G'day mercuryus. Sorry to resurrect an old thread, but your link no longer seems valid.

Is it possible for you to put the example up again?

Thanks,

Jibb

I cleared up my webserver (to free space).
Multithreading was downloaded too few in the last time - so I deleted it.

But I will put it on the server the next days again...

I think multithreading would be a good idea, but what about if one thread is busyier than the other, could glitches or "desync" issues occur?

@mercuryus: Thanks.

@CoburnDomain: One thread will generally be busier than the other. If it's the main thread, the secondary thread will probably just spend some time waiting for the main thread to be ready, but on a multi-core CPU that's still much better than having it all in one thread. If other threads end up being really busy (busier than the main thread) you'll either have the main thread wait for them (which really removes much of the usefulness of threading), or have the other threads notify the main thread when they're ready.

For example, if for some reason you needed an extremely expensive path-finding algorithm (it's not always about reaching a single position) you'd put it in another thread to avoid a pause whenever a new path is calculated. Each unit that requests a new path will continue what they were doing (or just wait) without pausing the main thread until they are notified by the secondary thread as to when they can move.

Or at least, that's my understanding. You'll still need to be really careful as to how the secondary thread will notify the main thread.

Jibb

download online again...

Many thanks!

Jibb

new download link

Isnt working

Stupid Google Chrome, will test it

Question?

I am correct in saying: The major benefit of multithreads is allowing a complex algorithm to run while also running the engine.

Would intermittent "wait(1)" commands in the algorithm perform a similar function?

Thanks

no, because the algorithm still eats the CPU time of the main thread