Article View: gnu.gcc.help
Article #51632gcc for 68040 replace line-f instructions with inline asm
From: Rainer Menes
Date: Tue, 27 Apr 1993 15:28
Date: Tue, 27 Apr 1993 15:28
12 lines
1350 bytes
1350 bytes
Yesterday I have run a test where saw the slowness of the line-f trap mechanism for unimplemented instructions hured my performance. I wonder if some has do a implementation of gcc 2.33 where for example 'fsin' is replace by inline assembler the 68040 know and without using the trap handler to emulate it. This is done for the instructions fintrz and so on but not for transcendetal functions. I know here the line-f trap is fast compared to the calculation time of the sin but if you do millions of them this will coast you alot of time. Thanks for your help Rainer Menes email: menes@statistik.tu-muenchen.de
Technical Details
Message-ID:
<1993Apr27.152852.7906@Informatik.TU-Muenchen.DE>Path:
arc1-ca.pugleaf.net!gmd.de!newsserver.jvnc.net!howland.reston.ans.net!ira.uka.de!math.fu-berlin.de!informatik.tu-muenchen.de!menesShow Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: 11318c,3f5892110af24b97,start X-Google-Attributes: gid11318c,public X-Google-ArrivalTime: 1993-04-27 08:39:06 PST Newsgroups: gnu.gcc.help Path: arc1-ca.pugleaf.net!gmd.de!newsserver.jvnc.net!howland.reston.ans.net!ira.uka.de!math.fu-berlin.de!informatik.tu-muenchen.de!menes From: menes@statistik.tu-muenchen.de (Rainer Menes) Subject: gcc for 68040 replace line-f instructions with inline asm Originator: menes@suniams2.statistik.tu-muenchen.de Sender: menes@Informatik.TU-Muenchen.DE (Rainer Menes) Organization: Technische Universitaet Muenchen, Germany Date: Tue, 27 Apr 1993 15:28:52 GMT Message-ID: <1993Apr27.152852.7906@Informatik.TU-Muenchen.DE> Lines: 13