have a c2d 7200 running a mathmatics modeling program but only uses 1 core. when running the program twice it will use each core 100%
is there a way of forcing both cores to work on the one program?
And that would choose which *one* core to operate on, surely?
What you're suggesting can't be done unless the program is written to be multi-threaded to deliberately take advantage of multiple processors. Nothing else can do it on its behalf.
I've never really understood this, why can't the OS take advantage of two cores in a "crude" way?
Like just run the program over both cores, meaning that as it calls for instructions to be carried out, the OS feeds them into both cores rather than just one?
Wouldn't that improve speed?
I know it's not possible, but I can't understand why.
Because CORE1 may be processing a conditional jump which may mean the MATH computation being carried out in CORE2 is redundant or that CORE2 is computing a subsequent conditional jump that should have occurred after the CONDITIONAL jump that core1 is processing simultaneously - and may break the algorithm if it wasn't designed for parallel processing.
The OS would then end up spending more computational power dealing with the above exceptions (and others like it) then would be gained by distributing the code between the CORES in the first place. Or the performance gain is negligible compared to the amount of effort to achieve it by code in the OS.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.