Weird CPU usage (E6600)

FH

Distinguished
Sep 23, 2004
58
0
18,630
This is not really a problem, I have my machine for a few months and it has only happened twice, namely one Core shows constant CPU usage. The first time it was close to 100%. The second time only 11%. The machine is running XP SP2 (see full specs in my signature). Task Manager shows the CPU usage on the performance tab, but not which process is causing the CPU usage on the Processes tab. The second time round I noticed jerkiness of the mouse approx. every 1 second and the Windows logon sound would stall at the same time. No, I'm not accidentally running TAT in the background. The condition is only cured by a reboot (actually I powered down the machine completely). Any ideas what might be the reason? By the way machine is not overclocked right now, instead I tried underclocking to 2GHz and reduced voltage to 1.125V, which is 14 hours Orthos stable. Junction temps were only about 25C at the time according to TAT.
 

abominator

Distinguished
Sep 29, 2007
7
0
18,510
I had this same problem today. I have Q6600 and one core was using 100% under windowsXP 32bit, mouse was lagging and none of the processes used this. I switched off computer just like you and everything was normal. It happend only once since i have this computer (1 month) after playing Sega Rally for about 15 min. My pc (i didn't overclock anything): Q6600 2.4ghz, 8800GTS 640mb, 2GB RAM, HDD WD4000AAKS, Motherboard GIGABYTE GA-P35-S3
 

FH

Distinguished
Sep 23, 2004
58
0
18,630
Yes, a fresh Windows install. Not sure software is the cause. Although Windows software can be programmed to use only one Core I actually know of no commercial software or games which do this. Even when an application is single-threaded by default Windows automatically distributes the load evenly between the 2 Cores, so you get 50% usage on both Cores, not 100% (or even just 11% as in the second case) on one and 0% on the other. Also Task Manager should show on the Processes tab which process is causing the constant load. Well, it didn't. That showed 97-100% of CPU unused. I have a funny feeling this could be a hardware issue - or something buried very deep in the heart of the Windows kernel. Why the periodic, regular sound glitches when only one Core is only 11% loaded (second occurrence of the problem); you know the way when the sound chip is not being fed any further information and keeps quickly cycling over the last sound it received like sometimes happens in games. I don't by the way think the Core was continuously 11% loaded - Task Manager averages the load over it's sampling rate, which is only every 2 seconds or so by default. I think that the one Core was probably loaded full 100% for very brief but regular spikes every one second or so during which time also the sound and the mouse movement stalled. I don't know modern CPUs that well, otherwise I'd say it was responding to an interrupt or something like that. Anyone else ever seen anything like this?
 

FH

Distinguished
Sep 23, 2004
58
0
18,630
Hi abominator! So, I'm not alone. It might be interesting to look for commonalities in our setups and anyone else who wishes to chime in. 32-bit XP here also and a Gigabyte motherboard (full details in my original post). Had a game on it from the beginning (Tomb Raider Anniversary) that uses SecuROM protection. Didn't even have a Virus scanner installed when this first happened. All the latest Windows patches and firewall enabled though and according to Nod32 Anti-Virus, which is on it now, the machine is clean.
 

No1sFanboy

Distinguished
Mar 9, 2006
633
0
18,980
I would consider your underclock as a cause. Is the problem fixed by upping the vcore? I've experienced a similar problem when overclocking.

In my case my e6600 is perfectly happy at 3.2 at both 8 and 9 times multipliers with 1.3125 vcore. Whenever I've tried for 3.6 I run into an issue where one core will peg itself at 100% usage and stays there till I reboot. I've tried different things and this one problem has been my sure sign of being unstable at 3.6. From my own experience I'm led to believe that your under clock is simply unstable.
 

FH

Distinguished
Sep 23, 2004
58
0
18,630
No1sFanboy, I guess that could be. For me there is no real reason to underclock or overclock at the moment other than curiosity, so it would be an easy fix. On the other hand look at abominator's experience, who seems to be running everything at stock. Also in my case the problem is very rare and my CPU, which is only Orthos (14 hour) stable up to 3.05GHz at stock voltage, has not regularly had this problem at speeds around or above that (only once in 5 months to be precise). Your input is interesting, but I wonder is there more to it than playing with Vcore and clock. Out of curiosity, what's your motherboard?