PlaneShift

Support => Linux Specific Issues => Topic started by: Davigetto on April 07, 2009, 07:04:41 pm

Title: Take advantage of Dual Core PC's
Post by: Davigetto on April 07, 2009, 07:04:41 pm
Hello,

I am playing a compiled-by-me version of Planeshift on a Intel Dual Core (using the -j3 option everytime I make or jam). When I play, I have noticed that only 1 of the processors is being used (only 1 of them keeps working while playing).

Several games like World of Warcraft take advantage of Dual Core, using the both of them for playing with a considerable higher performance. I wonder if there is any way in planeshift to play using both cores. I am sure that it will increase a lot the performance.

Thanks and greetings.
Title: Re: Take advantage of Dual Core PC's
Post by: Xordan on April 07, 2009, 07:44:56 pm
The next release with take advantage of multiple-cores for stuff like background loading. :)
Title: Re: Take advantage of Dual Core PC's
Post by: Davigetto on April 07, 2009, 07:59:21 pm
The next release with take advantage of multiple-cores for stuff like background loading. :)

Thanks for the fast reply :D
Title: Re: Take advantage of Dual Core PC's
Post by: Vonor on April 09, 2009, 10:34:30 am
also the -j switch for make and jam takes only advantage of multi cores / multi cpus at compile time. not at run time. in other words, the higher the -j is set the more source files will be build parallel. a good number for -j is number of cores/cpus + 1, but you can even do more.

If you only provide -j it will automatically take as much processes as possible.
If the number is too high it will slow down compiling again.

On my gentoo box I figured -j7 is a good choice for everything, but best is you try it out yourself. compile something with "time make -j#" and increase # every time, don't forget to make clean before :)
Title: Re: Take advantage of Dual Core PC's
Post by: enderandrew on April 09, 2009, 08:37:35 pm
I hope compilers are more intelligent in the future in automatically producing binaries that are more multi-process and multi-thread aware.  Right now the onus is certainly on the coder.