GTX580: 1.025VID: 950/1900/2300: stock cooler: guaranteed?!?

[NapalmV5]

@Dumo that close and still not the same vids! son of butch! lol

[Dumo]

@Dumo that close and still not the same vids! son of butch! lol

The 1.0V not that good on air, probly will shows it on cold

@ Asmola....Can you share the bios? Thanks

[Asmola]

@ Asmola....Can you share the bios? Thanks

http://www.mediafire.com/?ngcgkgkklkn1u6m

There you go.

[Dumo]

http://www.mediafire.com/?ngcgkgkklkn1u6m

There you go.

Thanks Asmola

[NapalmV5]

The 1.0V not that good on air, probly will shows it on cold

if you got a plate handy install it on the back of the gpu

some gtx580s badly need a backplate

one of my 1.025vid was doing 40+C above the other 2 @ load and 10+C @ idle

[Dumo]

if you got a plate handy install it on the back of the gpu

some gtx580s badly need a backplate

Will try, thanks Napalm
Asus intalled plastic back plate which probly make it worse

I'm looking for new gpu pots, MM's gpu pots already out of date. Probly both cards will show their hidden power when naked and cold

Heres out of the box sli run with mismathed bioses..

[Demo]

I got my Evga GTX580-SC a couple of hours ago and just installed it, stock VID is 1.082v which is pretty high from what I've seen from the rest. I guess 950Mhz might be very hard to reach or even 900Mhz ?


Makes me thing bad batches are sent to my country

How much should be the maximum 24/7 voltage that can be pushed on stock cooling, say something like limiting the fan speed to about 80% tops, 1.100v ?

[vern]

Bios vmod with NiBiTor worked fine and now im able to use 1.213v on my DCII without losing card's stability.

That cold slow bios also dropped my card's P0 Vid to 1.000v, so it's not constant.. but now using original DCII bios with unlocked voltage.

VID can alter in the BIOS

[OC Nub]

Nice results guys!

I am looking to get a GTX580. I saw where Giga will make an Soc version of the 580. They bin the GPU's for the Soc cards, so that seems to be the only choice for a guaranteed low vid.

Would it be best to wait for the Gigabyte, or does anyone know of another brand available in the US that has binned gpu's?

Also, there is a 480Soc available at newegg for 100.00 less than the reference 580. Is there that much difference between the 480 and 580 to make it worth spending the extra 100.00?

It will be used for 3d benches, not gaming. I need roughly 55k points in 05. Not sure if this is possible with a single card?

Thanks

[Dumo]

2 CUIIs retail samples from Newegg

The top one overclocked to 957 maxed (s/n 1213XXX) and the bottom one (s/n 1212XXX) can do a lot more with the same oc. settings...made me believe that same 580 CUII gpu will varies in term of oc on stock hsf.




CUII #0 with base Voltage @ 1.00V at 1.15 VGPU




CUII #0 with 1.80VGPU



CUII #1 with base Voltage @ 1.013V at 1.15 VGPU



Dang.....I started binning this CUII

[Demo]

Nice, both cards have pretty low VID.. maybe nvidia is binning 0.995v chips as well ?

[vern]

2 CUIIs retail samples from Newegg

The top one overclocked to 957 maxed (s/n 1213XXX) and the bottom one (s/n 1212XXX) can do a lot more with the same oc. settings...made me believe that same 580 CUII gpu will varies in term of oc on stock hsf.

CUII #0 with base Voltage @ 1.00V at 1.51VGPU


CUII #0 with 1.80VGPU

CUII #1 with base Voltage @ 1.013V at 1.51VGPU


Dang.....I started binning this CUII

Very nice card

[Asmola]

2 CUIIs retail samples from Newegg

The top one overclocked to 957 maxed (s/n 1213XXX) and the bottom one (s/n 1212XXX) can do a lot more with the same oc. settings...made me believe that same 580 CUII gpu will varies in term of oc on stock hsf.

CUII #0 with base Voltage @ 1.00V at 1.51VGPU

CUII #0 with 1.80VGPU

CUII #1 with base Voltage @ 1.013V at 1.51VGPU

Dang.....I started binning this CUII

Are those voltages correct?? 1.51vGPU?

[OC Nub]

2 CUIIs retail samples from Newegg

The top one overclocked to 957 maxed (s/n 1213XXX) and the bottom one (s/n 1212XXX) can do a lot more with the same oc. settings...made me believe that same 580 CUII gpu will varies in term of oc on stock hsf.



