(G92) 8800GTS 512mb thread

[retsgtcas]

To claim a "stable" clock, how long do you guys either run a loop of the first 7 test in 3DMark06 or ATI Tool Artifact testing utility?
I do realize that its a bit of a personal taste, but what should be, say forum standard for this card.

 

[Elluzion]

what should i use to overclock this card and increase fan speeed??? Riva Tuner??????

 

[mblue]

what should i use to overclock this card and increase fan speeed??? Riva Tuner??????

Yes, but you may have to make a slight modification for it to recognize that 8800.

 

[Elluzion]

Yes, but you may have to make a slight modification for it to recognize that 8800.

and what would that modification consist of? info?


i just installed my 8800 gts 512. anyone overclock using riva tuner>?


where to download latest version anyways>?>

 

[Mustanley]

It's noted on page 2 of this thread.

Here ya go...
http://www.ocforums.com/showpost.php?p=5422343&postcount=47

You can get the latest version (2.06) at guru3d.com

 

[mblue]

http://forums.guru3d.com/showthread.php?t=246531&highlight=8800gts+512

Easy and it work.

 

[jason4207]

To claim a "stable" clock, how long do you guys either run a loop of the first 7 test in 3DMark06 or ATI Tool Artifact testing utility?
I do realize that its a bit of a personal taste, but what should be, say forum standard for this card.

I like to go about 15-25 mins of ATI Tool w/ no artifacts. You can tell if you had an artifact while out of the room b/c the two timers won't match.

The GT, and GTS seem to have a memory avalanche problem that VJ has discovered the weak spot for, so a little additional testing is necessary for total stability. For the GT run 8+ loops of 'Battle of Proxycon' from 3DMark03. If you get through clean you should be pretty good. [For me F.E.A.R also brought out this issue very quickly on the GT.] For the GTS run 8...maybe 10 loops of 'Canyon Flight' from 3DMark05. Again, get through clean, and you should be rock-solid.

 

[leftheaded]

Can someone help me understand what I'm giving up if I should decide to replace my 512MB 8800GT SLI setup for a single 512 8800GTS?

I see plenty of people getting around 16k with a single 512MB 8800GTS & that's what I got with my GT in SLI. they bench the same, but is the real world performance going to be the same?

does it only depend on what resolution i want to run? anything else?

 

[Blazing fire]

I've got an XFX 8800GTS 512 Alpha Dog and the card is incredible. I love the thing so much. It also overclocks like a beast. Stock, the card is 650/1625/972 (core/shader/memory) and now it's running at 800/1900/1075. My 3DMark score is held back by my CPU, but I still posted 11,447 3DMarks. These cards are amazing.

But 3.2 ghz should not be a real bottle neck right?

 

[ViperJohn]

Can someone help me understand what I'm giving up if I should decide to replace my 512MB 8800GT SLI setup for a single 512 8800GTS?

I see plenty of people getting around 16k with a single 512MB 8800GTS & that's what I got with my GT in SLI. they bench the same, but is the real world performance going to be the same?

does it only depend on what resolution i want to run? anything else?

There is absolutely no reason to do that. It would be a step downhill.

Because the CPU score is included in the 3DM06 overall score you can not go by what any
body else's system does in it. CPU/System speed has a huge impact on the 3DM06 overall
score. Just having a Intel C2Q is worth 1200-1400 more overall points over a same clocked
C2D just from the better CPU scores the quad gives.

You can only go by relative scores on the same (your) machine.

Viper

 

[Hipcrostino]

Can someone help me understand what I'm giving up if I should decide to replace my 512MB 8800GT SLI setup for a single 512 8800GTS?

I see plenty of people getting around 16k with a single 512MB 8800GTS & that's what I got with my GT in SLI. they bench the same, but is the real world performance going to be the same?

does it only depend on what resolution i want to run? anything else?

umm looking at your sig, you system should be getting much higher scores than 16k. I get 15.5k with a single GT! i looks like one of your cards isn't doing anything!

 

[jason4207]

Can someone help me understand what I'm giving up if I should decide to replace my 512MB 8800GT SLI setup for a single 512 8800GTS?

I see plenty of people getting around 16k with a single 512MB 8800GTS & that's what I got with my GT in SLI. they bench the same, but is the real world performance going to be the same?

does it only depend on what resolution i want to run? anything else?

Just stay w/ what you have, and if you are looking for better 3DM06 scores then OC your CPU & GPUs more. As has been stated increasing your CPU OC won't have any real-world effect for games, but it will boost your score if you're looking for bragging rights.

I got over 16k w/ 1 GTS, but I also had my Quad at 3.8GHz when I got that score. I was getting 14.8k w/ 1 GT in my rig at 3.6GHz CPU.

 

[leftheaded]

thanks guys. the thing is, I have to RMA my XFX 680i mobo, but zipzoomfly no longer sells this mobo. They said my RMA could be for a refund only.

I also have to RMA one of my GT cards with Dell. They actually said I could exchange or return one or both GT cards for no fees.

So, this situation has me considering entirely different mobo/VGA builds - eg p35/x38 with a single card (GTS 512 or Ultra).

The kicker is I have to decide today or tomorrow as my RMA window is running out.

anyway, i've been thinking about getting abit ip35 pro, maximus formula, or x38 DQ6 at this point. I guess I should take one more look at the evga 780i...

thanks again gentlemen

 

[jason4207]

In that case I would send back the 680i, send back the GT's, and pocket the extra loot! I'm not a big SLI fan right now, though.

Get yourself a good GTS (eVGA), and you might want to look into the Foxconn MARS P35 MoBo. It looks like a contender, but I haven't looked into it too much.

 

[OC nub]

8800 gt

Picked up a BFG 8800 gt oc. Was quite impressed with the performance. Even running on Vista, scores are pretty good in 3dmark. I play at 1600x1200 res in COD4 most of the time and even with AA at x4 it runs very smooth which is more than I can say for my radeon hd3870's in crossfire. AA kills the performance, and without it the game looks like crap.

 

[reboot]

in case anyones interested, on my XFX
756 x 1890 / 1061 @ 60% fan speed gives me 54C idle and 72C load

with an hr-03 GT installed, and a silverstone 92mm fan on low ( not all the way down but cannot hear low ) and ICD7 diamond paste my idle temps went to 31C idle and 41C load.

same load, same ambient temp.

 

[majinwar]

Here are my results from stock speeds n stock cooler vs OC and HR-03 GT.

Stock (650c / 1625s / 972m): Idle-47 Load-69
3DMark06 12258
2.0 = 5631
3.0 = 5682
CPU = 2794

OC (760c / 1900s / 1062m): Idle-40 Load-59
3DMark06 12815
2.0 = 5797
3.0 = 6229
CPU = 2792

Does this look right? Or is my suspicion of my CPU holding it back correct?

Also the temps look a little high...

 

[leftheaded]

Here are my results from stock speeds n stock cooler vs OC and HR-03 GT.

Stock (650c / 1625s / 972m): Idle-47 Load-69
3DMark06 12258
2.0 = 5631
3.0 = 5682
CPU = 2794

OC (760c / 1900s / 1062m): Idle-40 Load-59
3DMark06 12815
2.0 = 5797
3.0 = 6229
CPU = 2792

Does this look right? Or is my suspicion of my CPU holding it back correct?

Also the temps look a little high...

yeah, i'd wager the cpu is definitely the limiting factor. my cpu score is around 4500 with a q6600 @ 3.2GHz.
my scores on page 2

 

[Jas]

Here are my results from stock speeds n stock cooler vs OC and HR-03 GT.

Stock (650c / 1625s / 972m): Idle-47 Load-69
3DMark06 12258
2.0 = 5631
3.0 = 5682
CPU = 2794

OC (760c / 1900s / 1062m): Idle-40 Load-59
3DMark06 12815
2.0 = 5797
3.0 = 6229
CPU = 2792

Does this look right? Or is my suspicion of my CPU holding it back correct?

Also the temps look a little high...

It's the CPU. Your scores are right on par with a GT with those clocks.

 

[majinwar]

It's the CPU. Your scores are right on par with a GT with those clocks.

I figured as much.

A little update though, I flipped the heatsink around, found out I didn't use enough paste last time, but now idle temps are 34 degrees and load is 42! Also have it at 795 core / 1987 Shader / 1062 Memory without artifacts!