Calling all pros, opinions needed.

[kikaida]

Here are the specs of my brand new build.

Gigabyte GA-X48-DS4 (rev 1.3) f2 BIOS
Pentum Duo 3.16
G-Skill DDR2 (1066) 2GB x2
2 Seagates, SATA 3g, 250GB (non-raid)
2 Samsung DVD-RWs
PSU Thermaltake 750watt
Video card(s)
First one was...ATI 4870 1GB, DDR5 (powercolor)
Second one was...Nvidia GTX260 886MB (EVGA)

XP home and Office 2007 (all updates applied, including DX 9.0c)
All drivers installed and updated. Chipset for the mobo (X48) was gotten from intels site.

No sound card was installed yet (to avoid conflict)

The system appears to be very stable and runs very nicely. Event Viewer clean.

The issue is when I try to run a game. In this instance Warhammer Online.

For the ATI card, I would notice stuttering right off the bat and then a lock and then the card would generate and error and reset it self. The error was to the effect "The gpu has stopped processing information and has been reset" or something to that affect.

Before I moved on to the Nvidia card, the harddrive was reformatted and everything was reinstalled and updated.

Upon running the same app (Warhammer), initially it ran for about a minute and then corrupted with polygons.
This keeps happening, I could not get more than a minute in-game without corruptions.

In both instances the GPU temps were under 50c and no attempts at over-clocking were undertaken anywhere.

I even so much as lowered the clocks on the cards and reduced the game settings to "balanced" and still corruption occured.

At this point this is what I'm looking at and where your experience and wisdom comes in.

I am going to RMA the PSU. Although the specs are beefy enough, my thoughts are the 12v rails @ 18amps might not be cutting it. Although it should, I was leaning towards getting a PC Power Cool with a single 60amp rail. What do you think of these single high-amp rails for these power hungry cards? Good idea, bad idea?
Also, could it be my PSU is just defective to some capacity? I just don't know.

Second, I am going to RMA the ATI 4870 and move over to the Nvidia GTX260 as the support from Evga is absolutly stellar and should the problem continue, I know that I can count on them to see to it that I get a known-good card should it get to that point.

Now, if the problem happens with the new PSU & Video Card...Should I RMA my mobo? I ran a memtest all last night with zero errors and have done the same to the CPU and its doing its thing. Can issues on a mobo cause this? It runs fine outside of a 3d environment, has any one ever experienced an issue with a PCI-Express 2.0 slot? Is it even possible? If so, is it an extremely remote possiblity or semi-common?

I've been in IT for 10 years and it's stuff like this that just drives me nutty. I don't build scratch systems but have a pretty good idea of whats going on. So for those of you scratch building gurus, please lend me your input. What would you do at this point.

Thanks!

 

[lunyone]

Well I'm not sure if it's your PSU, GPU, or mobo. What is the full specs on your memory? Is it at the required voltage (1.9-2.2v)?? I had a similiar issue on a new build where a game would run for 15-120 min. before locking/freezing up. I ran memtest for hours, no issues. Then I ran Everest and had the computer completely reboot, anywhere from 1-2 hrs. I found that the voltage for the memory should've been at 2.1v to run at the given speed. The mobo that I was using, didn't support anything over 1.8v (stock setting), so I installed everything into a mobo that could support the higher voltages required. I then ran Everest again and it ran for +14 hrs. with no reboots!! I'm not saying this is your issue, but it is something to double check. Have you ran any other burn tests, other than memtest??

 

[bobbknight]

One of the specs I look at when I buy ram is the voltage required along with the cas number.
I try to buy ram with a cas of 4 at DDR2 voltage spec of 1.8v.
The last ram I bought was DDR2 4,4,4,12 at 1.8-1.9v now this ram I have been able to push up to the speed that I want it to go to with only a small increase in voltage.

 

[kikaida]

Thanks for getting back to me, I did check the voltage and it was at 2.1v which is the manufacturers recomended.

The specs are

Cas Latency 5
Timing 5-5-5-15
Voltage 2.0V - 2.1V

I briefly ran Prime95 and all seemed well.

 

[lunyone]

I'd recommend running Everest Ultimate Edition for a few hours to see if that creates anything. If it doesn't than I'd suspect your PSU might be an issue. If you could borrow a PSU from a buddy than you could try his out and see if that is it. Of coarse, you will want a PSU that can support the 4870. Or you can go and buy a new PSU, with the idea that you might return it if installing it doesn't change anything.

 

[kikaida]

Thanks for the reads and suggestions folks.

Well I figured it out while chomping on a slab of BBQ ribs.

I've already bought a GTX 260 to test out if the 4870 was the issue and that bombed too. Went to Fry's this AM and returned it. Sort of like the PSU swap.


There are just too many variables at this point so here's my solution.
Mobo, PSU, Video Card and Sound Card are all going back.

I'm gonna wait for the credit then start over with those components.

I'm going with:

Evga 780i (one generation back to avoid "growing pains")
Evga Gtx 260 216 886MB
PCPC 750watt single 12v rail @ 60amps
Sound Blaster Xtreme Gamer 7.1

The cool thing is Evga's support was amazing. I can't belive I actually was talking to some one at 11:30PM last night!!! Not to mention that they are about a 45 minute drive from where I live. So if things get to out of hand, I can always drive down and pay them a visit. Lifetime warranty doesn't hurt either.

Well, thank goodness I work in IT and can work on this disassembly project while at the office.

Newegg has never let me down with returns before and I hope they don't start now.

 

[lunyone]

Well if your going that extensive, I'd just get another PSU and try that. Than you don't get charged for RMA and such.

 

[covertxt5]

I would say the problem is a setting in the bois.