Be Excellent To Each Other

"how do you take your tea?"
"in the bedroom"

Sounds like the psu to me, dodgy gpu rails.

Or the RAM. Or the case.

Or the screws holding the case together. Hmmm.

Glued computers are far superior.

Jittery artefacts in yo gwafix is usually a sign of a graphics card that is overheating or fucked. You're best sending it back for a replacement.

Also, based on what you're saying, it sounds like you struggled to fit it into your case, which will almost certainly affect air flow and cause the card to overheat. You need a bigger case! You tried running it with the side off the case? (The hood popped, yo.)

I'd say it's almost certainly overheating because as you say, you played half an hour of something, then it started happening 10 minutes into Burny Pee.

If you got a new case and transferred everything over, the card would probably be fine. But it might be worth sending it back for a new one anyway in case the overheating has done any permanent damage.

The case already seems fucking huge, it's kind of galling that I might need something even bigger (although I recognise that's a possibility); it's this one if anyone cares. Hmm, actually, looking at that page it is listed as a 'mid' case rather than 'full' so maybe it really isn't designed for the massively big card I'm cramming in it. I did have to take the funnel thing off the side of the door to fit a new CPU cooler on ages ago, although I assumed the funnel thing was gimmicky bullshit anyway. Oh fuck, I've no idea!

Anyway, I'll run it for a while with the side of the case off and see if that makes any difference. If so then I suppose I'll be in the market for a new case! Although, if it's a pure overheating issue, why did Portal show up graphical flaws the instant I launched it when everything was just sitting at idle temp?

Also, I only have the vaguest understanding of what this means; is there something I should/can check to rule out an issue with the PSU?

Just so you're aware "mid" and "full" is a height difference. You could eat your dinner off the top of a "full" case. They're more suited to home servers and stuff where you need lots of vertical room to stack myriad hard drives in there.

I had a similar problem when I built my current PC. I bought an expensive Lian Li case which is a chunky "mid" sized case and probably the biggest one I've ever had since I started building PCs. But graphics cards are so bloody wide now that it barely fits in.

Basically, look for a wider case than what you have! Or a narrower graphics card...

I'm not really up on current PC tech, but as I understand it, the cards were getting absolutely massive over the past few years, but more recently a new breakthrough has allowed for a newer generation of smaller processors whilst retaining/bettering the power of the older large ones, so maybe a slightly newer card might be an idea? Of course, I just read that somewhere and as I say I'm not really clued up these days. Also, I don't know if your card is a new model or a little older, I'm just speculating!

edit: Re the "full" and "mid" thing, looking on the Coolermaster website, they seem to think that "full" is just a slightly bigger "mid" case, so you're right, your case is probably a bit "small". Not sure why they've chosen that nomenclature when it traditionally means something else.

I doubt this is the issue based on what you've described to be honest. It's a possibility, but it sounds like a very simple case of a powerful card overheating in a cramped case to me.

A photo of the inside would be potentially useful in diagnosing whether the layout represents a likely airflow issue.

I bet you say that to all the girls.

If cooling is an issue, move to the Antarctic.

Also, add more LEDs to draw the attention of the heat away.

"Bend over."

WTB: the card I've got is this one, so coming up for a year since it was released and certainly the biggest one I've ever handled (as it were).

Craster: roger that, I'll post up a pic tonight.

Is anyone else as amused as I am by the whole Crossfire thing?

"We need to sell more graphics cards!"
"how are we going to do that?"
"People only need one graphics card per machine"
"Unless..."

6 months later:

"A hahahaha! they fell for it! trebles all around!"

It's a little fatty! I'm not sure what sort of cases they expect people to have...

So is it tight up against the side of the case? Suppose we could just wait for the picture! Anyway, if it does turn out to be a bit tight in there, try running it with the side off for a while and see if that helps. Might be a decent temporary solution until you get a new case or whatever.

Windows Vista / 7 GPU Temp gadget: http://nvidia-gpu-temp.en.softonic.com/

It's not tight up against the side of the case, but it's right up against the back of one of the hard drives (which are installed on the opposite side of the case) which just doesn't seem right. I'll post the pic up tonight though and folks can see what they think.

I think it's actually a really good idea. Generally speaking when your gfx card gets a bit too long in the tooth then you need to replace it entirely with a new one, but with Crossfire/SLI you can bolster the original card with a second one of the same type. So not only are you getting to carry on wringing benefit from your initial investment in the original card, but a second card of the same type should now be a lot cheaper than buying a new 'current' card so you get to essentially double the performance at much less than it would cost to replace one card with a new one that was twice as good (if you see what I mean). Obviously, yes, some people will actually go out and spend a fortune buying two 'current gen' cards to get ridiculous frame rates or whatever, but you could also use it to keep your upgrade costs down as well.

Trebles all around still. They clear out their stock of old, inferior cards to people Crossfiring them.

If the PSU is supplying wobbly voltage or current to the GPU's additional power plugs, non-fatal glitches might result.

I'm suggesting it as you said both the old and I-assume-brand-new(?) replacement are doing it, without the rest of the system being notably unstable.

The fact the new card is likely less of a current draw* and that it artifacts rather than falling over** adds weight to the theory***.

* Associated possible less-severe voltage sag, depending on the design of the PSU
** To the point nVidia's "rescue the card instead of blue-screening" thing kicks in
*** In my head.

Outside of gaming the machine is certainly rock-solid and, yep, both old and brand-new cards do start throwing artifacts after playing for a while. The only 'new' symptom is the new card instantly throwing up weird artifacts in certain Portal 2 graphics even though it's barely under load at the time, but I'm hoping that's just a driver glitch. If it was a PSU issue though wouldn't things be fucked right from the get-go, rather than getting progressively worse as Burnout does?

Anyway, so, the behaviour I'm seeing could equally be attributed to any one of the following issues:

• Fucked card
• Over-heating card courtesy of shit case
• Crap PSU

And the only way I can think of to properly narrow it down is to swap the gfx card into someone else's PC. Something I can't do because none of my IRL mates have decent PCs.

Joy!

How did your old card fit into the case? What were the symptoms with that one? How dusty is your case? Did this only start happening as the weather got warmer?

Sorry late to the party again just got home from work and I see the BEEX Hivers have been on the case already.

First off Bamba sorry you're having trouble, in all honesty PC building/upgrading/drivers/gaming is mostly hassle free these days, so you've been a bit unfortunate here

My thoughts:

OVERHEATING - Not likely to my mind, all modern cards have pretty good temperature control built into them, auto-throttling and suchlike to keep themselves 'safe', plus the thermal thresholds on the GPUs are insanely high anyway. Not saying it isn't overheating, but it's not my first thought though, especially since you'd expect the fan to get uncomfortably loud long before dangerous temperatures were reached.

CRAP PSU - Certainly possible, what make and model of PSU is it Bamba? The 6870 is a pretty tame card though in power requirement terms.

FUCKED CARD - Definitely possible, if you go back to the game that worked properly (Hard Metal) does that still work OK? Do certain games exhibit the same problems all the time, and other games not?

DRIVERS - You've already done the right stuff here, although with having gone from an Nvidia card to an AMD card, there's the possible for some problems there but they're a lot better then they used to be in this regard.

WHAT TO DO NEXT -

Download MSI Afterburner from here and install it - http://event.msi.com/vga/afterburner/download.htm

It'll give you a nice pretty graph panel showing you fan speed, GPU temperature, and suchlike.

Play a game in windowed mode instead of full screen, and keep an eye on the MSI Afterburner graphs at the same time, see how hot the GPU is getting, and if the artefacts correspond with certain temperature thresholds.

I recently upgraded to a GTX670 from my old GTX480, and the difference is astonishing.

The 670 brings about 50-60% extra horsepower to the table over the 480 but the card itself is physically shorter than the 480 and weighs an awful lot less, the cooler on it is whisper quiet even when it's flat out gaming (literally just a slight buzz over the noise of the case fans, where the 480 was a wailing banshee and really quite unpleasant), it uses far less power, and generates far less excess heat.

It can run BF3 at a frankly astonishing 2560x1440 with every single setting on ULTRA, and maintain over 60FPS even on the new 'Scrap Metal' map which is an absolute card-killer. (The 480 couldn't manage it at all so I had to play it scaled from 1920x1080 instead.)

I'm absolutely gob-smacked by the thing, the fabrication process is down to 28nm from 40nm which explains a chunk of it, but it's also a testament to how fucking efficient they're getting the architectures of these GPUs now.

Okay, here's what we're looking at. The gfx card is sitting pretty much right up against the top HDD, although I don't see that in itself causing any issues. Apart from the rear of the case, the only air vents are on the side panel that I've removed to take that picture so the fan on the card itself is really just blowing air against the side of the case that's right in front of it. Is that bad, I've no idea.

I'll go and run some games and take some temperature readings and whatnot and come back.

Oh, and the PSU is a WinPower ADK-700 if that means anything to anyone.

