i386 performance problem due to regmove.c change

John Wehle john@feith.com
Sun Jun 28 16:16:00 GMT 1998


> I would be very curious to know if this performance regression went
> away after my changes to recompute register set/reference counts.

Using crafty (a chess program) v14.13 with the assembler routines
disabled and running the internal bench mark (higher numbers are
better) on a 233 MHz Pentium II running FreeBSD-current results in:

egcs-19980608                                                : 1.01
egcs-19980621 with regmove.c & local-alloc.c from 1998-06-08 : 1.01
cvs 1998-06-28                                               : 0.95
cvs 1998-06-28 with regmove.c from 1998-06-08                : 0.99

It appears that the regmove regression is still present and that there
is possibly another new regression.

-- John
-------------------------------------------------------------------------
|   Feith Systems  |   Voice: 1-215-646-8000  |  Email: john@feith.com  |
|    John Wehle    |     Fax: 1-215-540-5495  |                         |
-------------------------------------------------------------------------




More information about the Gcc-patches mailing list