). The only funny thing I find is how small the motherboard looks in that case. Here, I am always short on full ATX cases (and all cases, in general), so any full ATX case I get always receives a full ATX mobo.
). The only funny thing I find is how small the motherboard looks in that case. Here, I am always short on full ATX cases (and all cases, in general), so any full ATX case I get always receives a full ATX mobo.
i think thats your problem! when running four sticks, the command rate must always be 2t or else the mch cant handle the addressing required for four sticks and will crap out.
i think thats your problem! when running four sticks, the command rate must always be 2t or else the mch cant handle the addressing required for four sticks and will crap out. addressing lots of memory modules coupled with the high density of 2 gb sticks will overload the mch and it wont be stable coping with the demands.
)
) I removed that “new” 4th RAM stick, pressed power button again, and PC POSTed fine. … here are the final specs of this PC that I think will stay like this for a while:
) – clack-y and tight, with deep keystrokes. Decently heavy too. ) ) I removed that “new” 4th RAM stick, pressed power button again, and PC POSTed fine.
Or perhaps I have to raise the RAM voltage? 
Instead, I set the FSB to 1300 MHz (from 1333 previously, since the board let’s me choose any value I want – at least between 400 and 1600 MHz, that is.) This lowered my CPU over-clock only slightly to 2.925 GHz instead of 3 GHz. I figured it shouldn’t make much of a difference in performance but might make a slight difference for stability purposes. (?) 
Made me wish the system would just explode into a fireball so at least I would know it’s dead. 
albeit with CMOS settings all lost for whatever reason. 
Seriously! (Except only if I run into issues again.) As it currently stands, the PC has been running for over a week now with the 8 GB memory config and no problems. Again, aside from YouTube giving me BSODs, the system has been pretty stable otherwise. Or is it? 
Time will tell with more testing. As I said, I’ve been using it mostly just to play games on every other evening or so. Haven’t tried it much with any other “daily” tasks, so we will see how it goes.
And what exactly? – The PSU or the GPU. Got normal picture on the screen and PC proceeded to boot to Windows just fine. PSU voltages were same as before too.
Time for another round then. 
But apart from that, everything else is plugged in properly. I even re-rerouted and tidied-up the front panel cables a bit. Of course, this being a non-modular PSU, there's always going to be that mess of cables in the HDD cage area. But again, I'm not after building a great-looking system. So all of the cable organization I did was purely for fun. And in terms of power draw, the Antec drew only about 5-10 Watts higher from the wall than the Enermax – at least at the same tests that I could run with the Enermax where it didn't shut down. This seems to make sense, as the APFC circuit in the Antec should actually make the PSU slightly more inefficient. But where I lost a little bit of power, I gained in near unity PFC (which I'm sure the power company will appreciate). And on the OCCT PSU test, the power draw was somewhere around 340-350 Watts, IIRC – right what I expected it to be with that CPU and GPU (and the silly hot-running 780i-SLI chipset.)
Unfortunately, that was all - it crashed as soon as I got to the Login screen (it did manage to load Windows, though.) That did it! She was stable in OCCT CPU test and OCCT LinPack CPU test. But just for a good measure, I bumped the CPU V_core up one more notch to 1.23125V. And now we were in business 
- no crashes in OCCT or games even after several 2-4 hour-long sessions in Fortnite, CS:GO, and a few older games like Mirror's Edge and Test Drive Unlimited. The only crashes I got (and I think I have this ruled out... but still verifying) is from the GPU dropping out at stock clocks, which I mentioned in that link I posted in my previous post above. ) And "CPU Core" temperature was just the average of CPU cores #0-3. Thus, I decided to take the "worst case" temperature, which was either CPU Core #0 or 1, and hence the 60°C max figure above. I suppose I did a decent job with that CPU fan curve. 
(It really is, though - it has tweaks for just about every voltage and clock one could wish for.) 
I turned off power and inspected my CPU cooler... particularly my ghetto mounting with the wooden thumb screws - were they not holding the cooler on properly? After thorough inspection, I determined that wasn't the problem. I then started thinking more: perhaps it's the cheap thermal compound I used? (Some eBay 30 gram tube for under $3 shipped. Makes me wonder sometimes if it even is thermal compound! ... but it really is, though - I've been using it on large chips with heatspreaders and low-power GPUs for a few years now, and it's never let me down.) But that didn't make sense either, as 105W TDP CPU with 150W max dissipation isn't something this thermal compound hasn't handled before (*looks at Pentium D830 CPU on shelf* ) I then started checking more things in BIOS and eventually found this: the FSB was set to 1600 MHz, making the CPU core clock @ 3.6 Ghz!!! Well, DUH, no wonder why that CPU fan was running full blast. From stock 2.4 GHz to 3.6 Ghz - that's some serious OC there! I'm surprised it even POSTed every time without any issues.
On that note, I wonder if that's what made the "PSU blow up" on the previous owner. Perhaps he/she tried to OC their CPU and were using a crappy no-name PSU that gave up the ghosts? Ah, who knows. 
The interesting thing is that the AMD PC mentioned above does not crash. Instead, the video continues to play after the short 0.5 to 1 second sound looping. But it happens completely randomly in time and never at the same spots in a given video. - In all of the 60 or so minutes it took me. Most of that went towards making the wooden thumb screws for the CPU HS. Otherwise, the OS install went pretty quickly to... Windows 7 Pro, by the way (as I had a COA sticker with that from another PC).
Leave a comment: