30 December 2002 GForth discussion list I compiled the snapshot gforth-0.5.9-20021227 with Redhat 7.0, using gcc-2.95.3, gcc-2.96-85, and gcc-3.2.1, with no apparent problems. Use of the configuration recommendations of the Known Problems page affected only the benchmark times.
Given the same configuration options, gforth-0.5.9 ran about the same speed with each compiler. gcc-2.96-85 gave slightly better speeds. The option --enable-force-reg affected speed of gforth-0.5.9 quite a lot. The other options did not affect speed much. gforth-0.5.9 ran the benchmarks much faster than the best gforth-0.5.0 benchmarks. The gforth-0.5.0 release compiled with gcc-2.96-85 and gcc-3.2.1 only if 'configure'ed with --enable-indirect-threaded. 'configure' options affected the speed of gforth-0.5.0 a lot. -fno-gcse (for gcc) and --enable-force-reg (for configure)each added some speed to gforth-0.5.0. gforth-0.5.0 ran much faster when compiled with gcc-2.95.3 than with gcc-2.96-85 or gcc-3.2.1. Test notes (about 350 lines) are at http://www.nemr.net/~dnarnold/gforth-listings/testlog.20021230.gforth-0.5.9.gz Is this a suitable way to report informal test results? Regards, David Arnold ===== --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
