Re: Thread-safety issues in QtHandles

[Søren Hauberg]

tir, 15 11 2011 kl. 04:07 -0800, skrev Joaquim Luis:
> For comparison too Matlab takes 0.02 sec. Being the difference so big
> resulting from the fact that there is no JIT in Octave this example is
> probably not representative of real cases overhead introduced by the
> Michael's changes.

I chose this specific example as it doesn't really include any
arithmetic, so I figured this would be the absolute worst-case
situation. Of course this does not show how things behave on average,
but it does provide some insight.

The exact numbers I provided might not be as good as they should; it is
possible that I compiled without optimisations. I'll investigate that
and report back.

Søren