Is this normal when running LinX?

[Highlow]

Hi, i'm new to this forum, i've tried asking this question in other places and have had no answers yet so I thought i might try here.

I'm encountering a very annoying problem, and was wandering if anybody has any suggestions. Please bear with me this is rather a long post. I think this might be a software issue but would really appreciate any input -

I recently installed FSX with the acceleration pack (which i think patches the software to use multi cores and dx10?), since then on the main menu and sometimes during the simulation the mouse will freeze intermitently for 1 to 2 secs over a 30 sec period then be ok for 30 then do the same again, fps goes to less than 5. This also occurs when opening a large speadsheet and occasionally the desktop resets - icons refresh v slowly and already open windows go blank momentarily. So i uninstalled all the FSX software thinking something has screwed up.

Now i'm noticing the same issues, sometimes on startup, or with a couple of apps open, more frequently with multiple web pages open within one browser (using firefox). The only way i can reproduce this effect in a controlled manner is to run LinX or Intel Burn test, freezing only occurs when all 4 cores have just reached 100%, then maybe once or twice while under sustained load - irrespective of the amount of memory being tested. Prime95 small FFTs or blends using upto 3GB ram don't cause any problems like this, even when the cpu load maxed at 4 cores 100%. I'm 90% sure this wasn't happening before the FSX install, but it may have been to a much lesser extent. I get slightly worse 3Dmark06 scores than before without any crash issues, but at times (i guess when all cores are loaded in a particluar way) everything grinds to a halt then speeds up again.

These are the things i have done so far:

1. Reduced a mild overclcock, back to stock settings, both are 24/7 P95 and LinX stable, max load temperatures 53-58C (coretemp) (23C ambient) - this made no difference (btw Vcc never above 1.23 under load).

2. Verified ram settings, all manually set to manufacture's spec.

3. Run memtest overnight at both stock and overclocked settings - several passes no errors.

4. Cleaned the registry with a registry cleaner, disabled AVG and the wireless card.

5. Turned off all the power saving features in BIOS, C1E and C3 etc, (turbo is disabled anyway), as well as turning off the LAN and Audio ports.

6. Resintalled the latest Directx SDK and C++ redistributable, re updated all graphics drivers (completely uninstalled old ones first), chipset drivers, flashed BIOS to latest - F4k.

7. Opened case, removed cleaned and reinstalled hsf, inspected for burned out components, reseated the ram and GPU.

None of these made any difference. Now i'm about to do a fresh OS install, because i thought FSX may have installed/modified some background process which is hogging resources - but i'm not convinced this will work.

Also checked the PSU voltages using HWmonitor, but i think there is some compatability issues because BIOS says +12V is 11.850 (never changes), HWM says +4.3 (which fluctuates +-0.4 under load). Tried using gigabytes easytune 6 to monitor temps but can't get it to install, get a dll error which has BIOS in the text string. Maybe thats a Win7 issue - or could that be a motherboard problem?

So what do you guys think? I'm at a complete loss at the moment. Does this sound like a software issue or a hardware related. Apologies for the long post. Thanks in advance.

Specs
Core i5 @ 2.66 or 3.44 o/c (bclk 172, 1.25)
Gigabyte P55m-UD2
2x 2GB Gskill DDR3 16000
HD 5870
Corsair 750W PSU
Win7 Home Premium
AVG free

 

[n7]

Welcome to the forums

This might be no help, but make sure you have updated to the very latest BIOS for your mobo.

There have been some weird issues they've fixed for my mobo series; might be some for yours too.

Aside from running stock, i'd be reinstalling the OS, like you mentioned.

 

[AdamK47]

Have you checked the event viewer to see if anything was happening behind the scenes during your FSX problems?

 

[aigomorla]

Have you checked the event viewer to see if anything was happening behind the scenes during your FSX problems?

+1 this..

go in the event viewer and see why its freezing.

 

[Highlow]

Welcome to the forums This might be no help, but make sure you have updated to the very latest BIOS for your mobo. There have been some weird issues they've fixed for my mobo series; might be some for yours too. Aside from running stock, i'd be reinstalling the OS, like you mentioned.

