Nazo
2[H]4U
- Joined
- Apr 2, 2002
- Messages
- 3,672
I keep noticing what I would normally think of as an artifact in a lot of games. Almost every single time without exception, the issues are with shadows. They do weird things like flickering in NFS-MW and in another game I'm playing, Sudeki, I often see this sort of pattern of shadows on people that I'm pretty sure either should be a clean single shadow or not there at all. Is this normal (as in possibly game engine issues or something else like that,) or is this a form of artifacting I hadn't seen in the past with my previous video cards? I'm kind of wondering if it might be temperature related as stock cooling lets my card run pretty hot even when ATITool forces the fan to higher speeds. Despite what people love to claim -- that GPUs can take a LOT of heat -- earlier ATITool screwed up my fan settings and was forcing the fan down to 14% until 80 something (my normal settings have the fan usually running around 50% for most games, which is loud) and the video card was running into the range of 74C or so in just Sudeki (which is no NFS-MW, with it's most powerful effects being a slight shine to all the metals.) I wondered why the system made a VPU recover once, and then, a bit later it crashes and I noticed it was running 74C around that time. Maybe the GPU won't fry until 90C or so, but, it's not good for it to run in the 70s clearly. I should also add that I am using ATITool, running it at the maximum temperature at correct fan speeds (about 63, maybe 64C with the fan at 50% and no ocing) and I get no artifacts detected, which is part of why I wonder if I should maybe blame the game engines. Then again, just how much does ATITool test? Does it run any shaders or anything? Or does it maybe just toss sheer raw numbers of polygons at the GPU?
I may already be too late to try to send back to newegg. For one thing, this card is no longer being produced by the manufacturer (VisionTek,) even if they did just simply rebrand a stock ATI card as nearly as I can see (didn't even toss in extra software or cables.) I'd probably have to rely on manufacturer warantee then, and even then I don't know if I can do anything since it looks like they've stopped even selling this card in favor of only selling the PCI-Express version (surprise surprise.) I'm pretty well planning on getting an ATI Silencer either way, but, legally I can't RMA this card if I change the cooling, and less legally, even if I swapped back after installing an ATIS, they'd surely notice the changed thermal paste.
So, the question is, is my card producing errors ATITool can't detect, is it just running too hot for full stability at the stock clocks, or am I just seeing problems in game engines (and Sudeki can't be that well designed -- even has known problems with some ATI cards, though mine isn't in the list so far as I know)? Do I try to RMA now if I even can, or go ahead and get that expensive cooling upgrade so I can finally start properly ocing this card?
EDIT: BTW, on subject but a little off, is there any way to change the BIOS fan speed setting without using a weird modified BIOS? Like some kind of tool that would let me just simply change my stock BIOS fan setting? I don't like having to rely on ATITool -- especially after this little screwup that forced my card to run so hot like that...
EDIT2: I noticed my memory won't hardly OC at all, but, when I ran ATITool, it managed to get the core all the way up to 576 before it locked up the system. The memory ran up to 596 or something along those lines before it started backing down to 591.55 or so. Stock being 590... It's not hard to believe the memory could produce errors more easily than the GPU since I'm running the card at stock until I improve the cooling (it's bad enough as it is -- I'm especially worried the way temperatures just practically jerk nearly 30 degrees in a matter of seconds sometimes. Not good for any electronics, even if they aren't supposed to fry until something far higher.) Under worst conditions right now, the card's "ambient" (eg memory) temperature is in the area of 42C, which isn't particularly bad, but, definitely isn't particularly good either considering chasis temperature is 28C or so tops.
Oh yeah, and I tried an underclock. I set the GPU to 450 and the memory to 475. I still saw some of the same sort of flickering and other such things that just felt like artifacts to me, and ATITool still says 0 artifacts.
I may already be too late to try to send back to newegg. For one thing, this card is no longer being produced by the manufacturer (VisionTek,) even if they did just simply rebrand a stock ATI card as nearly as I can see (didn't even toss in extra software or cables.) I'd probably have to rely on manufacturer warantee then, and even then I don't know if I can do anything since it looks like they've stopped even selling this card in favor of only selling the PCI-Express version (surprise surprise.) I'm pretty well planning on getting an ATI Silencer either way, but, legally I can't RMA this card if I change the cooling, and less legally, even if I swapped back after installing an ATIS, they'd surely notice the changed thermal paste.
So, the question is, is my card producing errors ATITool can't detect, is it just running too hot for full stability at the stock clocks, or am I just seeing problems in game engines (and Sudeki can't be that well designed -- even has known problems with some ATI cards, though mine isn't in the list so far as I know)? Do I try to RMA now if I even can, or go ahead and get that expensive cooling upgrade so I can finally start properly ocing this card?
EDIT: BTW, on subject but a little off, is there any way to change the BIOS fan speed setting without using a weird modified BIOS? Like some kind of tool that would let me just simply change my stock BIOS fan setting? I don't like having to rely on ATITool -- especially after this little screwup that forced my card to run so hot like that...
EDIT2: I noticed my memory won't hardly OC at all, but, when I ran ATITool, it managed to get the core all the way up to 576 before it locked up the system. The memory ran up to 596 or something along those lines before it started backing down to 591.55 or so. Stock being 590... It's not hard to believe the memory could produce errors more easily than the GPU since I'm running the card at stock until I improve the cooling (it's bad enough as it is -- I'm especially worried the way temperatures just practically jerk nearly 30 degrees in a matter of seconds sometimes. Not good for any electronics, even if they aren't supposed to fry until something far higher.) Under worst conditions right now, the card's "ambient" (eg memory) temperature is in the area of 42C, which isn't particularly bad, but, definitely isn't particularly good either considering chasis temperature is 28C or so tops.
Oh yeah, and I tried an underclock. I set the GPU to 450 and the memory to 475. I still saw some of the same sort of flickering and other such things that just felt like artifacts to me, and ATITool still says 0 artifacts.