Yes, the core readings all stayed within 1 degree C of each other as they went up and down.I'm guessing ambient temp right now is in the mid to upper 60's Farenheit (18-20 C according to the web converter I used). We do not have air conditioning, and the room with the computer faces south, so it's much warmer in the middle of the afternoon (it's 4:30AM here now). I'm thinking that it will be best to reserve serious FlightSimming for after sundown once I've got my overclock set up. And the BIOS does let me set the shutdown temp to 60 degrees C for safety's sake.Now that I've made the effort to move the computer (with a lot of help from my wife re-routing cables, I'm going to leave it where it is.So if my cooling is in good shape now, how do I proceed with overclocking? I am willing to absorb as much information as needed for me to understand the concepts before moving ahead, but I don't really know the best place to find that information.Once again, thanks for all your help in getting me this far on the road to my goal.

Chech this outThat's the guide I followed with my 1090T and it was the first time I overclockedWe'll need your motherboard's manual too

Cool, thanks for the overclock link and mobo manual!I will read the overclock pages after I get some sleep.What do you think about Jahman's suggestion of a heatsink with 6 heat pipes?

Cool, thanks for the overclock link and mobo manual!I will read the overclock pages after I get some sleep.What do you think about Jahman's suggestion of a heatsink with 6 heat pipes?

I agree, of course usually more pipes means better performance, but that doesn't tell the whole story. Your cooler is still very good, considered the best price/performance ratio HSF in the market and should be capable of getting you to 3.8 maybe 4GHz

I read through Dolk's Guide carefully. I increased my multiplier (in the BIOS Core FID field) to 18 with no other changes, and ran prime95 for a few minutes with new errors. Temperatures stabilized at values still in the upper 40s. Then I bought some Cooler Master Excalibur fans to replace my four 120mm fans, and my temps are down a bit (especially under light loads). I increased my CPU multiplier in up to 19 (3.8 GHz), and ran a longer stability test (nearly an hour). I did see in Dolk's Guide that an increase to the frequency for the CPU-NB and HTT was necessary, and he recommended using a multiplier increase for those. My BIOS does not have a multiplier for CPU-NB and HTT, so I increased the frequencies directly to 2.4 GHz. At that frequency, Dolkrecommended a voltage of 1.175, so I set that for CPU-NB. I didn't see any way in my BIOS to adjust the HTT voltage. If I read Dolk's guide correctly, the next adjustment that may be needed is an increase in the CPU voltage (if stability tests fail while cores are at reasonable temps). Please let me know if I'm making any mistakes here. Even at 3.8 GHz, I have a performance gain of almost 20%. And max core temp after the long stability test was 50 C.

You're doing great man.A couple of things: when you say you increased the frequencies directly to 2.4 GHz, you mean the NB FID right? I would say that exactly the same thing as upping the NB multi, if there was one. If you increased the HT link, that's not right and you should revert that to stock (2000Mhz)... just in case hehe. The HT link affects your memory, NB, and CPU clocks, and you don't want that. They have their own multipliers and you should OC them in isolation, one at a time. What you call the HTT voltage should be the NB voltage (CPU-NB voltage in the guide) and you can find it in your BIOS right there, with the rest of the voltage/frequency settings. I think it's the "NB VID" Can you post a creenshot of CPU-Z, including the CPU, Memory and SPD tabs, while Prime95 is running please?

Thanks for the "doing great man." Thanks, also, for the correction regarding HT. I had increased it, but now it's back to stock (2000MHz). I even had a go at trying 4.0 GHz, but that made things VERY unstable. I'm back to 3.9 GHz currently, and stability has (for a 15-minute stress test) returned. Here is the CPU-Z screenshot with CPU, Memory and SPD tabs:

Modified CPU-Z screenshots (hopefully more readable):

Well, you still have some headroom there with Vcore. If you have good temps and want to give 4GHz a shot again, you can bump your Vcore a bit more, to say 1.45V. Mind you, the important Vcore is the one you see under load in CPU-Z, the ACTUAL voltage running through your chip. what you set in BIOS is irrelevant.So you should increase your Vcore until you have 1.45V in CPU-Z, if you don't want to settle for 3.9GHz, that is What's your NB voltage and Vcore in BIOS right now?

NB voltage in BIOS was set to 1.25V, Vcore to 1.40V. Bumped Vcore to 1.4375 in BIOS, resulting in 1.440-1.452 (switches back and forth) in CPU-Z. Bumped Core FID back to 4.0 and things seem more stable there now. So, can I squeeze some more performance out of my machine by adjusting memory timings (one at a time, as you said)? Might as well do all I can, since I'm having some success so far. Time for some more sleep now. I'll keep the machine off while I do that, just in case. Then some more stress testing at 4.0 GHz when I wake up.