Thanks, yes i have the latest BIOS which i believe has more overcurrent protection and a few more o/c features. I will reinstall everything tonight see if that helps. Its really the last resort at the moment, i also need to figure out how to monitor PSU voltages because thats my next port of call. Could an unstable PSU cause random system freezes like this?

Have you checked the event viewer to see if anything was happening behind the scenes during your FSX problems?

Actually no, I'll do that before i reinstall everything, thanks. I guess whatever it is would be running all the time - i'm also puzzeled as to why p95 100% doesn't have effects, but LinX does.

Thanks for the replies, i had one further thought, could this be related to a wireless mouse/keyboard setup? Driver issues or something.

 

[aigomorla]

IBT has a different route in which it loads.

That's why the temps on IBT can be higher then prime95.

Your system should not freeze when running 1 alone @ 100%.
The fact its freezing when all your cores is loaded leads me to think you might have a voltage issue.

The only way we can find out is your event view and see what exactly bsod'd.

 

[Highlow]

Your system should not freeze when running 1 alone @ 100%. The fact its freezing when all your cores is loaded leads me to think you might have a voltage issue.

Ok, do you mean PSU Voltage or not enough Vcore, Vtt etc? I'll check the event view and reply here. Thanks

 

[aigomorla]

Ok, do you mean PSU Voltage or not enough Vcore, Vtt etc? I'll check the event view and reply here. Thanks

could be a bunch of things.

Dram Vcore
Vcore
VTT
PLL




If windows itself bsod'd then its probably a vcore issue.
Bump vcore up and see if you can repeat the problems.

If windows is reporting a .dll error, or msx error, then its probably software.
If windows isnt reporting any error.. well, i dunno then...

 

[Highlow]

Right, thanks. I'll see what the events are and go from there, probably do a clean install to eliminate that as well. Reading around it seems that there are quite a few people experiencing random freezing in Win 7, although i haved eliminated the cause of most of them already...:\

 

[Highlow]

Event log showing a series of CAPI2 errors during the slowdown/freezing. After enabling and logging a few they are associated with processes in order of increasing frequency:

wmp..(something).exe
explorer.exe
consent.exe

The wmp...(something) is media player and they all are of the "windows update signature is invalid" flavor. I disabled media player updates, which only served to increase the number of explorer instances. Can't find consent.exe or what it is associated with anywhere...

Also discovered the stuttering is not restricted to high CPU loading, i noticed major slowdowns happening when just browsing (perf monitor shows CPU ram not taxed at all).

I read around a bit seems this is a Vista problem, not seen anything relating to Win 7.

Is it worth running LinX in safe mode to eliminate drivers? Or just fresh install? (spent all evening messing around with this stuff so didn't get time for that)

Thanks again guys.

 

[myocardia]

Your problem sounds to me like what happens when trying to overclock, but not using enough vCore/vTT/vDRAM. When you reinstall FSX, do not reinstall Acceleration. It's total crap. FSX SP2 (which Acceleration contains and requires) seems to cause problems with ~50% of all systems its installed on. Here's the download link for FSX SP1 which is all that's required for multicore action.

 

[SonnyDaze]

Your problem sounds to me like what happens when trying to overclock, but not using enough vCore/vTT/vDRAM. When you reinstall FSX, do not reinstall Acceleration. It's total crap. FSX SP2 (which Acceleration contains and requires) seems to cause problems with ~50% of all systems its installed on. Here's the download link for FSX SP1 which is all that's required for multicore action.

Good to know. I have both SP1 and SP2 installed and it runs with no probs. I had read that SP2 had some .cfg fixes for minor performance increase.

OP do you have the patches installed?

 

[theevilsharpie]

Does this problem occur in safe mode?

 

[Highlow]

Thanks for your replies.

Your problem sounds to me like what happens when trying to overclock, but not using enough vCore/vTT/vDRAM. When you reinstall FSX, do not reinstall Acceleration. It's total crap. FSX SP2 (which Acceleration contains and requires) seems to cause problems with ~50% of all systems its installed on. Here's the download link for FSX SP1 which is all that's required for multicore action.

Well maybe I have a dud CPU/motherboard then because it behaved like that at stock Vcore, Vtt, and Vdram and even with everything on AUTO in BIOS. But i think i found a solution, see below.

