8800GTS-512 V-mod Results!

[VETDRMS]

So far this card has done pretty good. I also have a GT, which had a better core (918), but shader would not go as high(2214), and the memory was terrible on the GT(950). Memory turned out great on the GTS. However, I have not messed with timings yet.

I wish I had a decent quad core, but I am pretty happy with the results with a single "budget" card and dual core cpu. :up:

Vcore was 1.400v

[guess2098]

So far this card has done pretty good. I also have a GT, which had a better core (918), but shader would not go as high(2214), and the memory was terrible on the GT(950). Memory turned out great on the GTS. However, I have not messed with timings yet.

I wish I had a decent quad core, but I am pretty happy with the results with a single "budget" card. :up:

Vcore was 1.400v
air??

if it is air result, i can guarantee you this is a 20k 3Dmark06 8800GTS 512 air cooling card!

[VETDRMS]

Sorry I forgot to mention, no, it is on water - MCW60 in series with Apogee, single Swiftech 220 radiator. Loaded temps are around 40C at 1.4v.

[over@locker886]

Full air, only 1,35v load.

http://img512.imageshack.us/img512/1519/83650mr9.png

[SmokeyTheBandit]

Also done the vmod and must say VERY worth the hasle :D

Stock Air no mods max clock : Core @ 800
vmodded on air max clock : Core @ 880 (1.4V)

Going to try a bit more but not to shabby for now :)

http://xs222.xs.to/xs222/07506/g92_16577.jpg.xs.jpg (http://xs.to/xs.php?h=xs222&d=07506&f=g92_16577.jpg)

[VETDRMS]

The card will do 882core/2260shader at only 1.3v. That 900mhz and high shader took a LOT more voltage.

[T_M]

These card do crazy clocks under cold temps :)
Mine was easily doing 1080/2700 core/shader under DICE today, with more room to move!

[Gribo]

Hi all. I would ask you if it is any known pencil mod for 8800 GTS 512mb.
At the moment my 3d mark score is 17474 points (http://service.futuremark.com/compare?3dm06=4205911) Of course I want more :D

[Demo]

Those are impressive overclocking results!

[VoRtAn_MaDgE]

Nice clocks indeed.
My g82 gts is not modded yet...
Wich temps are you guys getting at load with 1.3/1.4v ?

Any considerable gains with vmem ?

[VETDRMS]

Temps are usually 34-38C.

Increasing vMem will just make it worse, decrease vmem to improve overclock. My memory is now at 1144 with 1.4ns timings and 1.900v.

[VoRtAn_MaDgE]

Temps are usually 34-38C.

Increasing vMem will just make it worse, decrease vmem to improve overclock. My memory is now at 1144 with 1.4ns timings and 1.900v.

With stock cooler those temps must be idle for sure.
Lowering the resistance of the 50Kohm will reduce the voltage of the memories ?

[VoRtAn_MaDgE]

Well... already hit 880gpu with 1.375 stock cooler :D
fan 100% - 53ºC max
Gonna try more after christmas.

What +/- the score difference running xp and vista?
I'm running vista and i can't upgrade my 3dmark 1.0.2 using the 1.1.0 patch...gives me an error saying that no 3dmark 06 found...

Any idea how to solve that?

I could download the full version patched i know :D

http://img205.imageshack.us/img205/937/8802200gts512kr8.th.jpg (http://img205.imageshack.us/my.php?image=8802200gts512kr8.jpg)

[scotty1991]

How does lowering the voltage on the memory help? and how do you ajust timings?

[4Qman]

MSI OC edition 730/1944 seems to be a good card.

810/2180 Ati Stable Stock volts/cooler which read 1.15/1.18v

Vmod and Cool air next few days :D

[VETDRMS]

Vortan: Those temps are loaded on water using a Swiftech MCW-60.

Nvidia ties the VDDQ and VDD memory together, GDDR-3 can be pretty finicky with high VDDQ. While a higher VDD may improve clock speed a higher VDDQ usually causes instability. 2.00+v seems to really bug out the Qimonda memory at higher frequencies. I have found (thanks to VJs generous help) that a lower VDD/VDDQ improves the stability and overclock of this Qimonda -10 GDDR3.

