**Gigabyte Super OC GTX 480's - VERY LOW PRICES!**

[AzZ]

Here's the update I promised.

Short version:

Brand new Corsair HX 750 did not solve artifacts. The card is unstable at factory OC. Gigabyte's 'GPU guantlet sorting' has failed.

Slightly longer version:

As I will be building a new system next month I was going to be getting a HX 750, so after some reading and recommendations regarding the 480 SOC, it seems the problems could be because the card was not getting enough power. I ordered my new PSU early so I could be 100% sure. Got the PSU, hooked everything up and began testing.

My results:

Heaven 2.5: Artifacts (video is being uploaded)
Crysis (1680x1050/DX9/custom 'very high' settings): Artifacts seen (tested on 'Assault')
Metro 2033 demo (max settings/DX9): Artifacts seen
MSI Kombustor: Artifacts seen

and the big one...

OCCT 3.1.0: 56284 errors after 2 minutes.



All those tests were done at my stock voltage of 1050mV. I have also increased the voltage to 1075mV and the card still artifacts.

As I'm sure you are all aware the HX750 is a quality unit so I cannot blame it. The card is therefore unstable and should be RMA'd.

What do you think..?

 

[whipgeez]

Here's the update I promised.

Short version:

Brand new Corsair HX 750 did not solve artifacts. The card is unstable at factory OC. Gigabyte's 'GPU guantlet sorting' has failed.

Slightly longer version:

As I will be building a new system next month I was going to be getting a HX 750, so after some reading and recommendations regarding the 480 SOC, it seems the problems could be because the card was not getting enough power. I ordered my new PSU early so I could be 100% sure. Got the PSU, hooked everything up and began testing.

My results:

Heaven 2.5: Artifacts (video is being uploaded)
Crysis (1680x1050/DX9/custom 'very high' settings): Artifacts seen (tested on 'Assault')
Metro 2033 demo (max settings/DX9): Artifacts seen
MSI Kombustor: Artifacts seen

and the big one...

OCCT 3.1.0: 56284 errors after 2 minutes.



All those tests were done at my stock voltage of 1050mV. I have also increased the voltage to 1075mV and the card still artifacts.

As I'm sure you are all aware the HX750 is a quality unit so I cannot blame it. The card is therefore unstable and should be RMA'd.

What do you think..?

i'd def rma that, that is unacceptable in my opinion

 

[Buckster]

not good Azz

whatmobo do you have BTW?

anyone else having problems happen to have a PCI-E 1.1 mobo like me ?

 

[AzZ]

I have an Asus P5NE32-E SLi Plus. I don't think it's PCI-E 2.0. I don't think that this should be a problem though as PCI 2.0 are compatible with earlier versions. Also if this was causing the problems then surely the 570 you had would also produce errors..?

These artifacts are simply being caused by an unstable OC. It doesn't error at stock 480 clocks. I've done enough testing to be confident the card is at fault here. If Gigabyte state NFF like when you sent yours in then I will seriously question their test procedure. This was my first GB purchase and I'm not happy at all.

Here is the video I recorded earlier:

http://www.youtube.com/watch?v=ZvR2pfFaPUI

 

[ToucHDowN]

Here's the update I promised.
Short version:
Brand new Corsair HX 750 did not solve artifacts. The card is unstable at factory OC. Gigabyte's 'GPU guantlet sorting' has failed.
Slightly longer version:
As I will be building a new system next month I was going to be getting a HX 750, so after some reading and recommendations regarding the 480 SOC, it seems the problems could be because the card was not getting enough power. I ordered my new PSU early so I could be 100% sure. Got the PSU, hooked everything up and began testing.
My results:
Heaven 2.5: Artifacts (video is being uploaded)
Crysis (1680x1050/DX9/custom 'very high' settings): Artifacts seen (tested on 'Assault')
Metro 2033 demo (max settings/DX9): Artifacts seen
MSI Kombustor: Artifacts seen
and the big one...
OCCT 3.1.0: 56284 errors after 2 minutes.
All those tests were done at my stock voltage of 1050mV. I have also increased the voltage to 1075mV and the card still artifacts.
As I'm sure you are all aware the HX750 is a quality unit so I cannot blame it. The card is therefore unstable and should be RMA'd.
What do you think..?

That's bad. I've been talking to Gigabyte the last days, via my vendor (he's forwarding me their emails). GB says that it is clearly a power (PSU) problem. When I told him to tell them that I own one of the best PSUs available, GB replied saying that (among other)

We suggest end user to check this issue from PSU first.
If he could find another higher WATT PSU to test it, it will be a faster way to verify this issue.
At the same time, end user have to choose the power connecter with higher currency.
For some high WATT PSU, it might have 4pcs 8pin connecter. But the currency might not the same.
For example, there might be 2 connecter with 18 A, and other 2 connecter with 25A.
If end user just install single N480SO, it’s obviously to use 25A to keep provide enough power for VGA.
If they connect with 18A connecter, it will not as stable as connecter with 25A.
Some connecter has been designed to share power to CPU, and it will also possible to make lower power to VGA.

