[concurrency-interest] Concurrency-interest Digest, Vol 52, Issue 3
tim at peierls.net
Wed May 6 12:03:08 EDT 2009
On Wed, May 6, 2009 at 9:58 AM, Doug Lea <dl at cs.oswego.edu> wrote:
> Cleber Muramoto wrote:
>> How can one estimate the trade-offs of running into possibly
>> unnecessary instantiations in favor of narrower synch blocks?
> If you don't know, measure. And when you measure, do so across as many
> kinds of workloads as you believe will arise. We measure a lot, which in
> turn causes us to usually choose a good approach before measuring. But even
> still, "usually" is far from "always".
> The JCiP (http://jcip.net/) testing chapter(12) contains some introductory
> discussions on performance measurement.
And see http://gee.cs.oswego.edu/cgi-bin/viewcvs.cgi/jsr166/src/test/loops/ for
many examples of measurement code. The techniques used are sketchily
documented at best, but in conjunction with the JCiP advice they might give
you a sense of what kinds of things you need to worry about.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Concurrency-interest