🚀 go-pugleaf

RetroBBS NetNews Server

Inspired by RockSolid Light RIP Retro Guy

Article View: gnu.gcc.help
Article #32

Re: cross compiling produces worse code

#32
From: H.J. Lu
Date: Tue, 03 Nov 1992 23:31
19 lines
1375 bytes
In article <Bx5M21.899@cs.vu.nl>, jwromei@cs.vu.nl (John Romein) writes:
|> I succesfully configured and build gcc 2.3.1 as cross compiler, running on
|> a sun4, producing code for a sun3.  I also build the compiler to run on a
|> sun3 and to produce code for a sun3.
|> However, a program compiled with the cross compiler runs almost twice as slow
|> as the program compiled with the non-cross compiler (on a sun3)
|> Both compilers were told to optimize with the -O2 flag.
|> Both compilers produced hardware floating point instructions.
|> Is it possible that -O2 does less optimizations on the cross compiler than on
|> the non-cross compiler ?
|> 
|> John


Same thing happens to me when I cross-compile from Ultrix to i386. Any
solutions? BTW, if I crossc-compile from the different OSes but running
on the same CPU, i.e., i386, everything seems fine.

H.J.
Technical Details
Message-ID: <1992Nov3.233117.21861@serval.net.wsu.edu>
Path: arc1-ca.pugleaf.net!sparky!uunet!ogicse!cs.uoregon.edu!news.uoregon.edu!news.u.washington.edu!serval!hlu
References: <Bx5M21.899@cs.vu.nl>

Show Full Headers
Complete Article Headers
Path: arc1-ca.pugleaf.net!sparky!uunet!ogicse!cs.uoregon.edu!news.uoregon.edu!news.u.washington.edu!serval!hlu
From: hlu@eecs.wsu.edu (H.J. Lu)
Newsgroups: gnu.gcc.help
Subject: Re: cross compiling produces worse code
Message-ID: <1992Nov3.233117.21861@serval.net.wsu.edu>
Date: 3 Nov 92 23:31:17 GMT
Article-I.D.: serval.1992Nov3.233117.21861
References: <Bx5M21.899@cs.vu.nl>
Sender: news@serval.net.wsu.edu (USENET News System)
Organization: School of EECS, Washington State University
Lines: 19