Dual or quad core?

Panda Features in Development
1

A long time ago David said that he plans/works on such multi-threading changes that would allow users to program for Panda in the usual way (without caring about threads), while the Panda’s engine would use all available CPU cores. Can you, please, give a quick update on that?
Today it so much upsets to see that Panda uses only half/quarter of processing power on our dual- and quad-core machines.

2

Still a back-burner project. Since it’s not a priority for Disney right now–we’re marketing our games to people who aren’t likely to have multi-core machines–it’s not something I can spend a lot of time working on. There are, however, other branches of Disney who are eager to see some progress made in this area, so I may be able to wrangle some time to work on it sooner rather than later.

David

3

Cool, quad cores support will be awesome!

4

hmm what about octi core?

5

Hi David,

Do you have any news on this feature? :wink:

6

Actively working on it now. It won’t be in 1.6, but it might be in 1.7.

David

7

Hi David,
I want to ask for little update on the subject :slight_smile:
Can we expect multi-core optimizations in 1.7?

8

not sure if it still works, but in the manual there is the prc option loader-num-threads. you could use that for a little optimization.

can’t test it here as i have only one core…

9

Sorry to say, the full multi-stage pipeline implementation has been put off a bit longer, to make room for the browser plugin, which is much more important to my employer at the moment.

As Nemesis points out, there does exist already a suite of tools for using multithreaded operations, especially to smooth load times and such. Using these multithreaded operations doesn’t require a multi-core CPU, though, and won’t particularly benefit from one.

Maybe I’ll be able to get the multi-core support done in time 1.8. It is pretty close.

David

10

With Python 2.6 it will be possible to use the new multiprocessing module, which offers a thread-like API for inter-process-communication. This would allow to put Python-based code on other cores, e. g. AI which is often written in Python

enn0x