1.45 and it still wasn't stable (*** mate.). But when I would run prime95, or any load, my vCore would slowly rise from ~1.46v at idle to 1.48+. I saw it go from 1.43 (1.425 in bios) to 1.48v. Why is my system adding .05v+ to my bios selection? that seems insane. I've disabled all the obvious power "features" that I know of, CnQ, Turbo, C1E, LLC. Should I turn them back on...? This kind of jump is common across almost all my vCore voltages, except some sweet spots like 1.275 which varies a lot less, etc.
/wall of text