Sign in with
Sign up | Sign in
Your question

Visiontek AtI 4850 preforming poorly...

Last response: in Graphics & Displays
Share
August 24, 2008 12:57:24 AM

Although I have completly cleaned all my old drivers with driver cleaner pro and have uninstalled all my previous drivers and replaced them with the latest ones from the ATI website, I continue to get poor results from my 4850. For example when trying to run Unreal Tournament 3 when I start up the game it gives me a warning saying that my system does not meet the minimum specs to run the game. Also when I play games like Crysis it sets the optimal system settings as everything low and the res as 800x600.
I have tried replacing my drivers numerous times yet it still preforms poorly. Can this problem be caused by a faulty graphics card? Could it be that my motherboard is not compatible with this video card? Or could there be another problem? If someone could please help soon I would be very greatful. Thanks in advance!

I have the following setup:
4X 1GB 667 Ram
AMD AThlon X2 4800+
Ati Radeon 4850
and a 22" LCD
550Watt Cooler Master PSU
My motherboard is http://h10025.www1.hp.com/ewfrf/wc/d...en&lc=en&cc=us

Thanks!
a b U Graphics card
August 24, 2008 2:32:43 AM

Have you downloaded the dual optimiser from AMD?
August 24, 2008 2:36:25 AM

Looks like driver problem, does the PC recognize it as 4850?

If the driver cleaner you ran is more than a year old, it probably don't have the full listing of entries made by newer drivers, possibly leaving some out. Try this:
http://downloads.guru3d.com/download.php?det=1655
Related resources
August 24, 2008 2:54:11 AM

cpu is a major bottle neck. the x2 4600 came out in november 2005.......
a b U Graphics card
August 24, 2008 2:58:56 AM

Whats the recommended for UT3?
August 24, 2008 3:42:39 AM

Thanks but I have the latest dual optimizer installed already and the Pc does recognize it as 4850. I also used the same driver cleaner as the link provided. Any other suggestion?
a b U Graphics card
August 24, 2008 3:47:07 AM

OK, coolnquiet downclocks your cpu, when it does, the game reads it at a lower actual rate than your cpu is capable of. Just hit play anyways. Is your fps that bad? Give us more info about game play/fps
August 24, 2008 4:32:41 AM

I have very similar specs as your computer and my 4850 is getting all the problems you have stated so far. Maybe we can figure this out together.
August 24, 2008 6:21:22 AM

Yeah I get a max of 20 fps in UT3 and Crysis is not playable at all at any level above low.
August 24, 2008 6:30:56 AM

I'm sorry to ask, but did u plug in the 6 pin power connection on the card to the psu?

Also in UT3 are you running the maps with the physics pack or mod on it?

Can we get a screenshot of your GPU stats from you computer? it might be a bad card

After reading the whole thread, I noticed that you didnt state the settings on which you are trying to play on for both games.

ALSO the link to your mobo is dead
August 24, 2008 8:08:32 AM

Sorry about the motherboard link didn't notice that. The 6 pin power connection is connected so far I have been trying to play on everything max in UT3 but I barely reach a max of 20fps in 1680x1050. And for Crysis it runs really low frames when I try on High but on 1024x768. I will get screenshots later on along with another link for the mobo.
August 24, 2008 8:16:32 AM

Also I have the fan speed set to 70% yet my temperatures remain well above 60 mainly around 65
August 24, 2008 3:14:37 PM

you might have to RMA it. :( 
August 25, 2008 7:44:15 PM

Thats what I think too but is there anything else I should try before doing so?
August 26, 2008 8:54:33 PM

Any suggestions before I RMA ?
I already sent the request?
August 27, 2008 12:52:39 AM

Suprised nobody susgested this, but it could be that motherboard. Odds are its running at 8x, or even 4x. There was a guy up here not too long ago with a Dell that had a 680i board in it, that was locked at 8x.

Have you tried any other games, or benchmarking tools such as 3Dmark. If not, I susgest running 3dmark and compare your scores with similar systems. If the card is faulty, youll easily be able to tell. I wouldnt necessarily tie low FPS to a faulty card so easily. Hold off on the RMA.
August 27, 2008 1:02:02 AM

The mobo link doesn't work for me. What motherboard do you have?
August 27, 2008 1:42:30 AM

70%? wow!!! is that healthy for the card?
And I are you talking 65 at idle with that fan speed?

I run my 4870 @ 38% and its loud but the card stays fairly cool. I couldn't Imagine the noise at 70%.
August 27, 2008 1:43:44 AM

I set my 4870 at 70%, it isnt very nice :(  I keep it at 50%, idle temps hover around 48-54c, never hits 70 under hours of load.
August 27, 2008 1:48:52 AM

spathotan said:
I set my 4870 at 70%, it isnt very nice :(  I keep it at 50%, idle temps hover around 48-54c, never hits 70 under hours of load.


Benchmarkers got 97C on the same 70% fan load.
http://en.expreview.com/2008/06/24/first-review-hd-4870...
What did you use to stress it? Typical games don't put gpus under full load. It's like how you'll use Prime95 to stress cpu instead of Microsoft Office. Use a gpu stress utility like Furmark instead. Use standard settings, (default resolution, no AA, no AF), stress mode for at least 30 minutes for temperature to stabilize.
http://www.ozone3d.net/benchmarks/fur/
August 27, 2008 2:03:35 AM

Well I dont play benchmark programs, so im not worried about that. After playing Supreme Commander on max for awhile it peaked at 69c. About the same with C&C3, and CoD4. I have a NZXT Tempest case with very good wire management. Have two 120mm fans blowing air to the fan and one hitting the side of the card, disconnecting these and the temps go up roughly 15c.
August 27, 2008 2:13:13 AM

spathotan said:
Well I dont play benchmark programs, so im not worried about that. After playing Supreme Commander on max for awhile it peaked at 69c. About the same with C&C3, and CoD4. I have a NZXT Tempest case with very good wire management. Have two 120mm fans blowing air to the fan and one hitting the side of the card, disconnecting these and the temps go up roughly 15c.


