Looking for some final input. I believe I've narrowed it down to the card, but need to be sure.
Here are the system specs:
asus p5l-mx 945g chipset
e6400
2x1gb buffalo firestix ddr2-800 5-5-5 @ 2.1v
evga 7900gs ko
ultra 500w / antec neo he550
stock cooling all around
Here's the situation. I have everything at stock. I can run orthos stable for 20 hours. Ram is powered by 4:5 (tried 1:1) as well, so effectively only 666mhz (vddr2 only goes up to 2.0v in bios, so wanted to rule that out). Also, I suspect that the other dividers are faulty, as running 266fsb and 711ddr2 2.0v is unstable and bsod's right on bootup. pcie/pci lock doesn't fully function on this board, so everything in that regard is kept at stock for the time being. Memtest ran overnight without issue at 666ddr2 spd 2.0v. Chips are believed to be d9gmh.
Here's the problem. I try running orthos in lowest priority, and 3dmark (03/06) on loop and it locks up after a while. I've tried stock evga ko speeds of 500/690, overclocked speeds of 580/800 (no voltmod yet and no artifacts), as well as reference 7900gs speeds of 450/660. All result in crashing. Load temps rise to 69-70'c, which is normal for stock evga cooling. I've gotten an antec neo he550 to test with for the time being to rule out the ultra 500w, and it didn't help the situation any at all.
Am I right to have ruled the evga 7900gs as the culprit? I know I could probably try a voltmod and see if I can any stability from it, but don't feel it's worth the effort, and don't want to jeopardize any rma. So in short, it won't be stable in 3dmark with stock everything. I've tried two installs of windows, and same result. I figure it shouldn't be the motherboard since pcie is powered mostly through psu, so should eliminate motherboard power supplying issues.
Here are the system specs:
asus p5l-mx 945g chipset
e6400
2x1gb buffalo firestix ddr2-800 5-5-5 @ 2.1v
evga 7900gs ko
ultra 500w / antec neo he550
stock cooling all around
Here's the situation. I have everything at stock. I can run orthos stable for 20 hours. Ram is powered by 4:5 (tried 1:1) as well, so effectively only 666mhz (vddr2 only goes up to 2.0v in bios, so wanted to rule that out). Also, I suspect that the other dividers are faulty, as running 266fsb and 711ddr2 2.0v is unstable and bsod's right on bootup. pcie/pci lock doesn't fully function on this board, so everything in that regard is kept at stock for the time being. Memtest ran overnight without issue at 666ddr2 spd 2.0v. Chips are believed to be d9gmh.
Here's the problem. I try running orthos in lowest priority, and 3dmark (03/06) on loop and it locks up after a while. I've tried stock evga ko speeds of 500/690, overclocked speeds of 580/800 (no voltmod yet and no artifacts), as well as reference 7900gs speeds of 450/660. All result in crashing. Load temps rise to 69-70'c, which is normal for stock evga cooling. I've gotten an antec neo he550 to test with for the time being to rule out the ultra 500w, and it didn't help the situation any at all.
Am I right to have ruled the evga 7900gs as the culprit? I know I could probably try a voltmod and see if I can any stability from it, but don't feel it's worth the effort, and don't want to jeopardize any rma. So in short, it won't be stable in 3dmark with stock everything. I've tried two installs of windows, and same result. I figure it shouldn't be the motherboard since pcie is powered mostly through psu, so should eliminate motherboard power supplying issues.