I changed the timing using Nibitor v3.6 and the latest NVFLASH. I used the 1.4ns timing set from the Qimonda spec sheet.

[4Qman]

VETDRMS
Did you see a benefit changing the timings to the 1.4ns set?

[VETDRMS]

VETDRMS
Did you see a benefit changing the timings to the 1.4ns set?

Minimal. Around 70 points in 3DMark05. Although it is more stable at 1144 with the tighter timings, go figure.

[scotty1991]

I think ill leave the timings alone and just go the volt mods, not till next month though, no point with my e2140 bottlenecking the whole system :D

[aussie-revhead]

Good stuff guys , can you tell me if the std mount for mcw60 fits these ? Or the GTx adaptor ?

:up:

[Boogerlad]

i think the std stuff that came with the mcw60 is compatible with the new g92

[Pacha]

mcw60 fit on the GT ref pcb, so as GTS uses the same, it will fit

[4Qman]

Running stock Air and V modded the VGPU.

Running 1.4v 880 ATI tool stable and artifact Free.

What you think safe 24/7 votage would be?

Also this card is DEFINATLY affected when i go over 52oC. Anythink under 50oC its sweet and no artifacts at 900core.

[tomati]

Temps are usually 34-38C.

Increasing vMem will just make it worse, decrease vmem to improve overclock. My memory is now at 1144 with 1.4ns timings and 1.900v.

Could you share your ram timing I will try them before doying the vmod, I like the idea that we have to decrease the voltage to doing better frequencies.

On stock volt but watercooled I'm limited @ 810 (GPU) 1998 (shader) and 1080 (ram). never go above 39°c most of time at 37°c in load and 32°c at idle.

[4Qman]

3Dmark 06 results :D

Q6600@4347mhz_Mach2GT_R410a-CrucalDDR2-Single_8800Gts512@913-2180-1035
8800GTS on Stock Air cooling + High CFM Blowing on back and Front of card.
Ambidnt 7oC

http://service.futuremark.com/compare?3dm06=4453078

EDITED - Futuremark link added :D

[aussie-revhead]

Happy New Year everyone !!!

Where is the score 4qman ? Mine has scored 16,050 but it should do more , volt mod time me thinks .

:up:

[4Qman]

Where is the score 4qman ? Mine has scored 16,050 but it should do more , volt mod time me thinks .

:up:

i edited the post man, sorry i was rushing hehe

[aussie-revhead]

Holy crap dude ... what am I doing wrong ? Sure you have 400MHz on me but 2500 points too . I couldnt even get 169.28 to work , did you get much of a gain from them ?

:up:

Edit - just noticed core clock , thats my next goal I think .

[4Qman]

Holy crap dude ... what am I doing wrong ? Sure you have 400MHz on me but 2500 points too . I couldnt even get 169.28 to work , did you get much of a gain from them ?

:up:

Edit - just noticed core clock , thats my next goal I think .

Ambient is around 7oC and i have hgh CFM on the card blowing at it.

Q6600 at 4.4ghz is helping however a massive jump with the core.

i run 1.4v idle 1.47v Loaded on GPU.
mem needs mod and ill get some better results.

the key is temps to get betetr core. anythink over 50oC and 900 core isnt stable even for 3d benching.

[Alex-Ro]

Currently having an 8800 GTS 512 mb running 24/7 800/2000/2100 all this on stock voltage only custom cooler from ZeroTherm,max temperature 54*c at full load.

http://www.zerotherm.net/eng/product/HC92CU-88.asp

I am currently considering a volt-mod only wondering a few things.I would like to have 900/2200/2100 for daily use.I am asking those who had already tested if this would be possible and if so at what voltage would it be safe? ;)

[Zeus]

1.41Vgpu idle, so 1.46V load, watercooled with EK fullcoverblock:

http://img.photobucket.com/albums/v620/Z3us/0615935.jpg

[gOtVoltage]

1.41Vgpu idle, so 1.46V load, watercooled with EK fullcoverblock:

http://img.photobucket.com/albums/v620/Z3us/0615935.jpg

Can you do one run @ 3.5ghz for me please ...I want o see what a E6600 does vs X2 6000+ @ 3.5ghz

Ill do bench with 901//149memory//2205 for comparison..:yepp:

[aussie-revhead]

Do you want one with my quad at 3.5 too ?

