Reverse approach: computer last depending of temperature

[krissM]

I've built my new computer with fanless components, also cooling capacity is constant. As the result I want not run prime95 in more than 1 thread under this hot summer.

Idea: Last should be depend of CPU temperature.

One could run more instances of prime95 if someone could write program which check CPU's temperature and controls (dynamically) the number of (prime95) threads.

[krissM]

I've written self a “dirty” manager. CPU temperature one can get using “Core Temp“ http://www.alcpu.com/CoreTemp/developers.html SDK. I've used example code from “.NET Framework 2.0 Dynamic Link Library for the shared memory interface”.

Then, manager starts and kills dummy programs and prime.txt has following row:

PauseWhileRunning=pauseP95_1.exe[1],pauseP95_2.exe[2],pauseP95_All.exe

Core Temp program must use (in my implementation) degrees in Celsius and “Display the distance to TjMax in temperature fields” - Advanced tab in Options | Settings.

I'm testing my manager and is seems to work correctly. I don't use more then 3 workers – so utility is not adapted for more workers.

I use Win7 and Intel Core i7 4770S CPU in fanless computer.

[Ken_g6]

On my laptop, I find it's more efficient to run one thread for each core, but throttled way back with ThrottleStop, than to run a single thread faster. But that may not be what you want if you run other software on that machine. And I haven't found a way to throttle based on temperature, unfortunately - you just have to fiddle with the multiplier manually.

[krissM]

I've tested my temp manager with 4 workers. It is too much load for my passive cooled computer, so first one and then second worker has stopped.

Strange (for me) is that load is constant 50% for two last cores – I have Core i7 4770S CPU.

For start condition with 2 or 3 workers load is dynamically divided between cores.

Can someone explain me such behavior?

[krissM]

Originally Posted by krissM

I've tested my temp manager with 4 workers. It is too much load for my passive cooled computer, so first one and then second worker has stopped.

Strange (for me) is that load is constant 50% for two last cores – I have Core i7 4770S CPU.

For start condition with 2 or 3 workers load is dynamically divided between cores.

Can someone explain me such behavior?

(can not edit, write new version of my last post)