 |
 |
On Mar 8, 2017, at 1:06 PM, Bradley T Yale <[log in to unmask]> wrote:We were using gfortran version 3.8.5.That should be changed then.If the benign errors are the same ones that I'm seeing, then they do stop the compilation process, if enough of them occur apparently:Error: Global name 'counters_out' at (1) is already being used as a SUBROUTINE at (2)
egs5job.f:25119.72:
subroutine annih......Fatal Error: Error count reached limit of 25.
Script stopped -- egs5job.f did not compile and/or linkFrom: [log in to unmask] <[log in to unmask]> on behalf of Maruyama, Takashi <[log in to unmask]>
Sent: Wednesday, March 8, 2017 12:49:30 PM
To: 'Maurik Holtrop'; Nathan Baltzell
Cc: McCormick, Jeremy I.; Bradley T Yale; hps-software
Subject: RE: egs5 compilationThe EGS5 maintainer told me about the gfortran issues,
http://rcwww.kek.jp/research/egs/egs5-compiler-problems.html
Takashi
-----Original Message-----
From: Maurik Holtrop [mailto:[log in to unmask]]
Sent: Wednesday, March 08, 2017 9:35 AM
To: Nathan Baltzell
Cc: McCormick, Jeremy I.; Maruyama, Takashi; Bradley Yale; hps-software
Subject: Re: egs5 compilation
Good afternoon,
I have a basic fix to get a running version of EGS5 on CENTOS7. There are still some warning messages when you compile using gfortran, however I think those are “benign”. They are warnings that the original Fortarn code is, well, not quite up to snuff.
My code is in /home/holtrop/egs5.
Best,
Maurik
PROCEDURE:
1) Recompile stdhep libraries with gfortran. This requires editing the src/stdhep_arch file and changing at the top (add -fd-lines-as-comments):
ifndef DEBUG
DEBUG = none
DFLG = -fd-lines-as-comments
SFX =
else
DFLG = $(DEBUG)
SFX = _g
endif
And further down, change g77 to gfortran.
Make the same changes to mcfio/arch_mcfio
2) Build EGS5 pointing to the new version of stdhep libraries by changing src/egs5run.
I did not actually run EGS5, I only compiled it.
On Mar 7, 2017, at 6:49 PM, Nathan Baltzell <[log in to unmask]> wrote:
I did look but didn't find a 77 compiler on centos7 in the usual places. I also noticed some recent ccpr's asking for one, without resolution yet.
Seems reasonable if someone that knows more about egs5/fortran can see if it really requires f77/g77 or can be built with gfortan (maybe it just requires some particular compiler options?). Or, yes, maybe just build fg77 for hps on centos7 instead of waiting on cc who seems to be backlogged.
On Mar 7, 2017, at 6:35 PM, "McCormick, Jeremy I." <[log in to unmask]> wrote:
If needed, it is possible to build our own f77 by downloading a specific gcc release and compiling it on centos7. Not that difficult actually...
But I would ask CC if they have an f77 compiler already available there, as there are a lot of site packages maintained by them and it might already be there someplace.
-----Original Message-----
From: [log in to unmask] [mailto:[log in to unmask]] On Behalf Of Maruyama, Takashi
Sent: Tuesday, March 07, 2017 2:04 PM
To: 'Bradley T Yale'; hps-software
Subject: RE: egs5 compilation
I am looking into it.
Takashi
-----Original Message-----
From: Bradley T Yale [mailto:[log in to unmask]]
Sent: Tuesday, March 07, 2017 11:07 AM
To: hps-software
Cc: Maruyama, Takashi
Subject: egs5 compilation
There is an issue with compiling egs5 using gfortran (4.8.5), that somebody familiar with it can hopefully shed some light on. JLab's egs5 normally uses g77, but is no longer present after the centos7 upgrade.
Switching out g77 for gfortran in egs5run and trying to compile generates a lot of "Warning: Interface mismatch in dummy procedure" errors. Does anyone know what could be causing this?
The issue can be replicated by going to
/u/group/hps/production/mc/egs5
and running
src/build.sh lhe_v1
The compiler and flags are set in line 189 of src/egs5run in the same directory. It may be easier to try to get g77 back at JLab, but was just wondering why gfortran doesn't work, or if the errors are related to something else.
########################################################################
Use REPLY-ALL to reply to list
To unsubscribe from the HPS-SOFTWARE list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=HPS-SOFTWARE&A=1
########################################################################
Use REPLY-ALL to reply to list
To unsubscribe from the HPS-SOFTWARE list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=HPS-SOFTWARE&A=1
########################################################################
Use REPLY-ALL to reply to list
To unsubscribe from the HPS-SOFTWARE list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=HPS-SOFTWARE&A=1
########################################################################
Use REPLY-ALL to reply to list
To unsubscribe from the HPS-SOFTWARE list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=HPS-SOFTWARE&A=1Use REPLY-ALL to reply to listTo unsubscribe from the HPS-SOFTWARE list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=HPS-SOFTWARE&A=1
Use REPLY-ALL to reply to list
To unsubscribe from the HPS-SOFTWARE list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=HPS-SOFTWARE&A=1