:up:

[gOtVoltage]

Do you want one with my quad at 3.5 too ?

:up:


Yes you too...

I believe with the exception of 2cores Vs 4cores the C2D and X2 should have simalar GTS SCORES on SM 2.0 and SM3.0...While the Quad @3.5ghz should be a few Hundred Higher on the SM 2.0 and SM 3.0...

This will give me a better Idea on the Scaling of the GTS on a system@ 3.5ghz //wether its Quad or Dual Core or X2 it should be real Close on the Sm Scores......

[aussie-revhead]

Yay !!!

I will test it for you later on , I am currently testing to see how low I can drop my vcore while remaining Prime stable at 3GHz .

:up:

[aussie-revhead]

OK I had a fiddle but I have cooling problems at the moment , I changed rad fans and I have lost some performance , so I couldnt quite pass at those clocks (though at 1.3vgpu) so I did this one at 3.5GHz 900/2150/1125

I will bump vgpu up and see if I get a pass at your clocks.

:up:

[nosboost300]

man.. all these g92 scores makes me want a 8800gts!

[aussie-revhead]

This one is better , not my best and not to the clocks requested , but nice anyway.

:D

[Zeus]

This one is better , not my best and not to the clocks requested , but nice anyway.

:D

Man, i'm really surprised you didn't break 18 yet.

I just took my SM2.0 and SM3.0 score and your CPU score, cause you have a quad.

I know you can easily run those clocks i did, i have a better one as well (16K2 :) )

I calculated with m3kk's 3d 06 calculator and here you have it:

[aussie-revhead]

Thats with all three of the best seperate scores together but I changed fans and have lost some stability now so I havent been able to beat that score until I sort out a solution for my cooling problems. Your card is stronger than mine , at least at the moment . My cpu isnt the best clocker either .

:shrug:

[Zeus]

Thats with all three of the best seperate scores together but I changed fans and have lost some stability now so I havent been able to beat that score until I sort out a solution for my cooling problems. Your card is stronger than mine , at least at the moment . My cpu isnt the best clocker either .

:shrug:

Bro, your CPU score is good enough for 18K, just check the 3d06 caculator screen i posted above.

I actually took my SM02 and SM03 scores and your CPU score.
Combined this gives 18K easy.

My card isn't stronger than yours, at least, i don't think so, mine needs lots of voltage and shader clocks aren't best.

The SM 2.0 and 3.0 scores i took were from an 15K9 run and were done at 901/2205/2300, you can run that too.

About your cooling problems, are you sure contact with the block is good, i had some trouble getting my EK fullcover mounted properly but eventually managed.

If you need help PM me. ;)

[aussie-revhead]

I could run those clocks with my huge single fan setup but I changed to 9x120 fans and its much hotter now , Im sure when I correct this it will run fast again.

:up:

[gOtVoltage]

I could run those clocks with my huge single fan setup but I changed to 9x120 fans and its much hotter now , Im sure when I correct this it will run fast again.

:up:

That is the Bomb......Youll hit 18000+

Now if you were to run @3.5ghz does the card score 13500-14000 for 3Dmark06....This will give me a good idea if the Oc scaling is the same on AMD and INTEL C2d and Qseries for SM 2.0 and SM 3.0..


Well my card @907 likes the shader Above 2160mhz...then when its at GP950nhz the shader like 2302-2350mhz...

Besides that mine at 3.5ghz gets an alright 13128+@ 907 and goes up wmore @950,,still working on it..Been playing with MOBo bioses....The newer ones on M2n32deluxe actually are benching a little slower,, In 3Dmark06

http://i168.photobucket.com/albums/u192/darrellsierra/13128bench30c_load.jpg

[Zeus]

I could run those clocks with my huge single fan setup but I changed to 9x120 fans and its much hotter now , Im sure when I correct this it will run fast again.

:up:

I don't see how 1 huge fan is better than 9x120, what rad do you have?

Got any pics?

[aussie-revhead]

My setup is .... unusual , Im using 3x360 rads in an external mount . Two rads are in series with an mcp655 for cpu , with a Fuzion. One 360 rad with just a DB-1 pump and EK full block on gpu. Before the 9x fans I was using a single car EOM thermofan off a jap car , adapted to fit , current limited and on a wire wound rheostat but the current draw still fried the wiring eventually , hence my experiment with 9x120 fans . The one car fan covers the three rads perfectly.

When I find a pic I will post it .

[ic3m4n2005]

@aussie-revhead: i know it's not ontopic but i have a short question. Did you notice any gains with using 2x360 rads for your quad ? I have a 2 loop system aswell and i'm considering adding a 240 to my cpu loop to get the max out of every cpu.

[aussie-revhead]

Now if you were to run @3.5ghz does the card score 13500-14000 for 3Dmark06....This will give me a good idea if the Oc scaling is the same on AMD and INTEL C2d and Qseries for SM 2.0 and SM 3.0..



Voltage - the run in post #36 is at 3.5Ghz or thereabouts .

Iceman - I did not test with the second rad so I cant tell you if it helped or not , I am thinking of converting to a single PA120.3 for each loop , or messing with pelts .

Zeus - using your sm2 and 3 scores is no use to me as I have never scored that high in either , but I will keep trying .

:up:

[Zeus]

Zeus - using your sm2 and 3 scores is no use to me as I have never scored that high in either , but I will keep trying .

:up:

That's what surprises me since you (used to) run the same speeds on the videocard.
Perhaps you need to do a little more tweaking?

How about running the big fan without rheostat, just running full throtle?

[aussie-revhead]

How about running the big fan without rheostat, just running full throtle?

Nah it would cool well but make a massive cyclonic noise at full speed , along with drawing 120W on its own lol.

:rolleyes:

I had the fan slowed down enough to be dead quiet on lowest speed and still not intrusive when flat out while rads were always cool , even when benching they hardly got warm . The extra heat has definately affected my max shader and mem clocks , but they are still decent .

:up:

[Zeus]

Okay guys, i had some time on my hands and decided to do some testing to see how the core and shaders responded to voltage.

Here's what i got, testing with ATItool, a single yellow dot meant not stable so i backed down:

http://img.photobucket.com/albums/v620/Z3us/scalingG92.jpg

So this is stable but i think i have a pretty weak core, especially the shaders won't clock very well. :(

Card was watercooled with an EK fullcoverblock.

Max temp at 1.525V loaded was 41ºC, ambient is 20ºC.

It's also interesting to see that the higher the voltage goes, the bigger the gap grows between idle and load voltage.

[aussie-revhead]

Those are nice temps , is that a single 240 rad for the whole rig ?

:up:

[Zeus]

That's right, one PA120.2 with 4x120mm Yate Loon in push-pull config.

A PA120.3 would be better but it does not fit in my case, or at least not very easy to say the least. :)

Does your card scale about the same?

[aussie-revhead]

Yeah something just isnt right with my water cooling , so Im going to dump the 3x360 rads and get a single PA120.3 then see what happens .

Anyway back on thread , my card scales funny I think , it clocks hard very easily and then hits a wall , so Im not sure if this is the 50 degree wall people are talking about or what . My next goal is to get temps down lower , like yours , and see how she clocks (along with getting a Q9550 to bump up my cpu score).

:up:

[STEvil]

aussie - i've noticed the same thing on mine.

But mine runs quite warm. 55c idle and up to 60c loaded at 1.25v idle.. using Half-Life 2 Deathmatch while measuring temps only so far.. some of the custom maps make GPU's scream ;)

[Lastviking]

aussie - i've noticed the same thing on mine.

But mine runs quite warm. 55c idle and up to 60c loaded at 1.25v idle.. using Half-Life 2 Deathmatch while measuring temps only so far.. some of the custom maps make GPU's scream ;)

Is that with stock fan?, what was your max 3dmark2006 clocks with stock volt and 1.25v?

[STEvil]

Yes, stock fan. Made it hit 67c today still all the same settings.

I havent tried pushing it farther on 1.25v yet. 850 was stable with 1.35 for quite some time but would eventually crash.