Checked my PSU and it doesn't even have connectors that produce 18A. They're all at 25A.

For all of you having problems (I know you're experienced users), please cross-check the possibility of having connected the 480 soc on a "18Amper slot" of your PSU.

For me, problems remain if only I use stock voltage (1.035v). At 1.088v everything is extremely good (amazing card, btw!!!).

 

[Buckster]

I have an Asus P5NE32-E SLi Plus. I don't think it's PCI-E 2.0. I don't think that this should be a problem though as PCI 2.0 are compatible with earlier versions. Also if this was causing the problems then surely the 570 you had would also produce errors..?

These artifacts are simply being caused by an unstable OC. It doesn't error at stock 480 clocks. I've done enough testing to be confident the card is at fault here. If Gigabyte state NFF like when you sent yours in then I will seriously question their test procedure. This was my first GB purchase and I'm not happy at all.

Here is the video I recorded earlier:

http://www.youtube.com/watch?v=ZvR2pfFaPUI

Touchdown- thats what I thought - my 12V railson my Tagan were only 20A each

I then bought a Corsair HX850....70 odd amps on the 12V rail- single rail - just as many errors

re the query re PCI-E 1.0/2.0 etc- I read that 2.0 can provide more power on its PCI-E slot than 1.0 - Gigabyte specs state 2.0 - normally or as Wikipedia says "most" 2.0 cards should be compatible with 1.0 - but perhaps the GB 480 SOC is relying on that extra power ? re my 570 - perhaps it didn't use as much current off the PCI-E slot? perhaps GB's fancy 14 or whatever it is phase voltage regulation does something different

otherwise I can't understand it - I could understand if GB just ran Furmark and didn't spot the odd error - same with Heaven

but even with my new PSU - I get very obvious flashing (very obvious) in Furmark- and lots of Artefacts in Heaven

wheras GB - and I believe them ran Furmark for 4 hrs with no artefacts, and ran Heaven looped with no artefacts, and they gave me the screenshots to prove this was at both stock core speed (820) and stock volts (1.038V)

which makes me think theres something more fundamental causing all our issues

 

[AzZ]

If that is the case then Gigabyte should clearly state that the card is not compatible with anything lower than PCI-E 2.0. They specify a 600w PSU and thats it. I don't think that they would make it so it's not backwards compatible and then not say anything about it, and surely if it was something to consider they would say "right check your PSU, what PCI version are you running? Version 1.0 you say? Theres your problem."

Buckster are you going to test your card in another system? I guess that will be the only way to truely know.

For me I guess the only thing left to do is RMA it and see what happens. Does anyone have the Gigabyte phone number handy? I'd like to see what the tech people say.

 

[Buckster]

If that is the case then Gigabyte should clearly state that the card is not compatible with anything lower than PCI-E 2.0. They specify a 600w PSU and thats it. I don't think that they would make it so it's not backwards compatible and then not say anything about it, and surely if it was something to consider they would say "right check your PSU, what PCI version are you running? Version 1.0 you say? Theres your problem."

Buckster are you going to test your card in another system? I guess that will be the only way to truely know.

For me I guess the only thing left to do is RMA it and see what happens. Does anyone have the Gigabyte phone number handy? I'd like to see what the tech people say.

I will try and test in another machine when I get time

PCI-E is only a guess - it may be a red herring - PSU was for me

thing is- GB have failed 2 cards sent in for RMA so far....mine passed...

they sent me 20 odd screenshots and photos as proof

yet in my system- with a now decent PSU - I'm getting artefacts at stock all over the place

therefore my only other wonder is my mobo - which is 4-5 years old- P965 chipset

nothing else it could be,RAM, CPU etc couldn't effect the GPU

 

[sid488]

i have 2 asus gtx 480s and my default voltage on both cards are different 1 is at 1.050v and the other is at 1.088v both cards are at stock settings but thats there default voltage for my cards so i force them to 1.000v in msi afterburner, but they easily do 840mhz on the core clock @ 1.050v cant remember what i set the memory to.

 

[therutz]

Buckster your more than welcome to try it in my rig still mate

I have a 1000w hyper M series psu

Look at the rails



I have 2 of the cables with two connectors on each, IE 6 pin and 8 pin. one cable to my sparkel 480 and the other cable to the 480 SOC and the third pic-e 6+8 is redundent

There is only 22a on the first two rails and yet i get no artifacting with a 1050v bump on the soc (did with it at stock 1025v's)

EDIT not sure if this is usefull or im understanding it right but the 12v 1 to 3 are for PCI-E and 12v 4 is for HDD and suchlike yes? if so then the 480 will be ok with a current of 22A on a 12v pci rail yes? (If im wrong please tell me as its all a little confusing to me)

 

[Atom]

Mine was artifacting on P67 chipset.. Downclocked to 760 core and it ran fine..

