[4166] in Athena Bugs
Possible VAX f77 compiler bug
daemon@ATHENA.MIT.EDU (nhdoerry@ATHENA.MIT.EDU)
Thu Feb 8 11:51:07 1990
From: nhdoerry@ATHENA.MIT.EDU
To: bugs@ATHENA.MIT.EDU
Cc: nhdoerry@ATHENA.MIT.EDU
Date: Thu, 08 Feb 90 11:50:34 EST
Hi,
For the past few days, I've been trying to get the latest version
of SHCP (Ship Hull Characteristics Program) running on Project
Athena. This program has about 600K of source code and is
located in the 13.ship course locker.
The program compiles and runs on the IBM RT, but suffers from a
floating point divide by zero error on a VAX Station II
which can be traced to the following line where
PSCALV = 4.29331e-39
CALL SCALF ( PSCALV, PSCALV, 0.0, -(YCEN/PSCALV) )
The following is taken from a DBX session showing that
PSCALV should be assigned the value of 0.125, but for
some reason doesn't :
*********************************************************
7868 800 CONTINUE
7869 IF ( IPLOT .LE. 0 ) GO TO 1000
7870 IF ( NAP .GE. 1 .AND. IPLTAP .LE. 0 ) GO TO 1000
(dbx) l
7871 C
7872 C STORE PLOT SCALE FACTOR IN PSCALV AND DON'T RESET WHEN PLOTTING
7873 C APPENDAGES.
7874 C
7875 IF ( NAP .GE. 1 ) THEN
7876 CONTINUE
7877 ELSE
7878 IF (PSCAL .GT. 0.0 ) THEN
7879 PSCALV = PSCAL
7880 ELSE
(dbx) stop at 7875
[5] stop at "shcpv3.f":7875
(dbx) cont
[5] stopped in table at line 7875 in file "shcpv3.f"
7875 IF ( NAP .GE. 1 ) THEN
(dbx) p nap
0
(dbx) n
stopped in table at line 7878 in file "shcpv3.f"
7878 IF (PSCAL .GT. 0.0 ) THEN
(dbx) p pscal
0.125
(dbx) p pscalv
4.29331e-39
(dbx) n
stopped in table at line 7879 in file "shcpv3.f"
7879 PSCALV = PSCAL
(dbx) n
stopped in table at line 7880 in file "shcpv3.f"
7880 ELSE
(dbx) n
stopped in table at line 7885 in file "shcpv3.f"
7885 IF ( IPLOT .EQ. 1 ) THEN
(dbx) p pscal
0.125
(dbx) p pscalv
4.29331e-39
(dbx)
******************************************************************
As a side note, this program takes up so much memory, that when it is
executed with the DBX debugger, I often get the following error
when trying to start new processes.
sorry, pid 28360 was killed in xalloc: no swap space
Could the memory limitations on the VAX station II's cause the compiler
problems? If so, what is the solution?
Thank you.
Norbert Doerry
Course 13A software admistrator
nhdoerry@athena.mit.edu
x3-0050
628-2153