Winpower? That came with the case, didn't it?

Believe it or not it's the second PSU for that machine after the first one died and Cyberpower replaced it in warranty. The first one must've been something pretty no-name as well because it's simply listed just as "600 Watts Power Supplies [+4] (CyberPower 600W Power Supply)" on the original order.

Airflow there is alright, I very, very much doubt that the card is overheating - you'd hear the fan go nuts long before it got to that stage and even then it'd throttle itself to prevent damage.

There are some low-end components in there, it's a cheap motherboard and a horrible power supply.

But that said, it should still work alright.

What is the model of motherboard and what BIOS revision is it on? Shouldn't really matter but I am curious. It looks like a cheap MSI jobby, that's for sure.

EDIT - I also note you're powering the graphics card with one 'proper' 6-pin PCIE connector and a MOLEX converter for the other, this is alright in itself but I really do have my doubts about that power supply. They probably only fitted it with one 'proper' 6-pin PCIE connector for a reason.....

It does look alright, but I'd be concerned with the position of your CPU fan which blows hot air directly into the graphics card's intake fan.

Hmm.

Tempted to side with AE on the crappy PSU side of things. People always cheap out on the power supply! And I'm not just talking about people who build their own systems - proper legitimate manufacturers do it as well. You could do worse than buying a decent PSU and seeing if that changes things. BTW: A decent PSU is £50-£70, rather than £10.

And a good one ought to last you across a couple of PCs at least. Unless they change something again. Bastards.

Not quite! First it goes over the Northbridge heatsink!

Even better! Ha!

I think my current CPU fan is like that actually - it's angled that way because the heatsink is so big that sticking a fan on top in the normal position would have it protruding somewhere around where I put my right leg under my desk.

Heh, so, y'know that weird 'fog' effect I mentioned that turned up after playing Burny P for a while? Turns out that was...fog! Whatever graphical effect the game uses to make smog just renders like shit and the roadside graphics are glitch to fuck. If you keep playing until the sun comes up properly to burn the smog off then the issues goes away. In my defence I've barely played the game so didn't know it had weather effects.

Anyway, there is still an issue because the jittery artifacts were not only still present on Portal 2 but, when I looked really closely, they were present at times on the car body in Burny P. It's just a lot harder to notice in Burnout due to the shiny nature of the card body and generally different graphic design from the cleaner Portal 2. It's worth noting also that as soon as I started playing Portal 2 everything suddenly went a shade of purple for no reason and stayed like that throughout so that's another minor glitch to add to the pile.

So, readings and whatnot. I've attached a screenshot of Portal 2 looking odd. You can see the weird purple tint and, if you look closely at the portal gun, the grey dots which were dancing around when playing. The next two are from MSI Afterburner; one from completely idle and another I dropped out of Burnout after about half an hour to grab. At this point the fan was noticeably running but not to a ridiculous degree or anything. Note that the screenshots have attached in the wrong order to the top one is the one from me playing Burnout and the second one is the idle readings.

Thoughts gentlemen?

Hmm. 74 degrees isn't massively hot, but it is a tad. In my experience, 60-ish is about as high as they'll go if you've got good ventilation, but 74 shouldn't be causing artefacts. But then I haven't bought a GPU for 2 1/2 years so maybe they're hotter now!

Also, the artefacts are appearing on specific objects - such as the black parts of the portal gun. Overheating issues usually present artefacts randomly.

Did you mention you're running the latest drivers? I must admit I'm totally stumped. What was the issue with your old card? The same thing or something else?

First thing I'd do is uninstall the drivers with one of those driver removal tools to properly get rid of them - I'm sure AE will recommend a piece of software to do that because I can't remember the name of the one I used to use. Then reinstall the drivers, or possibly a driver that is a couple of versions ago. If that doesn't work then it's almost certainly hardware related. Eliminate that possibility first.

Aye, that did cross my mind latterly. With the last card I'd get random black shapes sort of flashing in and out literally all over the screen and not in any way mapping to actual game objects, and it would definitely get worse as time went on and the temp rose. This issue seems to almost be a problem rendering certain in-game objects right from the start.



Yeah, I grabbed the latest set straight from the ATI site. I'm going to try redownloading and reinstalling them now though.

Also, to answer AE's previous question, according to CPU-Z my mobo is an MSI MS-7519 running v1.7 of an AMI BIOS. I've attached a screenshot in case there's anything else in there anyone wants to know.

