On Dienstag, Dezember 11, 2001, at 06:21  Uhr, John M McIntosh wrote:

> Also 161 million calls to ioMSecs or 272,923 calls per second! but I 
> need to investigate that more...

That's the primitive code figuring out wether to check for interrupts or 
something.

This is the reason for the low-res millisecond clock, which should be a 
lower-overhead mechanism, but in the standard VM gets mapped to the 
normal millisecond clock, which is quite slow in OS X.

> Mind what I'm seeing was 15% of the time in the mach-o was taken to 
> calculate macho-carbon milliseconds. Using gettimeofday is much more 
> efficient. 5%+  was taken in pthreads, that's fixed too.

In the CocoaSqueak VM, I am using an NSTimer running in the main-thread 
to update the low-res timer at intervals, similar to what the ITIMER 
code does in the Unix VM (except I don't do it with signals and I make 
sure to get the correct time each time the low-res timer is updated, 
instead of just incrementing by a fixed amount).

Anyway, your post motivated me to do some more checking (using the 
'sample' command line tool).  The result is another 10% off the Macro 
benchmarks.

Marcel

--
Marcel Weiher				Metaobject Software Technologies
marcel at metaobject.com		www.metaobject.com
Metaprogramming for the Graphic Arts.   HOM, IDEAs, MetaAd etc.