Re: maximum speedup bound on multi-core processors
Re: maximum speedup bound on multi-core processors
- Subject: Re: maximum speedup bound on multi-core processors
- From: Bill Bumgarner <email@hidden>
- Date: Sat, 31 Jan 2009 15:00:37 -0800
On Jan 31, 2009, at 2:48 PM, email@hidden wrote:
Can anyone tell me if the maximum speedup using parallel programming
on
multicore processors is BOUNDED by:
(A) number of processers (as on a single core processor).
(B) number of processors X number cores / processor.
If each processor runs numCore threads SIMULTANEOUSLY the answer
would be
(B).
If each procssor run numCore threads non-concurrently the answer is
(A).
If anyone REALLY knows please help a frustrated cocoa developer out.
Of far more importance is the algorithm employed and, quite
specifically, the dependencies between threads and exactly how the
dependencies are managed.
The question above can only be specifically quantified if your threads
operates in total isolation, do no simultaneous I/O, and there is
absolutely nothing else running on the system.
In the real world, the algorithm itself will have far more to do with
scalability than raw machine capabilities.
b.bum
_______________________________________________
Cocoa-dev mailing list (email@hidden)
Please do not post admin requests or moderator comments to the list.
Contact the moderators at cocoa-dev-admins(at)lists.apple.com
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden