Date: Sat, 13 May 2000 14:44:19 -0400 (EDT) From: Ken Seggerman <suleyman@echonyc.com> To: freebsd-questions@FreeBSD.org Subject: gcc's bad results on Celeron 500 running 3.3 RELEASE Message-ID: <Pine.GSO.4.21.0005131428510.1792-100000@echonyc.com>
next in thread | raw e-mail | index | archive | help
Greetings: I am running FreeBSD 3.3 RELEASE on a brand new Intel Celeron 500 mhz PC, as well on a much older Pentium 133 mhz PC. When compiling and running a mathematically intensive program in C on the Celeron 500 using both the gcc 2.7.2.3 that came with Release 3.3, and with the gcc 2.95.2 that I installed, I consistently get wrong results (I am #including <math.h> and linking in the math library with -lm). Compiling the same code in -verbose mode on the old 133 Pentium gives identical (except for the names of the temporary files in /var/tmp) verbose output to stderr, and when the binary is run it gives correct results. Compiling and running the same code on the three remote multi-user machines (two SPARCS and a PC running Solaris) under various releases of gcc where I have shell access and compilation privileges, yields correct results. Booting the Celeron 500 under Windows NT and compiling and running the code with a commercial (MSVC 6.0) compiler gives me correct results. Running a program of a similar nature (xephem 4.28) pre-compiled as a package for FreeBSD 3.1 on the two machines gives me identical (and correct) results. Is this a gcc problem, or a FreeBSD 3.3 problem? Is there anything I can do about it? Is it fixed in FreeBSD 4.0 Release? Thanks, Ken Seggerman ken_seggerman@suleyman.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.GSO.4.21.0005131428510.1792-100000>