 

[Buckster]

oh not good Atom

either so much for cherry picking

or the mobos they test them on are "unusual" in some way

 

[Cleeecooo]

Mine was artifacting on P67 chipset.. Downclocked to 760 core and it ran fine..

RMA

oh not good Atom

either so much for cherry picking

or the mobos they test them on are "unusual" in some way

mine was fine?
have you tried on another system yet?

 

[DTWD]

Well I got an email today. OCUK are sending my card back to me as no fault.

Apparantly the card topped out at 95 degrees in OCT, where I had it at 103 and still climbing.

Obviously I have a £10 bill.

Should I ask for the OCCT output results?

 

[Jeps]

Had mine returned by gigabyte and it looks like it passed thier tests. I'm wondering what setup they use to test it. I only get problems when my temps get up above 80 degrees and if they are testing in a nice air conditioned lab on an open bench i'm pretty sure they wont be hitting the temps that the average user will see at home.

Maybe i'll RMA it to OcUK and see what their testing department says? All they can do is send me out a new one though and there is no gaurentee that it will be any better.

 

[therutz]

I think its disgusting how they are sending cards back that are so obviously faulty
Even mine was at stock volts,

cleeecooo you were a very lucky puppy getting what you did shame others are being flobbed off with "theres nothing wrong with it"

 

[Buckster]

Had mine returned by gigabyte and it looks like it passed thier tests. I'm wondering what setup they use to test it. I only get problems when my temps get up above 80 degrees and if they are testing in a nice air conditioned lab on an open bench i'm pretty sure they wont be hitting the temps that the average user will see at home.

Maybe i'll RMA it to OcUK and see what their testing department says? All they can do is send me out a new one though and there is no gaurentee that it will be any better.

did you get a set of screenshots to prove like I did ?

perhaps overclock is very picky on overall system....

I mean I can't even getmine stable at 820 at any voltage.....

yet even the "worse-case" stock 480s seem to manage 850 odd< 1.11V

which makes me think something else is going on here

 

[AzZ]

I've sent an email to Gigabyte for my RMA. I also posted on the GB forums and asked if there was any know incompatibility with motherboards and there doesn't seem to be any known issues.

I meet the requirements to run the card therefore if I get errors only when running it at their factory overclock the cards faulty (technically the OC is unstable but I didn't pay to run it lower than what is advertised), simple as that. If they can run OCCT for 30 mins without a single error then either they are not using my card or have some super high tech setup and liquid nitrogen cooling...

I will keep you updated.

 

[yoink]

My PC now has a Q6600 @ 3.33Ghz and 4GB DDR2 RAM. Still not performing particularly well...

Also, my replacement 480 SOC still has thousands of errors in OCCT and artefacting. I've bumped the voltage to 1.075. I don’t really notice any artefacting in games now, (though I haven't looked too closely) and OCCT still has thousands of errors. I just can’t be arsed sending it back again. I have no confidence what so ever that I'll receive a working one :/

 

[Cleeecooo]

Had mine returned by gigabyte and it looks like it passed thier tests. I'm wondering what setup they use to test it. I only get problems when my temps get up above 80 degrees and if they are testing in a nice air conditioned lab on an open bench i'm pretty sure they wont be hitting the temps that the average user will see at home.

Maybe i'll RMA it to OcUK and see what their testing department says? All they can do is send me out a new one though and there is no gaurentee that it will be any better.

I would suggest you start a thread in the customer services section of the forum. If Gigabyte find the fault, OcUK are extremely quick with refunding you the money they charged.

I think its disgusting how they are sending cards back that are so obviously faulty
Even mine was at stock volts,

cleeecooo you were a very lucky puppy getting what you did shame others are being flobbed off with "theres nothing wrong with it"

I was certainly lucky by the looks of it. Is there something different about my system than any of yours?

I've sent an email to Gigabyte for my RMA. I also posted on the GB forums and asked if there was any know incompatibility with motherboards and there doesn't seem to be any known issues.

I meet the requirements to run the card therefore if I get errors only when running it at their factory overclock the cards faulty (technically the OC is unstable but I didn't pay to run it lower than what is advertised), simple as that. If they can run OCCT for 30 mins without a single error then either they are not using my card or have some super high tech setup and liquid nitrogen cooling...

I will keep you updated.

I was also told they didn't know of any errors regarding mobo incompatibility. just try it at GTX480 stock speeds and see what happens. I mentioned that I did this and maybe that was what got the to replace my card? I don't know

Good luck to all of you