OP do you have the patches installed?

Yes i think so, aren't they included in the acceleration add-on? I understood it also had DX10 support? Noticed a DX10 preview option in graphics settings, didn't really get to check it out.

However i think i have discovered what was causing my problems, (and your not going to believe this). After resintalling windows and testing with LinX before installing any other programs i discovered i had the same problem (although to a lesser extent). So i went out on the net from the same computer and started browsing around for any more help, to my surprise the system was grinding to a halt while just surfing. Disabled the wireless as before, no change. Then uninstalled the driver which Win7 helpfully installed on the first boot into OS, opened up a couple of saved web pages, slightly better. Then shut down, removed the WLan PCI card and removed all remaining bits of software relating to it. Much better now, connected via ethernet, browsing was cured! In addition my coretemp readouts were a couple of degrees lower.

However I still had the LinX induced stuttering (also occurs in safe mode). This was worst when the CPU was coming on and off load, once at 100% for a few secs no problem, as soon as the load drops off inbetween calcs - freezer central. Once, the system completely locked up after the test had ended, began to be sluggish as the load decreased, eventually grinding to a halt about a minute later. So - and this is the part your not going to believe - after playing with voltages for a while i decided to mess with the CPU fan control. Disabling auto so it runs flat out all the time, this made the issue disappear completely! I was very surprised and sceptical. So set LinX to use maximum ram 20000calc/3003MB/40flops, during that I could now do a multitude of tasks in the background, internet, openning closing office docs, changing startup items, search event viewer etc without any hanging/freezing or stuttering. As before - clocks 2.66 (20x133) / 1333 and BIOS set voltages were Vcc 1.25, Vtt 1.10, Vdram 1.5, which i believe is stock. Coretemp reported very similar (to the auto setting) high 40s on 100% load. Set fan to AUTO again, problems returned.

Reinstated the 3.4 o/c (BIOS profiles are great!) did the same test, everything ran nice and smooth. Basically spent 6 six hours stressing and playing in the bacground, so i didn't get a chance to install FSX last night. Ran P95 overnight @3.4 with no errors maxed 53C temps (ambient 22-23). I just can't believe the fan control is the culprit. Would you guys conclude the motherboard is to blame?

This time around i got easytune to install and was able to monitor PSU voltages, all were stable and within 0.03 of their datums. An added feature I found here is an "advanced" fan control, basically two levels, i set 25% below 40C and 100% above. Ran a bunch of stress tests without issue.

The event log doesn't have any of the errors I found before but they may return in due course as windows updates itself and i install FSX again!

Sorry for the the long post again

 

[myocardia]

Would you guys conclude the motherboard is to blame?

I would conclude that your motherboard's BIOS has a problem, and that you should send an e-mail to the company that makes your motherboard, with a link to this thread, asking them to "look into this". Believe it or not, this type of thing happens all of the time, more often with some companies, less often with others.

BTW, try reloading your current BIOS with your CPU at stock clocks, before contacting your motherboard manufacturer. There have been many instances over the years of people blaming a BIOS when the actual problem was having a not completely stable overclock when they upgraded their BIOS.

 

[Highlow]

Well i spoke too soon. I have the same problems - now they'll occur without any load, and the same in safe mode, very few basic processes running. Same at stock clocks, Upping Vcore makes no difference, tried flashing BIOS back to F3, and the F4 all the same deal. Ram?

 

[theevilsharpie]

RAM issues usually cause BSODs.

One of your hardware components may be in the early stages of failure, which could cause excessive hardware interrupts. Use your performance monitor (it's in administrative tools) to monitor the Processor/Interrupts\sec performance counter, and try to replicate the freezing problem while the counter is running.

 

[Highlow]

Ok, thanks. What sort of numbers should cause alarm bells when monitoring the interupts?

I think this maybe a heat issue. LinX and IBT run all day long never failing. CPU core temps are totally normal (realtemp/coretemp). However once this problem starts it never goes away, happens at low CPU load, in safe mode. After about 15mins of inactivity things get sluggish. I can accelerate the arrival by running a stress test for a few mins. If I just restart, the issue is still there immediately upon entering windows and stays (left overnight). But is I shutdown and don't restart for 30 seconds the issue has gone only to return a few mins later.

