|
|
 Up |
|
|
  |
Author: Carlie J. CoatsCarlie J. Coats Date: Nov 21, 2007 08:39
On a server I'm using for my work, I'm using"
#52 which idb
/ opt/intel/idbe/9.1.051/bin/idb
#53 which ifort
/ opt/intel/fce/9.1.051/bin/ifort
#54 ifort -V
Intel(R) Fortran Compiler for Intel(R) EM64T-based applications,
Version 9.1 Build 20070510 Package ID: l_fc_c_9.1.051
Copyright (C) 1985-2007 Intel Corporation. All rights reserved.
with debug-and-check-everything compiler flags:
FFLAGS = -C -g -O0 -traceback -debug-parameters all -ftrapuv \
-check bounds -check uninit -check noarg_temp_created \
-ip -tpp7 -ftz
with relevant code segment below, where I'm going into the
"CALL ADVHS" line, so that at that point CON(:,:) and CWORK(:,C,:)
should match, for second subscript in the range [1,NROWS].
But they don't, as shown by the debugger-output extracts following...
This is very disconcerting ;-(
Any ideas?
|
| Show full article (6.40Kb) |
|
| | 4 Comments |
|
|
Author: Herman D. KnobleHerman D. Knoble Date: Nov 21, 2007 09:05
Carlie: You didn't post compilable code, so I can't run it thru the Lahey compiler
with Polyhedron debug options:
lf95 myprog --chkglobal -g --co --f95 --lst --trace --nsav --warn
Regarding Intel compiler options, you might try:
ifort myprog -check all -d3 -g -fp -O0 -traceback -WB -V -ftrapuv -fltconsistency -fpe0
-no-ftz
or as Polyhedron uses for their diagnostic benchmarks:
ifort myprog -check all -warn all,nodec,interfaces -gen_interfaces -traceback -fpe0
-fpstkchk
Skip Knoble
On Wed, 21 Nov 2007 16:39:20 GMT, "Carlie J. Coats" jyarborough.com> wrote:
-|
-|On a server I'm using for my work, I'm using"
-|
-| #52 which idb
-| /opt/intel/idbe/9.1.051/bin/idb
-| #53 which ifort
-| /opt/intel/fce/9.1.051/bin/ifort
-| #54 ifort -V
-| Intel(R)...
|
| Show full article (7.49Kb) |
|
| | no comments |
|
|
Author: James Van BuskirkJames Van Buskirk Date: Nov 21, 2007 09:13
"Carlie J. Coats" jyarborough.com> wrote in message
news:YaZ0j.277$Xg.84@trnddc06...
> CON(NDYNSD,R) = MFAC
> CON(V,R) = DFAC * SBDY( V,C,L )
> CON(V,C) = DFAC * SBDY( V,C,L )
> CON(V,R) = CWORK( V,C,R )
> CON(V,R) = DFAC * NBDY( V,C,L )
> CON(V,C) = DFAC * NBDY( V,C,L )
> CALL ADVHS( NROWS, CON, VEL, DTS, DX2 )
I can't make heads or tails of your code, but why do you index array
CON with C a couple of times? Shouldn't that overwrite a just the
stretch of CON that the debugger is showing?
--
write(*,*) transfer((/17.392111325966148d0,6.5794487871554595D-85, &
6.0134700243160014d-154/),(/'x'/)); end
|
| |
| no comments |
|
|
Author: Carlie J. CoatsCarlie J. Coats Date: Nov 21, 2007 10:06
James Van Buskirk wrote:
> "Carlie J. Coats" jyarborough.com> wrote in message
> news:YaZ0j.277$Xg.84@trnddc06...
> >> CON(NDYNSD,R) = MFAC
>
>> CON(V,R) = DFAC * SBDY( V,C,L )
>
>> CON(V,C) = DFAC * SBDY( V,C,L )
>
>> CON(V,R) = CWORK( V,C,R )
>
>> CON(V,R) = DFAC * NBDY( V,C,L )
>
>> CON(V,C) = DFAC * NBDY( V,C,L )
>
>> CALL ADVHS( NROWS, CON, VEL, DTS, DX2 )
>
> I can't make heads or tails of your code, but why do you index array
> CON with C a couple of times? Shouldn't that overwrite a just the ...
|
| Show full article (1.27Kb) |
| no comments |
|
|
|
|
|
Author: dpbdpb Date: Nov 21, 2007 10:18
Carlie J. Coats wrote:
> James Van Buskirk wrote:
...
>> I can't make heads or tails of your code, but why do you index array
>> CON with C a couple of times? Shouldn't that overwrite a just the
>> stretch of CON that the debugger is showing?
>
> That did it! Thanks! Sometimes I "know" what I've done so well
> that I can't see my own typos ;-(
...
And, not just C, the subscript R is used twice in subsequent lines as
well so unless CWORK() is a function that modifies R instead of an
array, that looks suspicious, too...
--
|
| |
| no comments |
|
|
