I just built a new rig with the specs I'll list below. It's the first PC I've ever built so i 'm an absolute noob.
Gigabyte 3850 512mb
Antec Sonata III
Earthwatts 500w PSU
Seagate Barracuda 7200.10 ST3250410AS 250GB
A-DATA (2 x 1GB) 240 pin DDR2 800 (PC2-6400)
Windows Vista Home Premium 32bit
The first time I booted up it worked fine I changed a few ninor things in the BIOS but didn't OC or change any voltages. Vista installed without a problem too. I installed the drivers for the 3850, mobo, and monitor also without a hitch. I was then looking around at the different programs Vista comes with, this is the first time I ever used it, when I came across the rating system. Everything looked fin getting about a 4.7/4.8 on everything except the gaming graphics and it said "newhardware detected refresh score" so I start to redo the scoring and as it beginds to test the media player it crashes. I deleted my graphics card drivers and reinstalled them and got the same results.
I then decided the score meant nothing and went to play the NWN2 game that came with the 3850 and it crashed again right after I finished makig my character. It then proceeded to continuously crash in the BIOS. I finaly got it to stop and my mobo said my E8400 was running at 90 degrees celsius.
I suspect either bad RAM or something with the PSU.
thanks for any help.
Oh by the way I watched a movie and listened to a few CDs in the Media center so thats working fine. I don't know if any of this helps, I'm just trying to be specific.
Temperature is a major issues, get the corrected asap, reseat the heatsink and make sure it's staying cool.
It sounds like a power issue if you can get into windows and run fine, but once you're in a game, the 3d environment kicks up and the card goes to load which require more power. In the bios, they have no cpu idle software, so temp shoots up to full load, I believe the same would happen to GPU? I've heard people saying artifacting in games and in the bios, but no windows, so maybe it works the same way? (Something I really want to find out). Try to take out your 3850, test it in a friends rig, or if you have another computer.
Make sure you run memtest also just to be safe, but I doubt you're having any memory issues. This sounds power/temperature related, the 90 is definitely too high for your cpu, but this sounds more like a power issue.
Uhm, dude, if your mobo says your chip was running at 90 degrees celcius, there's your problem. Your system looks good on everything else. If it only does it after a while when you start stressing the CPU, then your cpu is cool enough to do basic things, but heats up and there you go.
Did you fit the heatsink on right? Did you use thermal compound? Also, do you have any case fans to create airflow? All of these could be factors. I'd recheck my heatsink and thermal compound if I were you.
It could be power, but I'm thinking that the Antec would be more than enough.
I already reset the heatsink and used thermal compund but I'll try again scince I doubt it can hurt. I'll also test my 3850 in my old computer to make sure it's not that. If nothing happens then I'll try another PSU and see if that solves it.
I agree heat is a problem. If you already "reset" the heatsink i would try using this as a reset method.... a NEW razor blade, NEW thermal compound, Artic silver 5 is supposedly the best, and the highest precentage alcohol you can get at the drug store (91% or above).
remove the heatsink then proc from your pc carefully, use the spounge or whatever packing your proc was touching to set it on so you don't shock it or whatever, make sure you're grounded out too. and use the alcohol and razor blade to remove all thermal grease from both surfaces, then use a lint free cloth dampened in the alcohol to remove any particles from both surfaces, and without touching or contaminating both reapply new thermal grease and install...
i'm sure you prolly followed a procedure similar to this... but this is honestly the best way that i have found to make sure it's done right... and if it's another problem, you'll have a heatsink/proc installation that is flawless. :-p
if this doesn't help, my next assumption that it may be your graphics card/psu. either the psu doesn't have enough power to run at full speed, or your graphics cards is bad (i doubt). hope this helps? let me know of your progress.
check to see what speed the ram is operating at and the fsb. A mismatch (asynchronous bus) in ram and fsb mode may work but can cause crashing even if it passes memtest. ie, ddr2-800 and 1066 fsb on cpu is asynchronos. Dont trust the mobo to automatically get it right.
I ran Core Temp and it says the proc is running at about 50c so I'm assuming the mobo is not reading it right or something. I used Arctic 5 and the stock heatsink, I'm going to remove the thermal compound and redo it in a few minutes. I haven't tested the 3850 yet but doubt that's the problem. Ill check for a mistake in the RAM fsb.
If it still doesn't work I'm going to get a new PSU with more Amps on the 12v rails and hope that solves the problem.
The Intel HSF uses those push pins to mount the HSF. The HSF may not be properly mounted. Unfortunately, the best way to check proper mounting is to mount and check the HSF BEFORE you install the motherboard in the case.
I took my mobo out and reset the heatsink making sure the pins went in properly. I also used fresh thermal paste. The crashing has stopped and I'm getting much lower temps, only hitting about 45c under load with the stock heatsink.
The push pins were a stupid idea on intels part, screws would be SO much easier.