CUII #0 with base Voltage @ 1.00V at 1.51VGPU



CUII #0 with 1.80VGPU


CUII #1 with base Voltage @ 1.013V at 1.51VGPU


Dang.....I started binning this CUII

Nice results, and great 2600k you have.

Im going to order in the next hour or so to get the card by the weekend. I have it narrowed down to 2 cards.

[Dumo]

Are those voltages correct?? 1.51vGPU?

I meant 1.15V

[vern]

Are those voltages correct?? 1.51vGPU?

I think 1.05V

[Asmola]

I meant 1.15V

You almost gave me heart attack! Any results with 1.213v?

[Dumo]

You almost gave me heart attack! Any results with 1.213v?

Haven't try your bios yet.

On stock air HSF oc this CUII really is a gamble (like I7 2600K), dunno if Asus binned it or not. Probly the one with high leakage, low air oc will shine when put it on cold

Imo the lower the vid, the higher the leakage. So, it can run on stock Vgpu?

[Demo]

My VID is 1.082 and its currently running @ 1.088 by default, would it be safe for me to lower those to say 1.050v or its already fixed to be stable at 1.082 at default clocks?

[nolonger]

My VID is 1.082 and its currently running @ 1.088 by default, would it be safe for me to lower those to say 1.050v or its already fixed to be stable at 1.082 at default clocks?

Sure, no problem in trying lower voltage.

[Demo]

Sure, no problem in trying lower voltage.

OC scanner seems to be not reaching 100% GPU usage, I've removed the usage limit under settings and it hovers around 90~97%

Or would there be another more effective app / method to test for undervolting ?

[nolonger]

OC scanner seems to be not reaching 100% GPU usage, I've removed the usage limit under settings and it hovers around 90~97%

Or would there be another more effective app / method to test for undervolting ?

Try gaming or Vantage. I wouldn't recommend fixed image tests (Furmark, OC Scanner, ATI Tool) to test for instabilities.

[donmarkoni]

I wouldn't recommend Furmark (and similar) application at all, especially with OCP turned off. You could burn your VRM.

[Chickenfeed]

I've found an extended Heaven Extreme Tes run (like 30min or so ) with max AA along with Vantage (max res, settings, aa, resolution pending monitor ) with feature tests 3x loop and 3DM11 on demo + test usually shows instability. There are never any guarantees with this stuff but in my own experiance if a voltage / clock is unable complete all these tests, instability is likely. From there I simple do the things I normally do ( games, decoding / encoding etc. ) and if all is well you have your answer

On a side note I've found Just Cause 2 to be quite effective at stressing hardware and ironing out instablility. Sadly There appears to be some inherent driver instablilty and it has ocassionaly crapped itself on me on more than 1 machine ( another with a 470 ) even at stock.

[Demo]

Try gaming or Vantage. I wouldn't recommend fixed image tests (Furmark, OC Scanner, ATI Tool) to test for instabilities.

Any specific test to loop under Vantage?

I wouldn't recommend Furmark (and similar) application at all, especially with OCP turned off. You could burn your VRM.

How can I check if my OCP is turned on or ff?

I've found an extended Heaven Extreme Tes run (like 30min or so ) with max AA along with Vantage (max res, settings, aa, resolution pending monitor ) with feature tests 3x loop and 3DM11 on demo + test usually shows instability. There are never any guarantees with this stuff but in my own experiance if a voltage / clock is unable complete all these tests, instability is likely. From there I simple do the things I normally do ( games, decoding / encoding etc. ) and if all is well you have your answer

On a side note I've found Just Cause 2 to be quite effective at stressing hardware and ironing out instablility. Sadly There appears to be some inherent driver instablilty and it has ocassionaly crapped itself on me on more than 1 machine ( another with a 470 ) even at stock.

Noted, Max Tes + AA under Heaven
Vantage Feature tests looped at max settings and 3DM11 demo

I want to perform some of these benchmark tests or synthetics before I go with my real world gaming test for stability.

EDIT:

BTW, how can the vantage feature tests be good since the down time inbetween tests is quite long.. its not really a single flow of different tests ?
Which game test on 3DM11 is very sensitive to overclocking, something that we can loop several times or so..

EDI2:

Initial testing for 1.000v I got 2 stable loops on 3DM11 Demo with performance preset, all sliders to max out side from MSAA which i left at 2x and running at 1680x1050.