You’re generally using one processor-intensive application at a time, and so the first core (or it could be the second core, whatever the windows scheduler decides) will be doing the heavy lifting and will be disproportionately hot. If you run a program that’s designed to spread the work around with multithreading (say, a good video encoder) you’d see more uniform temperatures.
That’s definitely possible, and perhaps likely. However, the temperature sensors that give those readings are uncalibrated and notoriously unreliable, so it could just be that Sensor 1 just reads a few degrees higher than Sensor 2. The readings they give are more useful on a relative sort of scale. (Things seem pretty stable at temperature T. Now it’s summer, and the computer is getting flaky, and I haven’t dusted it out since forever, and holy crap the processor is at T + 50, maybe I should do something about that…)
The sensors aren’t “notoriously unreliable” - they’re on-die and very accurate. Some confusion came from Intel hiding what the values given to motherboard sensors actually meant. Rather than reading a temperature, it actually gives a distance to the CPU throttling temp. IE a reading of “40” on a CPU where the throttling limit is 115c means the CPU is at 75c. Some confusion came because Intel wouldn’t release the throttle temp for their CPUs and each core or stepping could be different - so different programs would interpret the number differently, which could make it seem inconsistent.
The sensors are very good at what they do - accurately sensing when the CPU is getting dangerously hot and handling the situation. But Intel has not been user friendly about it for some reason.
Edit: There’s a program called RealTemp that is regarded as having the best guess as to what the number means on different processors.
My PC tower is situated next to an air conditioning vent, which is great for helping keep it cool but bad for dust.
When I’m running a non-intensive program (say, Firefox) and the A/C is running, core temps are usually about 29 degrees Celsius. After a couple hours hardcore gaming, I’ve seen temps as high as 62 degrees celsius.
I guess I’m wondering if that’s too hot. I did a little googling around and got the general impression that it was OK but may shorten CPU life.
Don’t modern PC’s have some kind of thermal overload protection so they will automatically shut down if temps get too high?
Most of them do, and most of the time it works pretty well.
I love this place 
Thanks for all the answers. Please continue with the side discussions, I’m getting answers I never knew I had questions to!
That large a difference between idle and load temp does indicate bad cooling (my core temps go from like 48 idle to 54 load, IIRC). But CPUs can safely handle temperatures that seem pretty high. Most of the auto throttle points on CPUs are above 100c (as read from the core). 62c, if you’re talking about a core temp, isn’t something I’d worry about.
This is something I’ve wondered about.
Short of peeking in Task Manager …How can we tell which applications make use of the multiple cores?
In particular, let’s say I bought a late model dual-core Intel machine with Windows Disgusta…err… Vista Home Premium, and the only thing I install on it is Microsoft Office 2007. If I’m running only one of the applications from Office, does it take advantage of the multiple cores?
How about my old copy of Office 2003 or even 2000? Are they any worse?
How about various versions (6,7,8) of IE that may have come with the machine or been upgraded on it?
Granted, most of my hunt-n-peck typing barely uses the capacity of any processor, but web browsing might need a bit of cycles if any graphics are involved (or maybe that’s a function of the graphics card driver?)
STASMQ
(Sorry To Ask So Many Questions!)