NB voltage in BIOS was set to 1.25V, Vcore to 1.40V. Bumped Vcore to 1.4375 in BIOS, resulting in 1.440-1.452 (switches back and forth) in CPU-Z. Bumped Core FID back to 4.0 and things seem more stable there now. So, can I squeeze some more performance out of my machine by adjusting memory timings (one at a time, as you said)? Might as well do all I can, since I'm having some success so far. Time for some more sleep now. I'll keep the machine off while I do that, just in case. Then some more stress testing at 4.0 GHz when I wake up.

Not really. From the SPD tab of CPU-Z it's obvious you are already running your RAM at it's rated speed and timmings. You could try to squeeze a couple MHz out of it, but the impact in performance is going to be minimal, and memory OC can easily lead to stability issues. Not worth it IMO. If you want to try, go ahead. There's not a huge risk in it, just not worth the hassle really

Okay, I'll leave the memory timings where they are. I'm currently running a stress test, and core temps have maxed at 59 C. I have the shutdown temp set at 60, so I've stopped the stress test. HSF is still pointing upward - is it time to reseat it facing backward?

Okay, I'll leave the memory timings where they are. I'm currently running a stress test, and core temps have maxed at 59 C. I have the shutdown temp set at 60, so I've stopped the stress test. HSF is still pointing upward - is it time to reseat it facing backward?

59ºC in Prime95 is perfectly fine. You're not going to fry your CPU inmediately if you hit 62ºC or anything like that, that's the máximum recommended by AMD and it shouldn't run at 62ºC or more for long periods of time. The CPU has it's own built in thermal protections and will shut down if it gets too hot anyway, and that happens at higher temps, some 70 - 80ºC off the top of mi mind. It's obviously not a good idea to push it that far, I'm just saying that you're not in the brink of the CPU's melt point at all just because it's getting close to those 62ºC. Prime stresses the chip a lot more than everyday apps. If you fire up FSX you should see temps in the high 40's / low 50's.What I do is to set speedfan/core temp to display the temp in the system tray, and a warning if temps get too high, shutting it down altogether at 62ºC is a bit too drastic IMO, but to each his own. About reorienting the HSF, I think you can squeeze a couple ºC doing it. Hard to tell though. The way it is now, the intake is right on top of the GPU. If you run something that uses both the GPU and the CPU, the airflow through the HSF could be significanltly hotter, thus making it less effective. You could try running Prime95 at a lower CPU speed (3.6 / 3.8GHz) so that you are not to close to the max temp for the CPU, and then start furmark or other GPU stress test utility and see if your CPU temps are affected and how much.

Good point about Prime95 stressing the CPU more than everyday apps. For the time being, I think I will put core temp in the systray. Temps are in the upper 40s when running FSX, so I'm in good shape there. Thanks for explaining the strategy about the HSF orientation. I think that reorienting the HSF is probably a good idea overall. Now for the real interesting part about my outcome so far: I am seeing significant reductions in CPU load when running FSX for cores 1, 2, and 3. But load for core 0 remains pegged at 100%, even with the AffinityMask set to 15. Frame rates have gone up, but not above 20 (with frame rate limit set to unlimited). I had read some posts about FSX not really using multiple cores very effictively. Is that the reason for what I'm seeing, or is there something else causing what I'm seeing?

Now for the real interesting part about my outcome so far: I am seeing significant reductions in CPU load when running FSX for cores 1, 2, and 3. But load for core 0 remains pegged at 100%, even with the AffinityMask set to 15. Frame rates have gone up, but not above 20 (with frame rate limit set to unlimited). I had read some posts about FSX not really using multiple cores very effictively. Is that the reason for what I'm seeing, or is there something else causing what I'm seeing?

That's normal. FSX runs in a single thread in core 0 by default, then it spawns threads in the remaining cores to load textures. The texture threads were introduced in SP1 and help with performance in single and dual core CPU's, but once you have 3 cores, adding more has virtually no impact in frame rates.Actually most of us run FSX on 3 cores using the AffinityMask tweak in the C:\Users\[yourUserName]\AppData\Roaming\Microsoft\FSX\FSX.CFG. This tweak and other useful ones can be applied using this http://www.venetubo.com/fsx.htmlGive it a try and see if it provides a perf boost, it's worked wonders for many of us. Are you familiar with the FSX.CFG tweaks?There's one you can try after that that could boost you FPS even more, but causes stability issues sometimes, not system stability issues, just FSX crashes. If it doesn't work you can always revert it back to where it was working ok. After you run the utility in that link, you'll find a secction added in your FSX.CFG that will look like this:

[bUFFERPOOLS]RejectThreshold=131072

the RejectThreshold value might be different. Change it to...

[bUFFERPOOLS]UsePools=0

If it's not stable, set it back to

[bUFFERPOOLS]RejectThreshold=xxxxxx

with the value suggested by the web app. Your temps are good in FSX and your GPU doesn't run too hot at all (it shouldn't considering it's a 6790) so you're all set for all I know. Nice 4GHz OC on a dirt cheap cooler hehe.Just one thing, what's the Vcore like in CPU-Z while running FSX? There's something called Vdroop that we should keep in check