8087 /287 /387 Math Coprocessor Bugs?

Article View: comp.sys.intel

Article #864

#864

From: PENELOPE E. SMITH
Date: Thu, 15 Oct 1992 16:19

19 lines
1184 bytes

I have been having trouble with the math coprocessors in my machines not
obeying instructions such as
    FCOMP  (floating point compare and pop stack)  with denormalized floating
point arguments.  The 8087 and 80287 do not set flags correctly, so sorting
fails to work correctly, and the 80387 seems not to pop the stack, so the
program eventually crashes when the 387 signals a stack overflow.
    Question: Is this a known bug?  What are the usual fixes?

BTW, please EMAIL me.
PS: The platforms are various PC clones running TURBOC 2.0 and the debugger
for verification.
    Thanks,
-- 
 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Penelope E. Smith               PS02@NS.CC.LEHIGH.EDU
Department of Mathematics       PS02@LEHIGH.BITNET
Lehigh University
Bethlehem, PA

Technical Details

Message-ID: <1992Oct15.161925.180185@ns1.cc.lehigh.edu>

Path: arc1-ca.pugleaf.net!sparky!uunet!ukma!cs.widener.edu!netnews.upenn.edu!netnews.cc.lehigh.edu!ns1.cc.lehigh.edu!ps02

Show Full Headers

Complete Article Headers

Path: arc1-ca.pugleaf.net!sparky!uunet!ukma!cs.widener.edu!netnews.upenn.edu!netnews.cc.lehigh.edu!ns1.cc.lehigh.edu!ps02
From: ps02@ns1.cc.lehigh.edu (PENELOPE E. SMITH)
Newsgroups: comp.sys.intel
Subject: 8087 /287 /387 Math Coprocessor Bugs?
Message-ID: <1992Oct15.161925.180185@ns1.cc.lehigh.edu>
Date: 15 Oct 92 16:19:25 GMT
Organization: Lehigh University
Lines: 19