games keep crashing, cpu bottleneck??

[pitman1]

hey guys i kept having problems with ghosts that keep crashing randomly (no surprise) so i thought meh id go and buy BO2 but that also keeps crashing! and a few times BO1? at times it would spaz out and everything, like firefox would crash after? BF3 and csgo have crashed a few times but very rarely kinda norm i guess :/.. no where near as bad as this. i have a feeling it might be my cpu?

amd phenom 2 975BE quad 3.7

asus M5A97

corsair 8gb 1833

gtx 770

antec 650w platinum

Tosh 500 gb HD

all but the cpu is about 4 months old.

 

[bigpinkdragon286]

A CPU bottleneck in a game simply means that your CPU can not feed your GPU as fast as the GPU can handle the workload.

A CPU bottleneck will not cause crashing, ever.

Something in your computer is malfunctioning, and sorting that out should take care of the problem, without the need to upgrade the CPU. It could be as simple as a bad setting somewhere.

 

[Onus]

Run Prime95 and see if your RAM is stable. If not, giving it another .05V may help.

 

[bigpinkdragon286]

Actually, rather than running a test which stresses multiple components at once, I would start with strictly a memory test:

http://www.memtest86.com/

http://www.memtest.org/

http://windows.microsoft.com/en-us/windows7/diagnosing-memory-problems-on-your-computer

While adding .05V to the RAM might not hurt, I don't generally recommend arbitrarily adding voltage to components we haven't determined to be malfunctioning, except as a last resort.

 

[bigpinkdragon286]

If your RAM comes back without errors, I would then run a utility such as AMD Overdrive, which can execute a processor stability test:

http://sites.amd.com/us/game/downloads/amd-overdrive/Pages/overview.aspx

Ultimately, you want to try and test each component with as much isolation as possible, so you aren't left with results that point in multiple directions.

 

[bigpinkdragon286]

Something else you may consider doing is downloading a utility to monitor your temperatures:

http://www.cpuid.com/softwares/hwmonitor.html

If your CPU is overheating, it could have symptoms such as you've described.

 

[pitman1]

just tried prime 95 and within second got a blue screen dumping memory, ill try isolating test. thanks for the fast reply guys!

 

[bigpinkdragon286]

You're welcome. Next time you blue screen, you might post the stop code for us, and see if there is an error message with a particular file. That can sometimes give us a direction to start hunting in.

 

[bigpinkdragon286]

When running Prime95 would be an ideal time to have your temperature monitor going. That sort of stress test can really heat up the CPU. Ideally you want to pay attention to both minimum and maximum. A minimum temperature that is high often points to the cooling of the CPU being insufficient for the given ambient temps. Sometimes it just takes a quick cleaning of the CPU cooler, but your original post seems to point to things not having been assembled and in service long enough for that sort of heat sink accumulation.

 

[pitman1]

ran prime 95 got a worker #4 fatal error hardware failure, ran again got blue screen stop: 0x00000101 (0x0000000000000031, 0x0000000000000000, 0xFFFFF988031D5180, 0x0000000000000003

 

[pitman1]

cpu running at about 52 after 10 mins of prime

 

[bigpinkdragon286]

52°C after 10 minutes of Prime95 sounds very reasonable, in my opinion. That is likely not the source of your troubles.

You might look into the configuration of your RAM modules. A bad setting can easily cause your symptoms.

CPU-Z is a great utility for reading the SPD of each installed RAM module:

http://www.cpuid.com/softwares/cpu-z.html

Read the SPD information on the SPD tab page in CPU-Z and then compare it with the actual speeds you are running on the Memory tab page.

 

[bigpinkdragon286]

From the looks of your configuration, you're running your 975BE with an overclock, and your Stop 0x101 points to a misbehaving processor core. You might consider backing off of the overclock and see if you can run Prime95 in a stable fashion, or for that matter, even an underclock might be worth trying, for the purposes of troubleshooting.

 

[a1blaster]

your mobo is 3 years old....gtx 770 is just a couple months old..i think is incompatibility issue...update bios/drivers/windows...

 

[jamesplaysgames]

what happens when ghosts crashes? Does your screen change to a random solid color and then freeze and make a buzzing noise? If so, I have been having the same problem since I started playing ghosts on my PC.

 

[pitman1]

whats the difference between JEDEC? jedec#4 is correct but 2 - 3 and xmp 1866 are different?

 

[pitman1]

motherboard has been all updated plus drivers from video card. and james kinda in BO2 i get some purple witch i know is like data failure and epic tearing like its glitching like mad and then i get noise.

 

[bigpinkdragon286]

JEDEC is the Joint Electron Device Engineering Council. They set standards for RAM, among other things. XMP is the Extreme Memory Profile SPD extension developed by Intel.

Only one column should match what your Memory tab displays, but one column should be matching, provided you are running the RAM at one of it's rated configurations. The various columns simply show what timings the manufacturer of your RAM modules determined that they should be set to, when running at the frequencies listed at the top of the columns.

It is generally advised against running RAM of different timings, unless all modules can be run at the same timings without error. You can not instruct your memory controller to use unique timings for each module, so all will be running at the same timings, whether they are okay with that or not.

When looking at RAM timings, a lower number is actually going to be a faster timing, in all regards except for the clock frequency. Running RAM modules with timings below what they are rated for will often result in bad data returning from the RAM, not because the module is defective, but because it wasn't given time enough to return the correct values.

 

[pitman1]

a lil brake though underclocked my cpu to 3.4 prime 95 after 10 mins no errors you might of been onto something here bigpinkdragon ill try some gaming tomorrow for a few hours and ill post on here the result! thanks a lot!

 

[bigpinkdragon286]

Sure, sounds good, and I'm glad you are showing positive results. Your Stop 0x101 is what makes me suspect your processor clock speed as the culprit. My quick reading on the error leads me to believe one of your cores has a clock signal that Windows is determining to be misbehaving.

 

[pitman1]

after a day of playing a few games i put y cpu back to default and the same problem happend again. im not sure whats going on but underclocking it sorts the problem. perhaps soon ill get a new cpu. thanks for the the help guys defo dragon!

 

[bigpinkdragon286]

Well, you could have overclocked too much, to the point your processor has degraded and can no longer run at it's stock clocks anymore. Another thought, actually, if you haven't touched the voltage of the processor, since your temps seem reasonable, you may try bumping up the core voltage slightly and see what happens. The 975 was near the end of the line when it came to Phenom II chips being made by AMD. It's possible you got one that doesn't like the high frequency. Bumping voltage, when not unreasonable, can help the transistors switch states a wee smidge faster, but it's really only useful when you're running a frequency that needs the voltage, otherwise you're just burning unnecessary power and turning it into waste heat. I would experiment in very small increments and see if you can bring your processor into a stable state. Watch your temperatures though, of course, to make sure you don't hit a point where the processor just starts dumping heat.