6800GS drops in speed a large amount in less than 24 hours.

G

GForce64

Gawd
Joined
Mar 7, 2005
Messages
788
I just installed a 6800GS yesterday, ran NFS: MW great, popped 4800 3dMarks in 05. Run MW today, it's slow as heck, so I run 3dMark again. 1700. What the heck could happen to my card in such a short period of time to make it so slow?
 
GForce64 said:
I just installed a 6800GS yesterday, ran NFS: MW great, popped 4800 3dMarks in 05. Run MW today, it's slow as heck, so I run 3dMark again. 1700. What the heck could happen to my card in such a short period of time to make it so slow?
Click to expand...

Throttled back due to overheating?
 
VSync's clear, idles at 44, about to get rthdribl to check load temps, not sure how I could check voltage.

Edit: Load temps are at 56 running rthdribl.
 
Use some kind of diagnostic to make sure the core and memory speeds are what they should be.
 
Coolbits reports 425 MHz for the core and 1 GHz for the memory, what it should be.
 
The culprit is something other than your video card. Check CPU temps. You are running a "hot" video card that dumps its waste heat into the case. This raises the case air that your CPU's HSF is using to cool the CPU which makes the CPU run hotter, which raises the case temp further, etc. etc. All of a sudden your CPU is throttling back to save itself.

A simple solution may be to purchase an Arctic Cooling GPU cooler which cools your video card better, and quieter, than OEM while dumping the waste heat out the back of the case rather than into it, thus relieving the stress on all the other components, including your CPU.

So , record the CPU temps and if they are higher then use the above $35 solution to get back to equilibrium.
 
Merc said:
The culprit is something other than your video card. Check CPU temps. You are running a "hot" video card that dumps its waste heat into the case. This raises the case air that your CPU's HSF is using to cool the CPU which makes the CPU run hotter, which raises the case temp further, etc. etc. All of a sudden your CPU is throttling back to save itself.

A simple solution may be to purchase an Arctic Cooling GPU cooler which cools your video card better, and quieter, than OEM while dumping the waste heat out the back of the case rather than into it, thus relieving the stress on all the other components, including your CPU.

So , record the CPU temps and if they are higher then use the above $35 solution to get back to equilibrium.
Click to expand...

Wouldn't this have been an immediate issue if it were to happen? Remember, the card ran perfectly fine for a long period of time.

Edit: Just reformatted, scored an 1850.
 
After a format and still a problem? Considering you installed the card's drivers right the first time thus getting the initial score it is likely that it is hardware and it isn't your graphics card. Make sure your cpu's fan still works. Make sure no heatsinks fell off your motherboard/ram. Make sure no other fans stopped on the motherboard (north or south bridge).

Check amps and voltage on your psu with a voltmeter (only do this if you know what you are doing, if you don't ask someone you know who does know to help you, I repeat, I am not responsible for your own idiotic mess ups; you've been warned).

~Adam
 
54YW4T said:
i say its PSU.....
Click to expand...

That would seriously suck, considering it's a 500w Antec.

Edit: This has the makings of a cooling issue. Just take case panel off, 2789.
 
GForce64 said:
That would seriously suck, considering it's a 500w Antec.

Edit: This has the makings of a cooling issue. Just take case panel off, 2789.
Click to expand...

If it was a cooling issue, taking the side panel off would alleviate 99% of the problems. There's something else wrong here. My guess is bad contact with the heatsink on the GPU, or bad thermal paste.
 
Bona Fide said:
If it was a cooling issue, taking the side panel off would alleviate 99% of the problems. There's something else wrong here. My guess is bad contact with the heatsink on the GPU, or bad thermal paste.
Click to expand...

Good point. I'm out of thermal paste, though, so I can't really check.
 
I am having the very same problem. I have been doing plenty of checks running temps and demos and benchmarks.

My first benchmark was on 3dMark03 and I scored 11591. 2 weeks later I tried again, and suddenly score 5800. Now, my new machine is no faster than my old Radeon 9700.

I ran rthdribl and got out the RivaTuner hardware monitoring. I am in my basement which is quite cool (72F) and my CPU runs about 90 degrees under load (stock A64-3200). My GPU temp runs from 36C-46C. 46C is max under load.

with rthdribl running, I see 46 Max on GPU temp, but the framerate shows a steady drop from about 70FPS all the way down to 23FPS. GPU temp never goes over 46C.

I noticed similar problems with GLExcess benchmark. If I pick one of the fillrate tests, it runs really fast by itself, but you can see it fall off as time goes by. If I run other tests (polygon counts, etc) the fillrate tests fall facedown. The longer the card runs in 3d mode, the slower it gets.

BTW, this is an eVGA 6800GS on an ASUS A8n-SLI (original model, not delux, premium or 32).

I am also wondering about temps. Maybe video ram temps?

If you have solved anything, let me know.

Thanx
Oddbawl
 
You must log in or register to reply here.
Back
Top