Crap

You know what sucks? Getting all your fans in your system only to realize that they are all backwards. Especially when one fan is contained in an apparatus that requires a pretty complicated removal process.

On the bright side though, I'm getting my 8800GTX today.

Abit Quad GT: So far, so good

First off all, the board is perty. It has a nice dark blue PCB and gives off a nifty little light show. I opted to ditch the somewhat gaudy CoolSleeves I had over my wires, and instead clean up the interior and go for a more understated, clean wiring look as you can see here:



Anyway, right now I've got the board running at 460mhz FSB, putting my processor at 3.22ghz. Should nicely compliment that speedy 8800GTS. I did update the BIOS so I do hope to hit even higher speeds, but this is undoubtedly already superior to the 680i. I like the uGuru utility, I positively *love* the clear CMOS switch on the I/O port, and overall my impressions are very positive.

The downside is that I've hit a snag with the IDE controller. I didn't realize the IDE and RAID controllers were the same, so I uninstalled the RAID controller (since I have no use for it) and now I can't get the OS to detect my IDE drive. Annoying. I hope to have it fixed quickly.

More headway on 8800GTS overclocking

Man, I think I'm going to have to take back the stuff I said about liquid cooling graphics cards not being a big deal for overclocking. My 8800GTS is now stable at... get ready... 690mhz core! How did I pull this off ladies and gents (aside from liquid cooling the core)? BIOS overclocking.

See, when you overclock in software (Rivatuner, nTune, etc.), the shader clock automatically increases concurrently with the core clock. The problem is that the shaders often hit a wall before the core does. But if you overclock in BIOS with NiBiTor, you can cap the shader clock and keep increasing the core.

The downside to this is that you have to re-burn a bootable CD every time you want to bump up the core clocks, but it's hard to argue with the end result. I found this really interesting because originally, I had tried 690mhz and it totally bombed in ATI Tool – froze the program every time. But when I used NiBiTor and capped the shader clock, I was able to hit 690mhz core no problem, and my shader clock is still the same as an 8800 Ultra.

Sweet!!!

Well... it's been a royal pain in the butt, but I've finally made some changes to my 8800GTS and I'm really happy with it all.

I needed to flash the BIOS of the card because last time I did it, I accidentally disabled temperature monitoring. But, my floppy drive is still not working, so I had to figure out how to make a bootable CD and get NVFlash to work on it. Let's just say it was a royal pain in the butt!

But after a few hours of confusion and headaches, I finally got it working. Now, overclocking through NiBiTor allows you to do something in the 8800 series you just plain can't do when you overclock in software via Rivatuner et al... you can manually set the shader clocks!

So, I bumped the shader clock from 1300mhz up to 1500, same as the 8800 Ultra. Obviously I don't expect to get quite the performance of that card, but I'm sure it's close enough that I don't need to sweat the $430 price difference. Anyway, current clocks on the card are 675 core, 1500 shader, and 975 memory. I actually did try bumping the shader clock up to 1600. I've found that you get roughly 1 point in 3DMark06 for every bump in the shader clock. So 1600 added about 97 points to my score, but it bombed in ATI Tool. No dice.

Oh, and by the way... I also took note of the card temperatures for the first time since I got my water cooling setup... with air cooling, the card had been pushing 80°. Now under full load the highest it got was 55°. Not bad! I expect the temps to be lower tomorrow because I'm actually going to swap my fans around. See, I looked up the specs on my stock Silverstone/Everflow case fans and found that they're the same RPM as the Arctic Cooling fans that cool my radiator, but they move a fair bit more air. So, I'm going to swap it around so that the Everflow fans are cooling the radiator, which should give me a nice temperature drop with no added noise.

Looking forward to tomorrow, when my new motherboard arrives. I just sold the EVGA 680i, and more than covered the cost of my new board in the process.

Slight adjustment to the GTS

Well, I scaled back my memory overclock to 975mhz, because I tend to see errors beyond that. However, it occurred to me that I had never really bothered seeing if the core would go higher. ATI Tool would detect 663mhz as its max clock, but was it really? Answer: Nope! Currently the card is stable in ATI Tool's algorithm test for 10 minutes (in my experience, that's more than long enough to allow it to get to full heat and detect any artifacting), with a core speed of 675mhz. Sweet! I'm going to keep testing and will update the blog accordingly. I did already try it out at 700mhz just for kicks, and although it passed in Rivatuner it totally bombed in ATI Tool. I'm thinking I might be able to hit around 685 or even 690. How crazy is that?

Fuggedaboudit

Geez, I tell you, this has definitely been a ricockulously frustrating build, and the frustration continues.

I received my replacement 680i board, and not only do I still have the FDD problem (it prompts me to insert a disk even when there's one in there), but this board is even less stable and a less impressive overclocker than the previous one. I've been unable to keep this board stable past 400mhz FSB. And I just recently had some problems at that speed, though it's possible some tweaking will solve the problem. On top of that, I was getting some "CPU squeal" under full load, which is apparently a common issue with this board but it's annoying nonetheless.

Well, dammit, I am not building this rig to be mostly happy with it. I want to positively love everything about this rig. So, this motherboard is up for sale and I'm heading over to Intel's p965 chipset with the Abit Quad GT, which I've heard nothing but praise over, particularly with their newer BIOS revision. I have zero interest in SLI anyway (because it sucks), so I don't know what I'd need an SLI board for. The Abit is headed to me rush delivery and should be at my door by Tuesday.

I'm also putting up my sound card (Creative X-Fi Xtreme Music) for sale. It's a decent card, but the lack of a connector for a front audio headphone jack is an issue, and frankly Intel HD audio sounds pretty decent. I will probably spring for a new soundcard before too long, but not until I've sold all the other stuff I want to.

New motherboard... again...

I'm getting a new motherboard, again. Except this time it's just a replacement 680i, courtesy of EVGA at no cost to yours truly (save the time of swapping out the boards, which is a pain in the ass). It's being replaced due to, of all the niggling things, a faulty floppy controller. Now, this is not a real big deal and by itself I would probably just keep the board and ditch the floppy. But, this board is a bit of an overclocking lemon compared to what I've come to expect from EVGA's 680i boards.

Recently, overclocking stability has gotten worse. I tried to push the board farther after getting my new RAM, and now it won't even stay stable at 400fsb, so I have the processor overclocked to a paltry 2.4ghz. Granted, it's still a fast chip, and there was a day when a 600mhz overclock would be considered pretty incredible, but this chip can do better. Hopefully the new board EVGA sends me will live up to the 680i's reputation as a leading overclocker.

Bigger is better

I just added the newest addition to my PC, a Samsung SyncMaster 226BW 22" widescreen monitor. To say I love it would be the understatement of the century. It's freakin' gorgeous! Great picture quality, great contrast, and I can finally agree that widescreen gaming really is much more immersive than standard aspect gaming.

I also received my new RAM today, and got my rig overclocked back up to 3.157ghz. Prime is running now (has been for roughly an hour) and it looks good so far. Interestingly, I've been able to get higher boots and better stability with lower voltages. If you leave the overclocked settings to Auto, it increases the voltage to 1.4v for the FSB, and the board wouldn't consistently boot at 450mhz. But I reduced the voltage to 1.3v, and it's rock-solid. Weird.

Still got great temperatures, too, and this is with the insanely quiet, low-speed Arctic Cooling 12025Ls. Using Core Temp (same program used by [H]ardOCP), idle is around 29°, load in Prime95's "maximum heat and power" torture test is 41°. Not too shabby if I do say so myself.