Is there a known bug in gcc 2.1 with -O?

Article View: gnu.gcc.help

Article #90

#90

From: John LoSecco
Date: Thu, 09 Apr 1992 17:50

26 lines
1642 bytes

	Is there a known bug in gcc 2.1?  I have recompiled some code that
ran fine under gcc 1.40.  It core dumps if I compile with -O4 or with -O.
It seems to be fine if I do not optimize.

	The code is a floating point intensive Monte Carlo but I have checked
and the problem is also found with smaller simpler programs.

	I am running gcc 2.1 on SCO 3.2.2.  I have a math coprocessor.  The
simple program dies with a Floating Exception.  No problem if I leave out the
-O.
	I have not checked the performance.  Is 2.1 without -O as fast as
1.40 for floating point?

	Thanks for any tips.


John M. LoSecco			Phone:(219)239-6044          Fax:(219)239-5952
Physics Department		HEPnet/SPAN: UNDHEP::LOSECCO or 47344::LOSECCO
University of Notre Dame		Internet: losecco@undpdk.hep.nd.edu
Notre Dame, Indiana 46556 USA		Bitnet: losecco@undhep
-- 
John M. LoSecco			Phone:(219)239-6044          Fax:(219)239-5952
Physics Department		HEPnet/SPAN: UNDHEP::LOSECCO or 47344::LOSECCO
University of Notre Dame		Internet: losecco@undpdk.hep.nd.edu
Notre Dame, Indiana 46556 USA		Bitnet: losecco@undhep

Technical Details

Message-ID: <1992Apr9.175039.24156@news.nd.edu>

Path: arc1-ca.pugleaf.net!sparky!uunet!spool.mu.edu!news.nd.edu!undpdk!losecco

Show Full Headers

Complete Article Headers

Path: arc1-ca.pugleaf.net!sparky!uunet!spool.mu.edu!news.nd.edu!undpdk!losecco
From: losecco@undpdk.hep.nd.edu (John LoSecco)
Newsgroups: gnu.gcc.help
Subject: Is there a known bug in gcc 2.1 with -O?
Summary: Code dies with -O or -O4.  OK with no optimize
Keywords: gcc 2.1 known bugs optimize 386
Message-ID: <1992Apr9.175039.24156@news.nd.edu>
Date: 9 Apr 92 17:50:39 GMT
Article-I.D.: news.1992Apr9.175039.24156
Sender: news@news.nd.edu (USENET News System)
Distribution: na
Organization: Notre Dame High Energy Physics
Lines: 26