I added more to my post - give it a read before you uninstall the drivers! You want to remove them properly with a removal tool. Just unsintalling them leaves things behind which can unfortunately make it a waste of time even bothering!

edit: So it sounds like your previous card was properly fucked. This could just be a driver issue - hopefully!

Driversweeper! That's the one! Download that and install it, then run it and uninstall your graphics drivers that way. Make sure you have the drivers you want to install after that handy to make the process pain free (in case your PC goes into 100x200 resolution or something without proper drivers making things difficult).

Ripped the drivers off and reinstalled them (eventually, the actual install failed and I had to manually do it from Device Manager because of course I did) and...it's made no difference! God motherfucking dammit. I can't face any further buggering around tonight so, failing any other suggestions, tomorrow I'll see if I can locate and install and older driver set to see if that makes any difference and if not then I've no idea what I'll do.

That aside though, thanks very much for all the patience and help you've all shown thus far; it's very much appreciated.

Ah bugger! And no worries. Happy to help - hopefully you'll get it sorted tomorrow. Give us a shout if you want more help! Well, "help".

Have you got your original OS install media available Bamba?

I note that you have two hard drives in the PC, so assuming you can get all your valuable data backed up to one of them, just safely unplug that drive, and then do a full rebuild.

It really doesn't take long in the Windows 7 era and is generally very painless.

Actually, before you do anything so drastic running with just 1 hard disk powered up would give a clue as to whether it is a PSU power output issue.

I do indeed have the full Windows 7 install disc, is a full reinstall the usual next step if I can't find a driver related fix?

ETA: I'll certainly try Trooper's suggestion before I go for the wipe.

For a masochist!

Personally I'd just throw money at it and buy a new PSU. Or I would if I had any money.

I suppose a full OS re-install is a possibility and something I've done in the past when I just got fed up trying to fix something, but it all depends how much of a pain in the arse that's going to be for you. How much software are you going to have to reinstall, etc?

Generally I set aside a whole day to get my PC back up to where I had it before on a clean OS install. That's a bit of a commitment if it turns out the problem is still there because your PSU isn't up to the job or the graphics card is faulty!

Spending money on a new PSU's not really a problem as long as people think that's likely to be the issue, but at the moment it does look like some kind of weird software issue. Setting everything back up again after a reinstall would be a pain obviously but I don't mind it too much if it's a likely fix. I'll just make a list of what I've got and spend an evening methodically whacking everything back on again. Buggering about with software is fine, it's hardware tinkering I can never be arsed with for some reason.

Personally, I'd attack it in the following order:

- Uninstall drivers (using removal tool), reinstall [done].
- Uninstall drivers (using removal tool), install slightly older drivers.
- Get angry.
- Search the internet for hours and hours. Finally find a barely comprehensible forum post on some ad-riddled arse-end-of-nowhere traction engine fan forum by a poster who has EXACTLY the same problem you have.
- Scroll down to look at the advice given (past the sarcastic responses and mum jokes from the ad-riddled arse-end-of-nowhere traction engine fan forum's resident Zardoz).
- Keep scrolling. Nobody has a clue how to help.
- Finally, the OP has returned! "I've fixed the problem, thanks for the help guys."
- Get angrier still.
- Complain some more on BEEX. Zardoz responds with a mum joke. Others suggest buying a Mac. An argument breaks out about how much of a pain in the arse PCs are and how consoles are better. John Coffee leaves for a few months. Everyone has forgotten about your original problem and nothing is solved.
- Build an entirely new PC.
- Keep your keyboard, mouse and monitor because they were expensive.
- Retain potentially faulty graphics card and start a three month "returns saga" with the company you bought it from.
- Give up on returning the graphics card, promise self that you'll sell it on eBay.
- Bin the rest.
- Install all of your software and games on your new PC.
- Download 15,000 "Windows Updates". Reboot computer 15 times.
- Get your mouse sensitivity just right because it feels "off".
- Get your web browser all sorted with your bookmarks and log back into every website you visit frequently.
- Mouse still doesn't feel right.
- Download more "Windows Updates" because despite telling you before that they were finished, they weren't. Reboot computer five times.
- Forget why you built a new PC in the first place, check your Steam games list and don't see anything you fancy.
- Play Pinball on your Xbox 360 whilst your Mrs reinstalls The Sims on your awesome new PC.
- Try to sell graphics card on eBay and discover it's now going for about £17.
- After PayPal fees, spend your tenner on a pack of fags.
- Months later, return to scratch that PC gaming itch.
- Update drivers.
- Counter-Strike is displaying some weird graphical glitch.