|
Boost : |
From: Sean Kelly (sean_at_[hidden])
Date: 2004-01-26 11:31:14
>> On Behalf Of Sean Kelly
>> Sent: Monday, 26 January 2004 4:38 PM
>> To: Boost mailing list
>> Subject: Re: [boost] Boost.Thread request: event object
>>
>> Beman Dawes wrote:
> <snip>
>> > How is this different, and safer, than the traditional event variable?
>> >
>> > (See www.boost.org/libs/thread/doc/rationale.html#Events)
> <snip>
>> My goal is really mostly improved efficiency for the Windows side. I
>> don't want to have to use a mutex if I don't need one.
>
> I think this is the crux of the matter. People will not bother with a
> library if there is overhead beyond what the native system provides,
> especially for something like concurrency. I can't pay for it as I have
> to keep my performance competitive.
Okay, I did some digging on usenet (I don't have an ACM sub and was unable
to read the paper that was referenced) and it seems that the Windows event
implementation may indeed be a tad broken. This is something I've never
encountered because I've never required complex use of the event objects
before, but that isn't an assertion I'd like to stand by when delivering
library code. There's definately a chance the existing convar
implementation could be optimized under Windows, but I'm coming to accept
that the overhead might be necessary. I only care about efficiency if
proper program behavior is already assumed :)
Sean
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk