4 x 18A 12v rails for a GTX 260 Black Edition?

[duffbeer666]

First off I have to say that this forum is amazing – such a wealth of knowledge!

Hopefully, someone will be able to help me with a frustrating problem. I should point out that I am not a newbie – I have been building my own PCs for a few years now but this is my first Intel system (I think my sig is up to date).

I have had a few problems with stuttering and slow HD access but I think (touch wood) I have sorted this out with a reinstall and newer raid drivers.

Everything seems to be running fine in Windows – Internet Explorer, Win Media Player, etc all running smooth. The problem comes when I play Crysis (the only game I have at the moment). I have been playing through the game all week and really enjoyed it, it seems to run pretty good at 1680 x 1050 4x AA with all other settings on high. I get the odd slow downduring intense battles but I attributed that to the screen being very busy at that time.
On Thursday night I reached the last level (I think! The one on the flight deck of the boat). During this level I get a complete failure at some random time but usually when there is a big explosion or something like that. The screen goes solid grey and the last second of sound is looped, at which point I can do nothing but hit the power button and reboot.

I have tried a few different things over the last 2 days. I have done a full chkdsk - takes a looong time on a 1TB drive! I have reinstalled the game and Nvidia driver (180.48) and I have done a full memtest but none of these things have made any difference.

The only other thing I can think to blame is my PSU. When I bought my components I checked the PSU amperage and it seemed to be OK so I thought I could save some money by keeping my previous unit. Nvidia recommend 36A for a GTX 260 and 40A for a GTX 280, and my PSU has 52A.

The PSU is a Seasonic S12 650w and I have just noticed that it has 4 12v rails at 18A each.
I assume the PCI-E connectors on the PSU would have a rail each which would give the GPU 36A, but I assume that as with most things Seasonic exaggerate their figures slightly so it would never quite achieve its maximum output. Also, as the 260 is heavily overclocked I am guessing it might have the same sort of power requirements as a 280.

So here is the real question: Do you guys think that the PSU cannot cope with the demand of Crysis + OC’d GTX 260 when the screen gets very busy with explosions and flying aliens? Or am I barking up the wrong tree?

I would be really grateful for some advice as I don’t want to spend £100 on a PSU only to find it does the same thing.

 

[duffbeer666]

Thanks for your replies.

So you really don't think the PSU is borderline? I reckon it has to be at least 2 years old.

I have the latest patch and all my drivers are as new as I can find. I hope you are right but I'm not convinced.

The Furmark website says it is for XP and 32 bit Vista - no mention of 64 bit.
I could run the Futuremark test, is that as good?

 

[duffbeer666]

I haven't tried to run Furmark yet due to the fact that Vista 64 bit was not mentioned. Do you think it is caoable of reproducing the power surge that is required when the game suddenly gets busy? Also, if the game was at fault surley it would just dump me back into windows rather than give me a complete failure?

Not trying to be argumentative - honest! I'm just trying to get my head around this one.

 

[duffbeer666]

Furmark does not work at all for me - the furry ring thing (or whatever you want to call it) does not move at all.

Any other suggestions?

What will really test my PSU?

 

[duffbeer666]

I just got another game to try - Call of Duty 4.
It seems this game is doing exactly the same thing as Crysis, anybody got any suggestions as to what might be happening here?
It's driving me nuts - I cannot play games!

 

[duffbeer666]

I have just tested my card and it seems that the max temperature during Far Cry 2 is 72 degrees. Surely this can't be simply a heat problem?? Gotta sort it out though coz it's driving me nuts.

 

[duffbeer666]

Any more news duffbeer666?

I am having exactly the same problem with my EVGA GTX260-216 SSC. In CoD4 and Crysis my screen goes a solid grey with looped sound and I have to restart. Also sometimes my motherboard gives an error beep - 1 long beep followed by 3 short beeps.
My system is 100% stable in Windows and passes IBT and Prime95. I'm starting to think the card is faulty.

I made a thread about my problem here; http://forums.overclockers.co.uk/showthread.php?p=13117706&posted=1#post13117706

It's nice to see someone else having the same issues because up until now everyone has just suggested changing drivers!
Maybe we should work through the problem together to see if we can solve it.

I have since played all the way through Crysis Warhead and had no crashes but other games still don't work reliably. I guess I was just lucky with Crysis Warhead.

It is just so random, I don't get it. I think the only way is to keep testing until I see some sort of pattern but to be honest it takes all the fun out of gaming becasue I am constantly waiting for it to die on me.

The only thing we seem to have in common is a very similar motherboard, maybe we should look at this?

 

[duffbeer666]

Totally sympathise mate. It's the completely sporadic nature of the problem that's really getting to me. I had nearly a week of no problems at all then last night BIA-Hells Highway crashes three times in a row in the same level. Then I can't get it to crash again for love or money even though I play the same level over and over.

Tried changing drivers, monitoring temps, etc. (never gets above 78C, mostly around 75C). I suppose some might live with it given it's fairly rare but after having such a completely stable system with my old 8800GTX this is niggling away at me like mad.

I know if I send it back they'll test it for 20mins and say there's no problem and probably charge me for the trouble.

Wow, you really do have an identical problem to me!! Thank god I'm not alone on this one anymore.

I notice your motherboard is different so I guess that blows my previous theory out of the window. I guess we are looking at a GTX260 problem, but I am worried in the same way as you that if I send it back they will not test it thoroughly and I will be no better off.

Drivers seem to make to real difference at all so maybe it is something to do with bios settings?

Has your motherboard got a P45 chipset?

 

[duffbeer666]

I think I might need to RMA it, but I bet you the shop will test it for ~1 hour and say it's OK and charge me.

That's exactly what I am worried about. How the hell am I going to be sure that they will confirm the fault. It's really not fair to be put through this hassle when you spend a lot of money on a graphics card.

 

[duffbeer666]

I haven't got BIA Hell's Highway so I can't test out your theory.
My crash is pretty predictable in Crysis if you have that.

Your feedback about how they will test the card is really bad news - I really don't believe they will be prepared to give it the time to prove the fault, and I cannot stand the idea of being without it for a potentially unlimited amount of time.

I guess I will have to replace the card myself - I can assure you that this is the first and last XFX card I will ever buy.
Has anyone experienced this issue with an EVGA? I thought they were supposed to be the best.

 

[duffbeer666]

I have had enough with this stupid XFX card so I replaced it with a 55nm EVGA SSC and all my previous issues are gone.

Now that I know for certain that the GPU was at fault I just have to find a way of getting the retailer to test it properly - it sounds like you guys are having some serious problems with that!