#721 closed Enhancement (wontfix)
Better CPU Time Implementation
Reported by: | rhughes2270 | Owned by: | davea |
---|---|---|---|
Priority: | Major | Milestone: | Undetermined |
Component: | Client - Daemon | Version: | 6.2.4 |
Keywords: | cpu, processor, cpu time | Cc: |
Description
I've noticed that when, for example, I set BOINC to use 33% of my CPU time, the BOINC uses 100% of my processor for 1/3 of a second and none for 2/3 of a second.
Would it be possible to break this down further so that it appears BOINC is running constantly, e.g. uses 1/3 of the processor for every .01 seconds? It would be much better for screen savers on computers that use less processing power.
Thanks
Change History (4)
comment:1 Changed 16 years ago by
comment:2 follow-up: 4 Changed 16 years ago by
I did think about that, and it seems like it would be a problem. But what I keep coming back to is the fact that, at least under UNIX, processor usage can be allocated using nice. This is far beyond my technical capacity, but it just seems like there has to be a way to efficiently grant a process usage of the processor at a given amount without simply splitting seconds.
comment:3 Changed 15 years ago by
Resolution: | → wontfix |
---|---|
Status: | new → closed |
comment:4 Changed 15 years ago by
Replying to rhughes2270:
But what I keep coming back to is the fact that, at least under UNIX, processor usage can be allocated using nice. This is far beyond my technical capacity, but it just seems like there has to be a way to efficiently grant a process usage of the processor at a given amount without simply splitting seconds.
BOINC already uses nice
to give project applications the lowest possible priority.
The only purpose of the 'throttling' feature is reducing CPU heat on laptops. Even at 100%, BOINC doesn't make your computer slow down.
What if telling applications to stop uses more than 0.01 seconds of CPU time? (writing to shared memory, parsing XML, etc) Also, I think apps only check for messages from the core client every second.
There are also some applications using a wrapper that take maybe *several seconds* to suspend (like when you click the Suspend button).