Mike Gerdts wrote: > $ uname -srvi > SunOS 5.11 snv_111b SUNW,SPARC-Enterprise-T5120 > > $ ptime pkg search -l ls > [snip] > real 34.130866085 > user 30.690454076 > sys 0.774897050
SPARC performance may be suffering mainly due to this bug: 10130 fmri object creation and string conversion performance issues on SPARC > the new software takes 12x longer for equivalent tasks. There is a > very high startup cost with pkg that does not exist with pkgadd. With > pkgadd I don't think too far ahead to be sure to group as many > operations into one invocation as possible. When I use pkg, I most > certainly try to lump as many operations as possible into each > invocation to avoid this startup penalty. Currently, startup performance primarily suffers from the following (from what I've discovered so far): * package data loaded on startup is not limited to what is needed (I'm working on this currently) * FMRI overhead issues (SPARC only as far as I know -- see above) > When I look at the publicly disclosed/speculated road map for CMT > systems, I don't see things improving for the simple operations > without fixing the software. I eagerly await the SAT solver and any > other improvements that are in the works. The main improvements coming are to: * startup times * info and uninstall performance * install / image-update (via SAT solver and catalog work) * file transfer (most of which is in place, but more is coming) > Right now I'm not complaining - I know the software is young and the > primary development platform is x86 where the regression isn't so > apparent. Once I start hearing that there aren't big performance > improvements coming, I will start opening support calls if the > performance is still worse than before. The pkg(5) project is currently focused primarily on achieving correctness and desired functionality, it just so happens that a lot of that ongoing work also gives us better performance. Significant performance work has also been done already, and more is planned. Cheers, -- Shawn Walker
