Alexander Gorshenev wrote: > On Thu, 12 Jul 2007, Danek Duvall wrote: > > On Fri, Jul 13, 2007 at 06:50:52AM +0200, Roland Mainz wrote: > > > >> With "-xO3" it takes ~31 seconds to execute this script. > >> With "-xc99=%all -xO4 -xalias_level=basic -xdepend" it takes ~39 seconds > >> to execute this script. > >> > >> Erm... the results are somehow upside-down from what I would expect from > >> these options... ;-( > > > > A typical response to that is that the function inlining that -xO4 turns on > > makes small programs bigger, and less likely to fit in cache or in a set of > > cache lines, resulting in memory thrashing and lower performance. I expect > > that cputrack(1m) would help you figure out if that's happening. > > This is in the userland, right?
Yes... this is about ksh93. Contrary to gdamore's fears I do not plan to convert it into a kernel module... =:-) > Sun Studio Performance Analyzer is your choice. Thanks! :-) .... right now I am building a test matrix to check which option combination trigger the issue... ---- Bye, Roland -- __ . . __ (o.\ \/ /.o) [EMAIL PROTECTED] \__\/\/__/ MPEG specialist, C&&JAVA&&Sun&&Unix programmer /O /==\ O\ TEL +49 641 7950090 (;O/ \/ \O;) _______________________________________________ tools-discuss mailing list [email protected]
