Low FPS in games, help please

I’m completely baffled by this, I’ve been trying to play Borderland and here’s a brief timeline -

Release day - Perfectly fine FPS, get up to Nine-Toes, die, computer locks up and I have to restart. I get POST code for bad graphics card (didn’t know which code it was yet). I turn off the power, turn it back on, everything seems fine. Go back in the game and everything is stuttering terribly.

Between then and now (about a week) - I try everything, stripping and reinstalling drivers, trying different releases, reseating cards, even verifying both cores of both of my cards are working (I have 2 NVidia 295 GTXs in SLI). I even deleted the game and reinstalled it. Throughout this occasionally tests failed during the splash screen, which caused the same reset-POST for reseat or replace Graphics card sequence. If I got through the splash screen, which I did quite often, the menu would be running at a pristine FPS, and so would the opening cinematic, but as soon as I got in-game it would slow to abysmal framerates. Never got it to crash again in game, but I couldn’t really play so it may have if I’d died again.

Today - To get Windows Update to stop bugging me (I’m going to get a non-RC version of Windows 7 tomorrow and do a full wipe so I didn’t bother) I install updates, in the hopes that maybe it will help. Surprise! I boot up the game and it’s perfectly fine. No FPS snags, same as release. Then I go to an area transition after the opening section (to those that have played the game, the quest to kill Nine Toes), screen blacks out, computer locks during loading. I reset it again and expect the POST code but I end up with an all clear and boot as normal. However, it’s back to abysmal framerates. GAH!

This isn’t completely new, when I got the computer in February up until one or two months ago it was fine. Now I have to do stuff just to get it to play stuff properly, to play a flash-based application in full screen I now have to deactivate Quad SLI, Machanarium has never worked, but it’s from well after the issues started. Cutscenes (of all things) in games have a tendency to stutter, like in the Mini Ninjas demo, the game plays fine but the opening cut… scene… goes… li… ke… this… bo…th au…dio and video. World of Warcraft doesn’t even do the menus right with SLI, with Single-card (which is still dual GPU) it works while I’m inside a building but as soon as I’m outside I hit 6 FPS or so. Most other games appear to work fine. The only thing that may have been a gateway was Champions Online, which was the first thing I experienced with really low framerates (there were a lot of issues on the forums), I tried tons of drivers, from homebrew to newer to older and it finally worked with 186.18, and everything else I was comparing it against noticeably improved too. It clearly wasn’t the trigger, just the first sign.

At first I thought maybe it was a botched driver install, I was having trouble installing them in the first place, it would keep telling me install failed until one time where it magically worked. But, I’ve uninstalled, booted in safe mode, stripped the drivers with driver sweeper, and reinstalled several different times once again and no improvement. So far the only thing that has fixed anything was the single Windows Update, and only for about an hour at that.

I checked all the usual (malware and the like) and nothing seems to be wrong, not anything detectable at least.

Anyway, I get Win7 (I run x64, by the way) retail in tomorrow, so I’m going to try wiping my HDD and doing a completely fresh install. But until then, any wild guesses? If that doesn’t work I figure I’m completely screwed and need to replace my cards, so this is partially out of curiosity now.

Wild guess: How’s your cooling setup? Maybe they’re overheating.

You know, I was going off my case thermo, you would think I would’ve thought of that. SpeedFan is giving me (I’m not even kidding) 81C (!!! :eek:) on my GPU (oddly, my CPU with the awesome fan is only 45). My case intake fans are apparently not cutting it correctly.

Odd considering I have almost the exact same setup as my friend.

That would explain my OCing issues too.

Kids, let this be a lesson: your case thermometer sucks eggs. The card fans are clearly working right now, so I guess my options are relocate or space the cards father apart on the PCI-E rack.

Thanks, not sure why I didn’t think of that.

Edit: I wonder why Windows Update worked for a second though, must have been a weird blip.

Well, thank me when and if it goes from hypothesis to solution :slight_smile:

As for WU, it may be that you weren’t using your graphics cards much during the update, allowing them to cool down just enough that they worked for a while.

Well, I mean, it still crashes from a fresh boot after being off all night, which is why I didn’t guess that in the first place.

My CPU with the awesome HSF is running at a nice 45C, and I have two giant 25C intake fans on the side of the case. My hypothesis right now is that the fact that the cards are:

  1. Adjacent
  2. Big and therefore
  3. Touching is doing bad, bad things with heat distribution.

If that doesn’t work I’ll try moving my case out onto the open floor, and if that doesn’t work I’m lost. Water cooling? Case fridge? Move to Antarctica?

And now it’s idling at 85 (oh wait, one is at 88), wonderful. My only other idea is to replace the TMI and make sure the heatsink is properly fastened on the card and I’m not sure how since I’ve never done graphics card surgery before. Tomorrow is going to be a fun day.

Okay, I had to download nTune and set my fan speed to 100%, hardly ideal, but it’s idling at mid-low 60s now, again, hardly ideal. But at least it’s theoretically usable now.

Replace that card. These are all symptoms of a hardware issue, not a software one. I think all you are doing is just reseting it with all these reboots and driver installs and getting a little extra life out of it.

Yeah, but this is two cards. What are the chances of two cards having the same issue? If anything it’s case airflow. Anyway, after a good night’s sleep for the computer, at 100% fan speed it still sounds like an airfield in here, but it seems to be idling at JUST below 60. I think I’ll have to try surgery.

I can’t really RMA these cards anyway, I completely depend on this computer and I don’t have a spare card, if I send them back to be replaced I have a brick and no way to do homework.

Oh. Lovely, I don’t have a warranty past the initial 30 days, just checked. So I’m screwed there anyway.

Okay, I tried Borderland again at my new temps just as a benchmark. Even though temps are a definite issue it’s not what was causing the original problem. The FPS is still abysmal even at a 15 degree temperature decrease, it’s acting no different than it was at higher temps. So either I have two faulty cards somehow or it’s a really esoteric software issue that’s causing undue load for some reason (I checked all the other hardware stuff, I have all the right amperage and board compatibilities that can cause heat issues sometimes).

What happens if you take one card out?

Do you get problems in other games?

Try using one card only? Maybe the SLI has issues with that specific game. Have you tried any other games?

Also, in my experience, case fans are only good for generating noise

Edit: LOL Quartz!

One card, doesn’t matter which one it is, same thing happens. Actually performance improves a tad (but not much) for older games, but not newer games.

ETA: And trust me, my case fans improve my temps by about 10degC. I figured that out a long time ago when one went out and I had to rush order one just so my computer didn’t start overheating (I really should have figured out the heat problem then).