WebMO - Computational chemistry on the WWW
Recent news

WebMO 13.0 is now available for free download!

WebMO 13.0 Pro and Enterprise have a variety of additional features and is available for purchase.

WebMO can now build bulk, periodic (1D, 2D, 3D) materials. Periodic calculations are supported in Gaussian, PWSCF, and VASP!

August 24, 2019

Trouble running GAMESS... Log Out | Topics | Search
Moderators | Edit Profile

WebMO Support Forum » Running WebMO Jobs » Other » Trouble running GAMESS... « Previous Next »

Author Message
joe cammisa
Unregistered guest
Posted on Friday, September 04, 2009 - 3:38 pm:   

using webmo_pro_9, my attempts to submit gamess jobs to a remote server invariably fail with errors like these (from raw output):
ERROR *** ATTEMPTING A BOGUS READ OF A DAF RECORD.
RECORD NUMBER 40 OF LENGTH 0 HAS NO LENGTH.

rungms on the command line of said host is not a problem; can also submit and run gaussian jobs to the same host with success.

while watching the scratch area on this host with my job "running", i notice that the *.F08 file (AOINTS) is empty--definitely not the case for other such files associated with the run, nor for the *.F08 of command-line submitted gamess jobs.

any ideas how i might resolve this issue would be appreciated gratefully. thanks...
JR Schmidt
Moderator
Username: Schmidt

Post Number: 115
Registered: 11-2006
Posted on Friday, September 04, 2009 - 4:23 pm:   

There is a GAMESS bug when running a calculation on a single atom. The problem does not appear for calculations on more than one atom. (You can also use Cartesian coordinates to solve the problem).
Daniel Gruner
Unregistered guest
Posted on Saturday, September 26, 2009 - 9:20 pm:   

I too encountered a similar problem. I am running gamess on a cluster, compiled with the Intel 11.1 compilers, and using Intel MPI 3.2.2. The job crashes with the following error:
ERROR *** ATTEMPTING A BOGUS READ OF A DAF RECORD.
RECORD NUMBER 113 OF LENGTH 22201 WAS NEVER PREVIOUSLY WRITTEN.

THIS ERROR IS LIKELY TO BE A BUG IN THE PROGRAM, BUT
PLEASE CONSIDER INPUT ERRORS AS A POSSIBLE CAUSE, TOO.
MANUAL CHAPTER -PROG.DOC- HAS A LIST OF ALL DIRECT ACCESS FILE RECORD NUMBERS,
WHICH WILL HELP YOU UNDERSTAND WHAT THE INCORRECT DATA IS SUPPOSED TO BE.

EXECUTION OF GAMESS TERMINATED -ABNORMALLY- AT Sat Sep 26 20:45:34 2009
514181 WORDS OF DYNAMIC MEMORY USED
CPU 0: STEP CPU TIME= 0.00 TOTAL CPU TIME= 261.7 ( 4.4 MIN)
TOTAL WALL CLOCK TIME= 263.1 SECONDS, CPU UTILIZATION IS 99.46%
A fatal error occurred on DDI Process 0.
[cli_0]: A fatal error occurred on DDI Process 2.
A fatal error occurred on DDI Process 1.
aborting job:
...

Here is the input file:

! File created by MacMolPlt 7.2.1
$CONTRL SCFTYP=RHF RUNTYP=TDHF MAXIT=99
MULT=1 NOSYM=1 INTTYP=HONDO ICUT=20
ITOL=30 $END
$TDHF NFREQ=2 FREQ(1)= 0.00, 0.04428
GTHG=.TRUE.
INIG=1 BEOPE=.FALSE.
BOR=.FALSE. GEFISH=.FALSE.
GIDRI=.FALSE. GOKE=.FALSE.
RETDHF=.FALSE. $END
$SYSTEM TIMLIM=525600 MWORDS=90
MEMDDI=360 PARALL=.TRUE. $END
$BASIS GBASIS=N31 NGAUSS=6
NDFUNC=1 DIFFSP=.TRUE. $END
$SCF DIRSCF=.TRUE. DIIS=.TRUE. $END
$DATA
Toluene_NLO
C1
C 6.0 1.90270 0.00000 0.00860
C 6.0 1.19960 1.20340 0.00210
C 6.0 -0.19420 1.20060 -0.00910
C 6.0 -0.91310 0.00000 -0.01150
C 6.0 -0.19420 -1.20060 -0.00910
C 6.0 1.19960 -1.20340 0.00210
C 6.0 -2.42310 0.00000 0.00950
H 1.0 2.98680 0.00000 0.01430
H 1.0 1.73600 2.14610 0.00170
H 1.0 -0.73190 2.14360 -0.01820
H 1.0 -0.73190 -2.14360 -0.01820
H 1.0 1.73600 -2.14610 0.00170
H 1.0 -2.80080 0.00000 1.03800
H 1.0 -2.82900 0.88440 -0.48730
H 1.0 -2.82900 -0.88440 -0.48730
$END


This job has run properly on a cluster of 12 Macs, running in 32-bit mode. The current test is on a single 8-core node, in 64-bit mode.
Thorsteinn Adalsteinsson
Unregistered guest
Posted on Thursday, January 26, 2012 - 12:17 am:   

I am having the hardest time to get GAMESS to run after I upgraded my webMO interface to version 12.004-pro.
I am running webMO and all the computational packages on Os X and have been doing so in the past 2 years without a hitch.