Leads me to believe something on the board is getting hot and not able to cool off. The components surrounding the CPU are very hot to the touch. This is at stock settings. So I think maybe the board has a problem. So I would say the CPU is ok because the stress tests are fine. By the way the chipset hs is not hot at all.

For all the hassel I should have got an 1366 setup.

 

[theevilsharpie]

Ok, thanks. What sort of numbers should cause alarm bells when monitoring the interupts?

It depends on how many devices are in the system, but it normally be about 1,000 at idle. A defective hardware component can cause that to shoot up to 20,000+.

 

[Highlow]

Thanks, not may devices, i've disabled firewire/floppy drive/LAN, only have the essentials, USB and a display adapter. When i checked this morning i'm seeing around 7-9,000 at idle.

More details about last nights testing:
Tried swapping out the GPU with an old HD 4650, and one stick of RAM. Flashed BIOS to F4n @ stock clocks. Pulled everything off and breadboarded to elimated case shorts. Same deal. Reinstalled OS again. No change. Changed keyboard and mouse.

Its the chokes and capacitors that get very hot (burn my finger hot). Chokes actually change color slightly after a couple of mins LinX (thought it was condensation but no) Manually set all voltages to stock Vcc 1.25, Vtt 1.10, 1.8, Vdram 1.5. Core idle temps 27-30C (Amb 21C), system (HWM) 25-28C. The hot chokes seem to coincide with the hanging problems, but i might be wrong because I would think if this was a problem the CPU would not be stable.

Ram is cool to the touch, GPU is cool. Chipset is cool, nothing looks burnt.

If i carry on trouble shooting like this much longer i'll be getting a divorce! Litterally taking up all of my time outside work. I really appreciate the input from everyone. Thanks again.

 

[theevilsharpie]

Interrupts/sec look OK.

Sounds like some mobo components are overheating. Take your side panel off, point a fan (a large one for rooms, not a case fan ) so it's blowing on your mobo, and try your stress test again.

 

[Highlow]

Yes, sorry forgot to mention I tried that, those components are so hot i'd need liquid nitrogen to keep things cool, the problem still occured but was somewhat less severe,

To be honest i think i'll pick up an EVGA board on the way home and try that (i'm hedging my bets and also hoping the CPU hasn't been damaged). If that works i'll RMA the gigabtye and sell the replacement. If its not any better i'm stumped... Probably take the evga back, sell everything else and buy an 1366 bundle, I should've done that in the first place, instead thought i'd save a few bucks (I would have to buy another i5 and another set of ram to eliminate them anyway so why not upgrade the lot) I guess you live and learn...

 

[Highlow]

Yes, sorry forgot to mention I tried that, those components are so hot i'd need liquid nitrogen to keep things cool, the problem still occured but was somewhat less severe, so i think that has something to do with the board, or socket maybe, and this is definitely getting worse - when i tried the older BIOS (F3) again, the freezing was absolutely ridiculous.

To be honest i think i'll pick up an EVGA board on the way home and try that (i'm hedging my bets and also hoping the CPU hasn't been damaged). If that works i'll RMA the gigabtye and sell the replacement. If its not any better i'm stumped... Probably take the evga back, sell everything else and buy an 1366 bundle, I should've done that in the first place, instead thought i'd save a few bucks (I would have to buy another CPU and another set of ram to eliminate them anyway so why not upgrade the lot) I guess you live and learn...get what you pay for etc

 

[Highlow]

This issue is now solved, theevilsharpie you were right, some components were getting too hot - i ended up using two house fans and turning the AC down to 65C to eventually cool things down enough.

So I replaced the board which did the trick, everything is as it should be now, in fact its considerably better than before - all the stuff i plugged in works fine and I was able to better my o/c by 350MHz with no temp or voltage increase... so i may hang on to this EVGA. Btw interrupts/sec are between 5-700 at idle now. Not really sure what was up with the old one. Its getting RMA'd...

What a lot of stress... Thanks for your help guys, really appreicated the suggestions and learnt a lot about software/hardware monitoring, the interrupts/sec compared to system calls/sec test is very useful.