[Ns-developers] [RFC] Worker Infrastructure - Parallelization in Large

[Hagen Paul Pfeifer]

* Mathieu Lacage | 2009-01-06 11:41:02 [+0100]:

Point 1 and 2 are brilliant hints, I will commit these changes soon.

>3) You would be my personal hero if you would modify the waf
>build/regression system to run the regression suite using a small C++
>program based on the WorkManager when it is available on the target
>platform.

An alternative way is something like "if sys.platform == 'win32'". The
interesting question is _how_ to react for systems where fork(2) is not
supported.

Should "WorkManager::Run()" blindly ignores that a parallelized execution is
not supported and execute all registered task consecutively. This is my
favorite solution because the simulation file is independent from the platform
and the only disadvantage is a slower execution. My personal wish is that the
WorkManager is totally opaque to the user.

Any hints, ideas?

>thanks again for your code,

You are welcome!

Hagen