Re: performance....

On 04/14/2002 at 08:13:29 PM EST, Corey Minyard <[email protected]> wrote:

> It's a little disturbing that it's being advertised that NGPT is faster
> than Linuxthreads for virtually everything, and common operations like
> mutexes and conds are being measure with these values.  So, I'm from
> Missouri, show me.  (Actually, I'm not from Missouri, but I know people
> who are :-)  What are the test platforms, where is the benchmark code,
> and what are the actual measurements?
>

Most of our benchmarking and testing has involved the same test cases that
are included in the LinuxThreads package.  In these tests, NGPT is faster
in virtually every area.  We've still got some hot spots (such as those
Richard pointed out) that are being worked on.  I'm a little concerned
about pthread_self since we use almost the same technique as LinuxThreads
and I'll have a look.  I could use more info and if you'd like, please send
me the test cases you've run and the platforms (type, number of processor,
memory, etc.) you've run these tests on.  This way we can run those tests
and see if maybe you've found some things we've overlooked.  We're going to
running an Apache benchmark very soon and will post the results (good or
bad) to this list and on the website.  We're waiting to get a working JVM
and at that time, we'll be posting volanomark results as well.  We're
hoping to have the JVM shortly.  So hopefully by then you and the other
folks from Missouri will be satisified.  Thanks for the input.  We really
appreciate it.

Regards,
     Bill Abt
     Senior Software Engineer
     Next Generation POSIX Threading for Linux
     IBM Cambridge, MA, USA 02142
     Ext: +(00)1 617-693-1591
     T/L: 693-1591 (M/W/F)
     T/L: 253-9938 (T/Th/Eves.)
     Cell: +(00)1 617-803-7514
     [email protected] or [email protected]
     http://oss.software.ibm.com/developerworks/opensource/pthreads