[Search for users]
[Overall Top Noters]
[List of all Conferences]
[Download this site]
| Title: | Digital Fortran |
| Notice: | Read notes 1.* for important information |
| Moderator: | QUARK::LIONEL |
|
| Created: | Thu Jun 01 1995 |
| Last Modified: | Fri Jun 06 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 1333 |
| Total number of notes: | 6734 |
1245.0. "%SYSTEM-F-HPARITH w DFAV on Alpha V7.1 / DPML" by CSC32::D_SANFORD () Fri Mar 28 1997 19:50
Digital Fortran 77 V7.1-107-3313
OpenVMS Alpha V7.1
The following example demonstrates a %SYSTEM-F-HPARITH on OpenVMS Alpha
V7.1, the same example will work with V6.2. I compiled using Fortran
77 V6.3 and Fortran 90 V7.1 as well, same results. Something to do with
the comet? The failure is in the DPML.
Regards, Drew Sanford
Customer Support Center
C970328-4277
real k/0.153144e-2/, v1/0.3369278/
real ugamma
type 100, k, k
type 100, v1, v1
100 FORMAT(Z, F)
ugamma = v1**(1.0/k)
type *, ugamma
end
$ fort/noopt test
$ link/deb test
$ run test
OpenVMS Alpha DEBUG Version V7.1-000
%DEBUG-I-INITIAL, Language: FORTRAN, Module: D$MAIN
DBG> g
BA993BC8 0.0015314
81CD3FAC 0.3369278
%SYSTEM-F-HPARITH, high performance arithmetic trap, Imask=00000000,
Fmask=00400000, summary=02, PC=FFFFFFFF804FAFA8, PS=0000001B
-SYSTEM-F-FLTINV, floating invalid operation, PC=FFFFFFFF804FAFA8, PS=0000001B
break on unhandled exception preceding SHARE$DPML$SHR_CODE0+290728
DBG>
| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 1245.1 | Appears to be a problem in the DPML | RTL::ARNOLD | The Math Library Group | Mon Mar 31 1997 12:08 | 7 |
| We can reproduce the problem.
It appears we've got a problem in the DPML with a "borderline" result
which underflows to zero, and we end up trying to work with a VAX
"dirty zero."
I'll reply here when we've got something definite.
|
| 1245.2 | | CSC32::D_SANFORD | | Tue Apr 22 1997 14:10 | 6 |
| A formal IPMT case will be rasied, using an older DPML$SHR.EXE on
Alpha VMS V7.1 caused the application to ACCVIO. Due to the size
of the application we feel it is best to get this problem fixed
in a ECO until we can find a better workaround.
-drew
|
| 1245.3 | | QUARK::LIONEL | Free advice is worth every cent | Tue Apr 22 1997 14:47 | 3 |
| Ok - just make sure that it's raised against DPML/VMS and not Fortran.
Steve
|