Supreme Commander is cpu bottlenecked, and C&C3 is hardcapped at 30fps by the game itself. Like I said, those don't nearly reach full load. You may not care for solid performance testing, but others do. It's established practice to post full load results. Others will assume you're adhering to the same standard as themselves. Don't measure yourself by a different rule and then post flawed results to mislead others.
August 27, 2008 2:21:04 AM

Real world results are not a flawed result. Sorry. Not amount of forum expertise is going to sway me otherwise. No offense, eventhough I understand what you are saying.
August 27, 2008 2:25:22 AM

spathotan said:
Real world results are not a flawed result. Sorry. Not amount of forum expertise is going to sway me otherwise. No offense, eventhough I understand what you are saying.


Nothing wrong with posting your temperatures. You should just have made it clear that those are not full stress results, which you didn't do. Stressed temperatures are the assumed norm. If you don't say so, they'll assume those low numbers are from benchmark stress.
August 27, 2008 2:30:09 AM

I understand, like I said. I know benchmarks are important obviously, but I just hate for so much weight to be put on them, and very little put on real world performance.

Weve gone a little off topic :D 
August 27, 2008 5:13:59 AM

Thanks for the replies well my mobo is Asus A8M2N-LA A.K.A. (NodusM3-GL8E) I have tried a few other games including Assassins Creed and Gears of War which both run at barely playable fps. I currently have taken my card out of my system but if I am correct I remember the GPU scoring in the low 3K range in my last 3dMark Vantage test. Also before removing my card I tried setting my fan speed to 80% yet on idle it was still about 62C-63C which is pretty hot. Anymore help would be appreciated. Thanks
a b U Graphics card
August 27, 2008 7:20:39 AM

Quote:

Nothing wrong with posting your temperatures. You should just have made it clear that those are not full stress results, which you didn't do. Stressed temperatures are the assumed norm. If you don't say so, they'll assume those low numbers are from benchmark stress.
Dagger you should listen to people who actually own an ATI card before saying the number they post are not correct. I get 56 degrees max with 96-97% GPU load after 10 hours of F@H (GPU client). I have decent case airflow but i must say that without the fan fix (45%) the GPU temp would reach 90 degrees after 10 minutes. I have a friend that has the 4850 card and he hits temps higher that me with max 10 degrees with the same fix. So stop annoying people with comments based on only benchmark sites, because you did not witness the tests they performed or the conditions they were performed.
August 27, 2008 12:49:39 PM

hallowed_dragon said:
Quote:

Nothing wrong with posting your temperatures. You should just have made it clear that those are not full stress results, which you didn't do. Stressed temperatures are the assumed norm. If you don't say so, they'll assume those low numbers are from benchmark stress.
Dagger you should listen to people who actually own an ATI card before saying the number they post are not correct. I get 56 degrees max with 96-97% GPU load after 10 hours of F@H (GPU client). I have decent case airflow but i must say that without the fan fix (45%) the GPU temp would reach 90 degrees after 10 minutes. I have a friend that has the 4850 card and he hits temps higher that me with max 10 degrees with the same fix. So stop annoying people with comments based on only benchmark sites, because you did not witness the tests they performed or the conditions they were performed.


Folding at home is not a good utility to test ATI cards. It bottlenecks due to the architecture of how ATI stream processors are set up. That's why comparable Nvidia cards perform so much higher in folding at home despite trialing in gaming. You have to use a real stress utility that does not bottleneck anything, like Furmark and rthdribl.
http://www.ozone3d.net/benchmarks/fur/
http://www.daionet.gr.jp/~masa/rthdribl/
I'm guessing you've already tried, as they're the most commonly used, but decided to not mention the results you got. :sarcastic: 
a b U Graphics card
August 27, 2008 1:05:42 PM

Nope...I haven't tried them. But if GPU-Z considers the load on the card to be 96-97%, I tend to believe it. This is a stress test so I guess the type of operations I perform don't matter as long as the card is at that maximum load. :kaola:  . I loaded Crysis with very high setting but failed to load the GPU to 100 % :ange:  and never exceeded 53 C temp. So I guess F@H stressed the card more than a game which is known to keep all GPU on their knees :D  . So again why shouldn't I use F@H as a stress test? :bounce: 
a b U Graphics card
August 27, 2008 4:53:32 PM

Theres a new folding client coming out for ATI cards. Ive heard itll be very good
August 29, 2008 12:36:18 AM

So should I RMA? And yes I did disable the onboard graphics in Bios but how can I tell if my board is locked to 4x or 8x?
August 29, 2008 12:40:57 AM

It might tell you under "onboard devices" or "advanced" tabs in your BIOS.
a c 187 U Graphics card
August 29, 2008 7:33:00 PM

If it`s active, try disabling `cool `n quiet` by changing the power saving options in Windows to anything but Minimal Power Management.
a c 187 U Graphics card
August 29, 2008 7:36:21 PM

OOPS!
hamzanaeem002, download GPUZ, it`ll tell you what pcie speed your system is using
!