Re: Rendering and physics thread repartition(scheduling?)
- From: "Marcus Gambit" <yacine.chahrour@xxxxxxxxx>
- Date: 7 Sep 2006 05:12:17 -0700
Hello David,
No it wasn't.
All the measures I did on that sample confirm the theory (even the 6
single frames).
It was about your physics thread being coded to run even
when it had no work to do. You should simply have coded the physics
thread to run only when it was appropriate for it to run.
As I said in the first post, I did delay physics routine to do the work
only when needed (every 20ms). It made things better but it did not
solve the problem.
Alternatively, you could have reduced the priority of the physics
thread.
As I also said in the first post, I did that. It did not help (the
result was even worse).
The problem is simply that you gave your physics thread an infinite
amount of work to do. Then you expected the scheduler to figure out how
to deal with that.
So how to tell the scheduler to run physics every 20ms?
.
- Follow-Ups:
- Re: Rendering and physics thread repartition(scheduling?)
- From: David Schwartz
- Re: Rendering and physics thread repartition(scheduling?)
- References:
- Re: Rendering and physics thread repartition(scheduling?)
- From: Martin James
- Re: Rendering and physics thread repartition(scheduling?)
- From: Marcus Gambit
- Re: Rendering and physics thread repartition(scheduling?)
- From: David Schwartz
- Re: Rendering and physics thread repartition(scheduling?)
- Prev by Date: Re: Rendering and physics thread repartition (scheduling?)
- Next by Date: Re: Rendering and physics thread repartition (scheduling?)
- Previous by thread: Re: Rendering and physics thread repartition(scheduling?)
- Next by thread: Re: Rendering and physics thread repartition(scheduling?)
- Index(es):
Relevant Pages
|