On Wed 18 Apr 2007 02:03, Greg Ungerer pondered: > Sven Johnsson wrote: > > I have executed a Dhrystone 2.1 benchmark on my uClinux system and get > > an extremely low result, about 0.0259 DMIPS (I also used a stopwatch > > to confirm it). The datasheet of the processor says 211 DMIPS! > > It looks ok in the 2.6.20 code base, but you need to check it > in your older code base.
Greg: What results did you get on 2.6.20? We posted some various tests (both with flat and fdpic - if you are interested in the overhead of fdpic) http://docs.blackfin.uclinux.org/doku.php?id=dhrystone#results_on_a_blackfin Where we measured around 328.8 with gcc under Linux - according to the folks who don't work on Linux, the results are sub-optimal - the architecture is capable of over ~450 with no OS and the proprietary compiler - and I was just wondering how close gcc/Linux on your platforms is to the non-OS/proprietary compilers. -Robin And yes - for those that are interested, we are going through the output of all the compilers to see how to make gcc better. _______________________________________________ uClinux-dev mailing list [email protected] http://mailman.uclinux.org/mailman/listinfo/uclinux-dev This message was resent by [email protected] To unsubscribe see: http://mailman.uclinux.org/mailman/options/uclinux-dev
