Greetings, I'm new to forums, but I'll do my best to be as explicit as possible with my problem. System: (built myself ~January 2005) OS: Windows XP Professional SP2 Asus K8VSE Deluxe Motherboard (VIA chipset) AMD Athlon 64 3000+ ATI Radeon 9800 Pro 128 MB (manufactured by ATI, it's the real deal) 512 MB pc2700 RAM CompUSA PCI 5.1 soundcard (placed one extra slot below video card) Netgear Etherfast PCI nic ANTEC TruePower 550w PSU I bought the processor OEM so I had to purchase a heatsink. I opted to get a Speeze Heatsink (certified for my proc). I also recently added two Antec LED case fans, which have really helped reduce my case temperature. I came home from college this week for spring break, and I began to notice my problem during a night of gaming. I was playing KOTOR 2 (pretty graphic intensive if you beef settings up; however I was only running 1024x768 ~medium graphics with no AA or AS) My game completely froze up. I thought I would need to manually reboot, but I waited awhile and it finally kicked back to desktop and I was given ATI's VPU error. (The exact error msg is unimportant, as the event described is clear.) My "graphics accelerator was restarted by the driver because it stoppd responding to commands from the driver." That's it. Thinking the problem was occurring only because of KOTOR 2, I tried loading up CS:S. Sadly, the same thing happened within ~2 mins. I can also tell that my games feel sort of sluggish, definitely short of the capabilities of my system. Naturally, my first reaction was to completely remove my video drivers via uninstall and driver cleaner, then to reinstall the latest ATI Radeon drivers. This was no success. I then began to browse forums. Apparently quite a bit of people have had this problem. Suggested remedies include: Turning AGP Fastwrite off, Reinstalling MOBO drivers, making sure the GPU isn't overheating, increasing AGP aperture size, turning off VPU reports, flashing/updating BIOS, and making sure RAM is not bad. Also, some individuals reported that the problem was related to power. (voltage problems / insufficient power) I have tried all of these remedies save flashing and updating my BIOS, which are already the most up to date ASUS BIOS for my mobo. I'm positive heat is not a problem, as I have an extra slot underneath my card, and I can see the fan on the heatsink spinning. I don't want to turn off VPU reports, because I could risk damaging my card / I would just have to reboot anyway. Here's where my question comes in. I think I'm having a power issue. Even though I have 550w PSU, I fear that the power cable im using to plug into it is not good enough. (I left the one that came with my PSU in the dorm room) On the box for my PSU, for input current it says: 12.0A for 115VAC, 6.0A for 230VAC. I'm no electrical engineer / electrician, but I think I may not be giving it the correct input. On the cable I'm using it says: 10.0A 250V. I think the cable at the dorm says 12.0A. So would this slight difference be a possible reason for my graphics card going on the fritz? The odd thing is that I didn't have the problem the first few nights I was home. It only popped up the night I started playing KOTOR 2. Who knows, perhaps it's coincidence or maybe I got lucky the first few nights. Thanks for taking the time to read this incredibly long rant, but I'm growing desperate for feedback. Ian

Just because you have an extra slot under your vid card doesn't mean that heat can't be a problem. Try opening it up and allowing a house fan to just blow inside while you play and see if the problem stops. You even made a note about adding extr fans to help cool it down...maybe the fans are just not up to speed or are bogged down with enough dust to slow them down. That graphics card by itself runs extremely hot when used with intense gaming. Any way it will ellimainate any doubt about being caused by heat. The other issue is that you may have a faulty ram stick and is locking you up. If you have two sticks...remove one and try the game and see what happens...then switch it out and try again. As to the PSU being the culprit then about all you can do is check for voltages and that may or may not tell you anything as it could be failing under load. You can either switch it out or unconnect everything except that which is absolutely needed for game play. Unplug the floppy and any other CD-Roms or DVD drives or slave drives that you are using and then test your card and see if the loads change that way. About all I can think of so hope it helps. Keep us posted on the fix. IN THE MATTERS OF STYLE, swim with the current; in matters of principle, STAND LIKE A ROCK

Thanks for the quick response. I had tried gaming with the side of the case off, and I looked at the GPU heatsink and saw the fan spinning. However, I hadn't tried stationing a fan right next to the case. I took your advice and tried the remedy, but unfortunately I still got the error / lockup. As far as the power issue, my old setup was having the rear case fan piggy-backed with the power cable connecting to the video card. (I know this seems stupid, but the Antec instructions said that it would be fine to do this.) So in an attempt to fix the problem, I disconnected the fan and reconnected it to it's own "fan connector." Sadly, this also proved to be a failure. Since I was out near CompUSA earlier this evening, I decided to buy a "loaner" power cord. I think my problem might be in the cords. As I mentioned in my original post, (and I'm clueless on this subject) when I look at the female connection of my cords, they say 10A.(which I'm assuming is amps) However, the required input for my PSU is 12A. Do the actual power cords themselves have any influence on the flow of current/power throughout the system, or are they pretty much standard? The cord I bought from CompUSA is a Belkin generic one. It didn't help my problem at all, so I'll be returning it. The only thing left to do is test my RAM. I have three sticks of RAM, one 512MB pc2700, and two sticks of 256MB pc2100. I was puzzled when I ran dxdiag earlier today. For my amount of RAM it tells me 510MB instead of 512MB. However, when I go to system properties or run a 3rd party benchmarking program like cpu-z, it says 512MB. Is this anything to worry about? I do remember a few years back when I had a bad stick of RAM the reported value was WAY off, something on the order of several 100 I believe, not just 2 MB. I really am completely puzzed by my problem, because heat was never an issue, even before I added the two additional case fans (which are clean and blowing plenty of cool air; front fan in, rear fan out). In addition, I seem to remember my RAM being reported as 2MB shy at the dorm room also, but I have never had a VPU error, or video card error for that matter, over the entire time I've had the card. Which makes me think it almost has to be the fact that I'm using a different cord than the one that came with my Antec TruePower 550w PSU. Anywho, I'll test the RAM and get back to you on my progress. I won't be able to try the proper cord until dorms open this Sunday. Ian

For a Antec TruePower 550 you should be using a minimum 16gauge (16AWG) cable. My PCPOWERANDCOOLING 510ATXPFC requires a 14gauge. I have the opposite problem in that my Athlon 2800+ gets overheated because I went with a cheap Antec "Quiet" HSF combo and the PC speaker starts beeping and I have to quit playing. My 9800 128MB Pro however doesn't overheat, then again I have 5 case fans.

So, did you fix the problem? I bought a Club 3D Radeon 9800 Pro the other day and experience the same problem in games like Flatout and Halo. Works fine for a minute or so and then - blip - it kickes me back to desk top. I think I've tried almost everything now. Loop

I had a similar problem with being locked up in games, and this fixed it. Dunno, its worth a try. Add this key to your registry. Its a problem with AMD and AGP. Key: [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Session Manager\ Value Name: LargePageMinimum Data Type: REG_DWORD Value: 0xFFFFFFFF in Hex / 4294967295 in Decimal