Computer continuously locks up while playing games

Computer continuously locks up while playing games though not when surfing the web/using general apps. My first experience with this issue occured when I built the pc with 64 bit vista ultimate. At the time, I was using an 8800 GTS and first noticed the error in AvP 2 (quite an old game), next I attempted Star Wars Battlefront 2 with the same results. Finally, I tried to use WoW on the machine and still had the issue. After a few BSODs it seemed that the graphics card was at fault.

As a result, I bought a new 4870 card from ATI and encountered the same issues. Next, I checked and changed all RAM voltages and timings to manual settings after which the errors still occured. Next, I installed XP and attempted to game out of a 32 bit xp pro. However, I STILL encountered the issues there. I know it is not the video card, or the PSU, I am in the process of testing the RAM though the errors occurs regardless of the amount in the computer or the DIMM they are in.

If anyone has any input I will gladly hear it because I just want to get this problem solved.

My system Specs:
Vista Ultimate 64 bit/XP Pro 32 bit
Foxconn A7DA-S Mobo
AMD Phenom 9950BE (not overclocked yet)
4 X 2 1066 corsair dominator RAM
ABS 900W modular PSU (tested with PSU tester)
3 500G WD HDDs in RAID 5
Creative X-Fi Elite Pro PCI sound card
Visiontek 512MB Radeon 4870

Alright I did manage to resolve that issue as far as I can tell. However, a new, much uglier one has arisen. My system appears to have become unstable with constant lock ups in Vista with a BSOD here and there. When I attempt to use the AMD OverDrive tool to auto OC it freezes, same thing when I try to stress test with that app. Also, I attempted to run 64bit prime 95 but after just a few minutes all of my cores had failed. I reverted to stock speeds to eliminate the vista 64bit bsod errors. Ht ref clock is at 200mhz, nb speed is 2000, hyper transport link is going at 2000mhz, 1.36V to cpu core, 2.1V and manual clocking for ddr2 1066 ram, no other extra voltages. Any input?

How did you resolve the first issue? I would swap the RAM out with new ones and see where that takes you. Don't use the auto windows based OC tool either. Do all of your OC'ing via the BIOS but I would leave everything at stock settings until you are fully sure that your hardware is good and there are no software issues.

PSU would be a possible issue... unless you happen to have a issue with the motherboard...

Well I did two things, first I disabled Spread Spectrum (Foxconn bios) which is said to "significantly reduce electromagnetic interference generated by the system," if overclocking is used it says to disable it. Next, I took off the auto CPU Multiplier adjust option and set it to the stock x13 2600MHz. I do have CPU voltages up to 1.36V up from that auto setting which had it at around 1.28, I have it at 1.36V because I intended to OC it to 3GHz. Earlier, I manually adjusted the RAM to 533 MHz (x2=1066) since they auto ran at 800Mhz, for that I also manually adjusted ram voltage to 2.1V which is stock. Now I did memtest on each stick and then paired them up and ran it. I gave each of the 6 test cycles 4 passes and there were no errors. As a result, I don't THINK the RAM could be at fault. However, I would like to note, the CPU-NB HT Link Speed is still set to auto in bios as are the NCHT Incoming and outgoing Link Width options. Furthermore, the Advanced Clock Calibration is set to disabled.

I have a PSU tester and I tested it for a number of hours and the voltages were always in range before I hooked everything up. They appear to be in range in the BIOS as well.

When the BIOS was at factory settings before OC'ing did you have any issues? I would still try some different memory if you have access to some just to rule it out. As evilsizer said it could be psu related but I would rule out the RAM first. Reset the CMOS and see what happens, then try different RAM. If you think the psu is ok then maybe you have a bad motherboard. A fellow member in another thread had all sorts of issues with his rig and even tested his psu, but in the end, it came down to being a bad psu so you want to swap out the easy stuff before you go spinning your wheels so to speak.

Before overclocking I had the aforementioned constant glitching and sometimes BSOD while playing ANY game. Technically, I do not have anything overclocked at all right now, all I have is manual settings for a lot of the components since the auto (factory settings) didn't pick up on some components being faster than they were (RAM modules). On another note, I will be trying some of the tips given but I still doubt it being the PSU since I took it from another build that was fine. I could be wrong though. I am currently testing 1 core at a time on my processor and then will play around with it some more. Would manually clocking the rest of the mobo specs and components help do ya think? I mean like the NB, HT and all that good stuff. Auto just seems to be a bad setting, maybe it is the mobo though I am no stranger to messed up mobos.

maybe its your temps? what are your idle and load temps?

GPU idles at 40 and doesn't ever get above 50.
General system idles and loads in the 30s
CPU idles at like 20 and loads in the high 20s-low 30s (watercooling).

could be mobo but i dont know how the IMC reacts to 4 dimms being populated vs that of a intel NB. have you tried more V to the IMC, not sure if that is possible.

You're gonna have to help me there because I don't know what IMC stand for.

unknown0153 said:

You're gonna have to help me there because I don't know what IMC stand for.

Click to expand...

IMC = intergrated memory controller
the new AMD cpu's moved the memory controller to the cpu. on intel Core 2 its still on the northbridge, on intel you would need more NB votlage. as it get stressed more with 4 dimm populated. since i dont own any AMD rigs with a IMC, i dont know how it reacts with 4 dimms being populated on the board.

either that or im mis-reading how much memory you have, 4gigs x2 sticks or 2gigs x4 sticks.

RAM may not be the issue, but testing with 2 sticks may rule it out. If the 2 sticks work but 4 fails then you may have to change the drive strength which controls the signals going to the DRAM. On some BIOS's it's a simple type 1 or type 2 setting, others it's level based from weak to strong.

Ruling out RAM I would next move to a NB issue. Increasing the NB voltage slightly may help but one thing I've seen with these K10s is they are sensitive to voltage. Just a little too much will bring it down. I would drop the CPU core back to the 1.28-1.30 range until you are ready to overclock.