I currently have Mopac and Gaussian running on the interface.

I started having problems getting Gamess 2010 to run, so I recently updated the program to 2010 to Aug112011 version of the program

Gamess runs correctly in the GamessQ, but I haven't bothered getting the rungms file set up.

After submitting jobs I get a log-file that looks reasonable (basis sets are defined etc) until the error:
"
ERROR OPENING PRE-EXISTING FILE ERICFMT,
ASSIGNED TO EXPLICIT FILE NAME $GMSPATH/auxdata/ericfmt.dat,
PLEASE CHECK THE -SETENV- FILE ASSIGNMENTS IN YOUR -RUNGMS- SCRIPT.
"
gets thrown. The previous version (2010) had the same problem with "$GMSPATH/ericfmt.dat" in stead of the auxdata sub-folder. The ericfmt.dat file is in the proper place and I have a feeling that the error is in the run-script that webMO submits, since the program runs properly in the GamessQ.app environment.

Any suggestions?

Thanks


Thorsteinn
JR Schmidt
Moderator
Username: Schmidt

Post Number: 221
Registered: 11-2006
Posted on Thursday, January 26, 2012 - 10:14 am:   

This issue is resolved in WebMO 12.1 -- to be released in a few days.
Thorsteinn Adalsteinsson
Unregistered guest
Posted on Thursday, January 26, 2012 - 12:31 pm:   

Great news!

Thank you for the prompt reply
Megan
Unregistered guest
Posted on Wednesday, March 21, 2012 - 3:02 pm:   

I just installed gamess-OSX.Current.x86-64.tar and I am receiving the same error. I'm not 100% sure it is in the right place but webmo is finding the rest of the games files so I am hesitant to change the location. Any suggestions?

Thanks
Megan
JR Schmidt
Moderator
Username: Schmidt

Post Number: 239
Registered: 11-2006
Posted on Wednesday, March 21, 2012 - 3:06 pm:   

Make sure to update to the most recent version of WebMO (12.1.002), which I believe should resolve this issue.
John Glenn Ramon
Unregistered guest
Posted on Tuesday, June 26, 2012 - 6:27 pm:   

I just updated my gamess-OSX version to May12012R1. I also use the most up to date version of WebMO 12.1.004. I have run the tests for gamess successfully in the command line. However, when I used it in WebMO, it gave me this message:

YOU MUST ASSIGN GENERIC NAME PUNCH WITH A SETENV.

When I looked at run_gamess.cgi, it would seemed to me that these are being read from rungms. In the latest version of gamess (above), I noticed that they have separated all these SETENV in a separate file called gms-files.csh. I don't know when they did this but it would seemed fairly recently and I think its the source of the problem. Please advise.

Thanks!
John
JR Schmidt
Moderator
Username: Schmidt

Post Number: 263
Registered: 11-2006
Posted on Tuesday, June 26, 2012 - 10:22 pm:   

Argh. It seems the change this in EVERY release. I just sent a note to the developers, hoping that we can arrive at a more satisfactory solution in the future. I hope to be able to use rungms in future release, contingent upon some tweaks to rungms by the GAMESS folks.

The simplest solution is to just to copy these assignments back to rungms from the separate file. Note that this ONLY effects the OS X version, since rungms on Linux is unmodified (at least of the most recent release).
John Glenn Ramon
Unregistered guest
Posted on Wednesday, June 27, 2012 - 4:13 pm:   

Well, I was already contemplating on that solution (transferring all the assignments back). I was just looking for a second opinion and see if you were looking into it already. I'll try it and see what happens. Thanks!
John Glenn Ramon
Unregistered guest
Posted on Wednesday, June 27, 2012 - 5:48 pm:   

Ok, I got it to work! First, I copied the whole gms-files.csh back into rungms without any editing right after the comment block:

#
# ---- the top third of the script is input and other file assignments ----
#

Second, just to play on the safe side, I commented our the

I also had to change setenv AUXDATA to explicitly point to where the auxdata directory was located because having $GMSPATH/auxdata gave me the following error:


ERROR OPENING PRE-EXISTING FILE ERICFMT,
ASSIGNED TO EXPLICIT FILE NAME /auxdata/ericfmt.dat,
PLEASE CHECK THE -SETENV- FILE ASSIGNMENTS IN YOUR -RUNGMS- SCRIPT.


I guess it has something to do with how run_gamess.cgi reads and assigns the value of AUXDATA from rungms.

Well, I hope this helps for anyone who is having trouble with running gamess through WebMO on a Mac OSX. At least, until the next update!
Btw, this works for both PowerPC and Intel-based Macs coz' I have both.

John
Anonymous
 
Posted on Friday, August 10, 2012 - 3:33 pm:   

Thanks John! It worked also in linux32 with gfortran (had the same error with PUNCH).

Mihai
Anonymous
 
Posted on Friday, April 05, 2013 - 1:30 pm:   

What does the error message "Unable to generate principal axes" mean and how do I fix it?

Add Your Message Here
Post:
Username: Posting Information:
This is a public posting area. Enter your username and password if you have an account. Otherwise, enter your full name as your username and leave the password blank. Your e-mail address is optional.
Password:
E-mail:
Options: Post as "Anonymous"
Enable HTML code in message
Automatically activate URLs in message
Action:

Topics | Last Day | Last Week | Tree View | Search | Help/Instructions | Program Credits Administration