Boost logo

Boost :

From: Peter Dimov (pdimov_at_[hidden])
Date: 2008-02-08 17:42:08


> Hey, would you mind running against lock-in-order? I'd love to
> confirm what I've been preachin'. ;-) Or if you see lock-in-order as
> better, then better for me to hear it sooner rather than later...

No difference on single core, ~20s (vs ~14s for try/backoff) on dual core
Pentium-D with lock2 changed to:

void lock2_( mutex & m1, mutex & m2 )
{
    if( &m1 < &m2 )
    {
        m1.lock();
        m2.lock();
    }
    else
    {
        m2.lock();
        m1.lock();
    }
}


Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk