From owner-chemistry@ccl.net Sat Oct 31 01:59:01 2009 From: "Abhinav Kumar abhinavmarshal*_*gmail.com" To: CCL Subject: CCL: Suitable basis set for Hg atom Message-Id: <-40577-091031014935-32317-gTiDZFcRGYazRq3OTiQFQA{:}server.ccl.net> X-Original-From: "Abhinav Kumar" Date: Sat, 31 Oct 2009 01:49:32 -0400 Sent to CCL by: "Abhinav Kumar" [abhinavmarshal-,-gmail.com] Dear All, I am using CEP-121G or LANL2DZ basis set for Hg atom in one of my calculations. But the warning Warning! Hg atom 34 has 80 valence electrons but only 34 basis functions.This is less than a minimal basis set! Warning! Hg atom 67 has 80 valence electrons but only 34 basis functions. This is less than a minimal basis set! is coming with both the basis sets. I shall be grateful if anyone kindly suggest me the appropriate basis set for the Hg atom. Dr Abhinav Kumar Department of Chemistry Faculty of Science Lucknow University Lucknow 226007 INDIA Email: abhinavmarshal{=}gmail.com From owner-chemistry@ccl.net Sat Oct 31 04:12:01 2009 From: "Marcel Swart marcel.swart(_)icrea.es" To: CCL Subject: CCL: Suitable basis set for Hg atom Message-Id: <-40578-091031040631-5868-XhkwBJnMfZZXHxBnRJjVlA*server.ccl.net> X-Original-From: Marcel Swart Content-Disposition: inline Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset=ISO-8859-1; DelSp="Yes"; format="flowed" Date: Sat, 31 Oct 2009 08:59:00 +0100 MIME-Version: 1.0 Sent to CCL by: Marcel Swart [marcel.swart,+,icrea.es] You made a mistake with the input, in which you specified only the valence part, but not the ECP part. If you used GEN, the CEP-121G or LanL2DZ should be specified on two lines, if I remember correctly. Quoting "Abhinav Kumar abhinavmarshal*_*gmail.com" = : > > Sent to CCL by: "Abhinav Kumar" [abhinavmarshal-,-gmail.com] > Dear All, > > I am using CEP-121G or LANL2DZ basis set for Hg atom in one of my =20 > calculations. But the warning > Warning! Hg atom 34 has 80 valence electrons but only 34 basis =20 > functions.This is less than a minimal basis set! > Warning! Hg atom 67 has 80 valence electrons but only 34 basis =20 > functions. This is less than a minimal basis set! > > is coming with both the basis sets. I shall be grateful if anyone =20 > kindly suggest me the appropriate basis set for the Hg atom. =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D dr. Marcel Swart ICREA Research Professor at Institut de Qu=EDmica Computacional Universitat de Girona Parc Cient=EDfic i Tecnol=F2gic Edifici Jaume Casademont (despatx A-27) Pic de Peguera 15 17003 Girona Catalunya (Spain) tel +34-972-183240 fax +34-972-183241 e-mail marcel.swart^^^icrea.es marcel.swart^^^udg.edu web http://www.icrea.cat/Web/ScientificForm.aspx?key=3D372 http://iqc.udg.edu/~marcel =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D From owner-chemistry@ccl.net Sat Oct 31 04:47:01 2009 From: "Morad El-Hendawy m80elhendawy()yahoo.com" To: CCL Subject: CCL: Suitable basis set for Hg atom Message-Id: <-40579-091031043702-19304-wqbBvo1KvyyKDN6vxFYqyQ(a)server.ccl.net> X-Original-From: Morad El-Hendawy Content-Type: multipart/alternative; boundary="0-1552807525-1256978207=:95470" Date: Sat, 31 Oct 2009 01:36:47 -0700 (PDT) MIME-Version: 1.0 Sent to CCL by: Morad El-Hendawy [m80elhendawy:_:yahoo.com] --0-1552807525-1256978207=:95470 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: quoted-printable Dear Abhinav Kumar, I think that CEP-121G or LANL2DZ basis set for Hg atom are not suitable. Yo= u can use the following website to choose the suitable basis set for almost= elements in periodic table: https://bse.pnl.gov/bse/portal I checked CEP-121G or LANL2DZ basis set for Hg atom, but they are not liste= d. Good luck ***************************************************************************= ***********************Morad M. El-Hendawy*********************************= *****************************************************************=A0If you = want a happy life, do four things:=A01- Contemplate and be thankful, 2- The past is gone forever, 3- Today all that you have, 4- Leave the future alone until it come.=A0********************************= ***************************************************************************= ******** --- On Sat, 10/31/09, Abhinav Kumar abhinavmarshal*_*gmail.com wrote: > From: Abhinav Kumar abhinavmarshal*_*gmail.com Subject: CCL: Suitable basis set for Hg atom To: "El-Hendawy, Morad Metwally " Date: Saturday, October 31, 2009, 7:49 AM Sent to CCL by: "Abhinav=A0 Kumar" [abhinavmarshal-,-gmail.com] Dear All, I am using CEP-121G or LANL2DZ basis set for Hg atom in one of my calculati= ons. But the warning=20 Warning!=A0 Hg atom=A0=A0=A034 has 80 valence electrons but only=A0 34 bas= is functions.This is less than a minimal basis set! Warning!=A0 Hg atom=A0=A0=A067 has 80 valence electrons but only=A0 34 bas= is functions. This is less than a minimal basis set! is coming with both the basis sets. I shall be grateful if anyone kindly su= ggest me the appropriate basis set for the Hg atom. Dr Abhinav Kumar Department of Chemistry Faculty of Science Lucknow University Lucknow 226007 INDIA Email: abhinavmarshal[A]gmail.com -=3D This is automatically added to each message by the mailing script =3D-=A0 =A0 =A0=A0 =A0 =A0Subscribe/Unsubscribe:=20 =A0 =A0 =A0Job: http://www.ccl.net/jobs=20=A0 =A0 =A0=0A=0A=0A --0-1552807525-1256978207=:95470 Content-Type: text/html; charset=iso-8859-1 Content-Transfer-Encoding: quoted-printable
Dear Abhinav Kumar,
I think that CEP-121G = or LANL2DZ basis set for Hg atom are not suitable. You can use the followin= g website to choose the suitable basis set for almost elements in periodic = table:
https://bse.pnl.gov/bse/portal
I checked CEP-121G or LANL2DZ b= asis set for Hg atom, but they are not listed.
Good luck

******************= ***************************************************************************= *****
Morad M. El-Hendawy
<= font style=3D"background-color: rgb(230, 230, 230);" color=3D"#00407f" face= =3D"Arial" size=3D"4">***************************************************************= ***********************************
<= /strong> If you want a happy life, do four things:<= /div>
 
1- Contemplate and be thankful,
= 2- The past is gone f= orever,
3- Today all that you have,
4- Leave the future alone until it come. <= /font>
**********************= ***************************************************************************= ******************
3D""
=
--- On Sat, 10/31/09, Abhinav Kumar abhinavmarshal*_*gmail.com &l= t;owner-chemistry/./ccl.net> wrote:
From: Abhinav Kumar abhinavmarshal*_*gmail.com <owner-chemistry/./ccl.ne= t>
Subject: CCL: Suitable basis set for Hg atom
To: "El-Hendawy, M= orad Metwally " <m80elhendawy/./yahoo.com>
Date: Saturday, O= ctober 31, 2009, 7:49 AM


Sent to CCL by= : "Abhinav  Kumar" [abhinavmarshal-,-gmail.com]
Dear All,

I = am using CEP-121G or LANL2DZ basis set for Hg atom in one of my calculations.= But the warning
Warning!  Hg atom   34 has 80 val= ence electrons but only  34 basis functions.This is less than a minima= l basis set!
Warning!  Hg atom   67 has 80 valence = electrons but only  34 basis functions. This is less than a minimal ba= sis set!

is coming with both the basis sets. I shall be grateful if = anyone kindly suggest me the appropriate basis set for the Hg atom.

= Dr Abhinav Kumar
Department of Chemistry
Faculty of Science
Luckno= w University
Lucknow 226007
INDIA
Email: abhinavmarshal[A]gmail.co= m



-=3D This is automatically added to each message by the ma= iling script =3D-
To recover the email address of the author of the mess= age, please change
the strange characters on the top line to the /./ sign.= You can also
E-mail to subscribers: CHEMISTRY/./ccl.net or us= e:
      http://www.ccl.net/cgi-bin/ccl/send_ccl_messa= ge

E-mail to administrators: CHEMISTRY-= REQUEST/./ccl.net or use
      http://www.ccl.net/cg= i-bin/ccl/send_ccl_message
  &nb= sp;   http://www.ccl.net/chemistry/sub_unsub.shtml

Before = posting, check wait time at: http://www.ccl.net

Job: http://www.ccl.net/jobs
Conferences: http://server.ccl.net/chemistry/announcements/conferences/

Se= arch Messages: http://www.ccl.net/chemistry/searchccl/index.shtml=
  &n= bsp;   h= ttp://www.ccl.net/spammers.txt

RTFI: http://www.ccl.net/c= hemistry/aboutccl/instructions/



=0A=0A --0-1552807525-1256978207=:95470-- From owner-chemistry@ccl.net Sat Oct 31 05:34:01 2009 From: "Soren Eustis soreneustis]![gmail.com" To: CCL Subject: CCL: Suitable basis set for Hg atom Message-Id: <-40580-091031050359-5495-fyhWZNQKAlIcVgHzneDzpA++server.ccl.net> X-Original-From: "Soren Eustis" Content-Language: de-ch Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="iso-8859-1" Date: Sat, 31 Oct 2009 09:14:38 +0100 MIME-Version: 1.0 Sent to CCL by: "Soren Eustis" [soreneustis!^!gmail.com] Abhinav, It appears your effective core potentials (ECP) are not being read properly. If you are using external basis sets, be sure to use the = GENECP keyword. If you still get the error, it is likely that it is a simple spacing issue. Try adding or removing a space between the valence and = ECP basis set inputs. Soren Soren N. Eustis, Ph.D. ETH - Z=FCrich Institute of Biogeochemistry and Pollutant Dynamics CHN.F33 soren . env.ethz.ch +41 44 632 93 48 (office) +41 44 632 14 38 (fax) -----Original Message----- > From: owner-chemistry+soreneustis=3D=3Dgmail.com . ccl.net [mailto:owner-chemistry+soreneustis=3D=3Dgmail.com . ccl.net] On Behalf Of = Abhinav Kumar abhinavmarshal*_*gmail.com Sent: Samstag, 31. Oktober 2009 06:50 To: Eustis, Soren Newman Subject: CCL: Suitable basis set for Hg atom Sent to CCL by: "Abhinav Kumar" [abhinavmarshal-,-gmail.com] Dear All, I am using CEP-121G or LANL2DZ basis set for Hg atom in one of my calculations. But the warning=20 Warning! Hg atom 34 has 80 valence electrons but only 34 basis functions.This is less than a minimal basis set! Warning! Hg atom 67 has 80 valence electrons but only 34 basis functions. This is less than a minimal basis set! is coming with both the basis sets. I shall be grateful if anyone kindly suggest me the appropriate basis set for the Hg atom. Dr Abhinav Kumar Department of Chemistry Faculty of Science Lucknow University Lucknow 226007 INDIA Email: abhinavmarshal[A]gmail.com -=3D This is automatically added to each message by the mailing script = =3D-http://www.ccl.net/cgi-bin/ccl/send_ccl_messageSubscribe/Unsubscribe:=20Job: http://www.ccl.net/jobs=20http://www.ccl.net/spammers.txt From owner-chemistry@ccl.net Sat Oct 31 07:51:00 2009 From: "Joe Leonard jleonard42(-)gmail.com" To: CCL Subject: CCL:G: Be a bit cautious... Message-Id: <-40581-091030203429-24128-5tHrVfTFXE04V4vaWBcYyQ(a)server.ccl.net> X-Original-From: Joe Leonard Content-Type: multipart/alternative; boundary=Apple-Mail-1--532829046 Date: Fri, 30 Oct 2009 20:00:45 -0400 Mime-Version: 1.0 (Apple Message framework v936) Sent to CCL by: Joe Leonard [jleonard42],[gmail.com] --Apple-Mail-1--532829046 Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes Content-Transfer-Encoding: 7bit On Oct 30, 2009, at 6:04 PM, John McKelvey jmmckel^^^gmail.com wrote: > Folks, > > I have a dilemma: Significantly results depending on the compiler. Couple of things come to mind: 1) How does the compiler treat uninitialized variables? Is it vax- like, making them default to 0 or C-like making them whatever the stack happens to have, or...? Does your compiler have flags for this to try? 2) Are you reading the numbers in with a single-precision or double- precision format field (%f vs. %lf in C-land)? I recall machines having issues with low-order precision/noise if one read a double with a float format field. You might be able to test this by making all of the special code float (real *4) and see whether similar machine/ compiler differences appear? 3) Can you do noise elimination on the input? Something like multiplying them by 10**5 or 10**6, taking an int and then dividing them back? If you're getting 7th digit noise from the I/O library this might reduce it. 4) Do you see the same behavior/results with -O0 and various optimization levels? You probably tried this, but that depends on whether you consider -O0 an optimization level or "no optimization"... 5) Can you randomize the order of operations to see whether you get different behavior, or sort the numbers in increasing magnitude to check (or reduce) the effect of precision? Can you change the G09 log to dump exponential rather than float (ie do you have the source to recompile)? 6) Do any of the results below match the expected answer? I'm sure others have more/better ideas... Joe jleonard42-.-gmail.com > > All copies of source are identical and compiled for real*8 (double > precision for 32 bits), and all answers are independent of > optimization level. The gfortran is the latest available download > version. ( I chose it because it is available across several OS's) > The g77 came with RHEL. > > The calculation involved reading output from a G03 log file where > Mulliken populations were computed from selected printed open shell > MO's and printed AO overlaps. [log file numbers had 5-6 significant > figuresafter decimal] Again, all calculations done in the special > program were done in real*8, and results are independent of > optimization level. > Compiler Answer Hardware, OS > 1) Linux Intel compiler (ifort) ~16 AMD64 opteron > RHEL4 (~=Centos 4) > 2) Linux g77 compiler ~10 " > 3) Windows Digital Corp compiler ~16 Intel P4 Windows2000 > 4) Windows gfortran compiler ~10 " > 5) Linux gfortran " ~10 Xeon > Centos 5.3 (~= RHEL5.3) > > Thoughts? > > Thanks! > > John > > -- > John McKelvey > 10819 Middleford Pl > Ft Wayne, IN 46818 > 260-489-2160 > jmmckel%a%gmail.com > > > > -- > John McKelvey > 10819 Middleford Pl > Ft Wayne, IN 46818 > 260-489-2160 > jmmckel%a%gmail.com --Apple-Mail-1--532829046 Content-Type: text/html; charset=US-ASCII Content-Transfer-Encoding: quoted-printable
On Oct 30, 2009, = at 6:04 PM, John McKelvey jmmckel^^^gmail.com wrote:

Folks,

I have a dilemma: Significantly results = depending on the compiler.

Couple of = things come to mind:

1) How does the compiler = treat uninitialized variables?  Is it vax-like, making them default = to 0 or C-like making them whatever the stack happens to have, or...? =  Does your compiler have flags for this to = try?

2) Are you reading the numbers in with a = single-precision or double-precision format field (%f vs. %lf in = C-land)? I recall machines having issues with low-order precision/noise = if one read a double with a float format field.  You might be able = to test this by making all of the special code float (real *4) and see = whether similar machine/compiler differences = appear?

3) Can you do noise elimination on the = input?  Something like multiplying them by 10**5 or 10**6, taking = an int and then dividing them back?  If you're getting 7th digit = noise from the I/O library this might reduce = it.

4) Do you see the same behavior/results = with -O0 and various optimization levels?  You probably tried this, = but that depends on whether you consider -O0 an optimization level or = "no optimization"...

5) Can you randomize the = order of operations to see whether you get different behavior, or sort = the numbers in increasing magnitude to check (or reduce) the effect of = precision?  Can you change the G09 log to dump exponential rather = than float (ie do you have the source to = recompile)?

6) Do any of the results below = match the expected answer?

I'm sure others have = more/better ideas...

Joe

All copies of source are identical and = compiled for real*8 (double precision for 32 bits), and all answers are = independent of optimization level.  The gfortran is the latest = available download version. ( I chose it because it is available across = several OS's)  The g77 came with RHEL. 

The = calculation involved reading output from a G03 log file where Mulliken = populations were computed from selected printed open shell MO's and = printed AO overlaps. [log file numbers had 5-6 significant figuresafter = decimal]  Again, all calculations done in the special program were = done in real*8, and results are independent of optimization level.  =
           = Compiler =             &n= bsp;        = Answer      Hardware, OS
1) Linux Intel = compiler (ifort)            = ~16   AMD64 opteron  RHEL4 (~=3DCentos 4)
2) Linux = g77 = compiler           =           ~10 =             = "
3) Windows Digital Corp compiler    = ~16    Intel P4  Windows2000
4) Windows = gfortran  compiler         ~10 =       "     
5) = Linux      = gfortran        = "            &= nbsp;  ~10      = Xeon      Centos 5.3 (~=3D RHEL5.3)
=
Thoughts?

Thanks!

John

--
John McKelvey
10819 Middleford = Pl
Ft Wayne, IN 46818
260-489-2160
jmmckel%a%gmail.com



--
John McKelvey
10819 Middleford Pl
Ft = Wayne, IN 46818
260-489-2160
jmmckel%a%gmail.com

= --Apple-Mail-1--532829046-- From owner-chemistry@ccl.net Sat Oct 31 11:42:00 2009 From: "John McKelvey jmmckel===gmail.com" To: CCL Subject: CCL:G: Further observations... on different results for different compilers Message-Id: <-40582-091031113145-29267-r3qUKURJQRlj1qccjdqLvA^server.ccl.net> X-Original-From: John McKelvey Content-Type: multipart/alternative; boundary=0016e647060e81c0c904773cd6d4 Date: Sat, 31 Oct 2009 11:31:27 -0400 MIME-Version: 1.0 Sent to CCL by: John McKelvey [jmmckel^^^gmail.com] --0016e647060e81c0c904773cd6d4 Content-Type: text/plain; charset=ISO-8859-1 Thanks to all for their comments so far. I also was notified by a major quantum chemistry software developer that they had similar results and said that while the Intel compiler had better optimization than gnu, some of the resulting code was wrong, and that now the gnu compiler is chosen over the Intel one. I have not tried optimization levels other than O0 and O3; haven't tried no-optimization as such (just assumed, correctly or other wise, that O0 and no-optimize would be the same, which is definitely true for Intel according to the ifort documentation) A further comment in answer to several questions. First, the G98 (not G03) log file was sent to me by someone else, i.e. not generated by me. Second, the program I used extracted the overlap matrix and selected MO's, and thjen computed the Mulliken pupulations from these MOs in the standard way. Third, the overlap matrix is printed in the format such as 1.234567D(+/-) 89 and the MO' such as as 1.23456. Fourth, for Intel's ifort the O0 and O3 results are the same, ~16, while for gnu's gfortran O0 and O3 the results are also the same, ~10. Fifth, for Intel's -O0 and -O3 gave the sane results, ~16, as well as -O3 with all calcs done in real*4 Sixth, gfortran -O3 gave the same results for -O0, -O3 and all calcs done in real*4, ~10 John -- John McKelvey 10819 Middleford Pl Ft Wayne, IN 46818 260-489-2160 jmmckel!A!gmail.com --0016e647060e81c0c904773cd6d4 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable
Thanks to all for their comments so far. I also was notified by a major quantum chemistry software developer that they had similar results and said that while the Intel compiler had better optimization than gnu, some of the resulting code was wrong, and that now the gnu compiler is chosen over the Intel one.=A0 I have not tried optimization levels other than O0 and O3; haven't tried no-optimization as such (just assumed, correctly or other wise, that O0 and no-optimize would be the same, which is definitely true for Intel according to the ifort documentation)

A further comment in answer to several questions.=A0

First, the= G98 (not G03) log file was sent to me by someone else, i.e. not generated = by me.=A0

Second, the program I used extracted the overlap matrix and selected MO's, and thjen computed the Mulliken pupulations from these MOs in the standard way.

Third, the overlap matrix is printed in the format such as 1.234567D(+/= -) 89 and the MO' such as as 1.23456.

Fourth, for Intel's ifort the O0 and O3 results are the same, ~16, while for gnu's gfortran O0 and O3 the results are also the same, ~10.

Fif= th, for Intel's -O0 and -O3 gave the sane results, ~16, as well as -O3 = with all calcs done in real*4

Sixth, gfortran -O3 gave the same resu= lts for -O0, -O3 and all calcs done in real*4, ~10

John

--
John McKelvey
10819 Middleford PlFt Wayne, IN 46818
260-489-2160
jmmckel!A!gmail.com
--0016e647060e81c0c904773cd6d4-- From owner-chemistry@ccl.net Sat Oct 31 12:17:00 2009 From: "Sean Zhu batrachian*gmail.com" To: CCL Subject: CCL:G: Ti Charge Message-Id: <-40583-091028165044-30845-LXcERUtk7XbyZ617YxAkWA ~ server.ccl.net> X-Original-From: "Sean Zhu" Date: Wed, 28 Oct 2009 16:50:41 -0400 Sent to CCL by: "Sean Zhu" [batrachian/./gmail.com] I was simulating a system of one T3Cl3 sit on top of MgCl2 cluster using Gaussian03. In the input file, the charge and mutiplicity were specified as 0 and 2. The optimization converged, however the NBO charge on Ti was shown to be 0.756. Was there anything wrong with the calculation. Thanks From owner-chemistry@ccl.net Sat Oct 31 13:00:01 2009 From: "Sudarshan Wadkar wadkar^_^gmail.com" To: CCL Subject: CCL: ADF and SGE(Grid Engine) : How To ? Message-Id: <-40584-091030173427-21728-dNE3NGwe0JO6veqr1QVh3A*server.ccl.net> X-Original-From: Sudarshan Wadkar Content-Type: multipart/alternative; boundary=001636e0b9c65418a804772d5c82 Date: Sat, 31 Oct 2009 02:33:13 +0530 MIME-Version: 1.0 Sent to CCL by: Sudarshan Wadkar [wadkar-,-gmail.com] --001636e0b9c65418a804772d5c82 Content-Type: text/plain; charset=ISO-8859-1 Dear CCL subscribers, I have few queries regarding Grid Engine's queue system and ADF (2008.01e). First a little background : I have a 4 node mini-cluster(4 x Core 2 Quad) using Rocks 5.1/CentOS 5, where I have managed to install and run ADF2008.01e in parallel. The important file for parallel execution, ".machine" or the "machinefile" is working properly. It's contents are -- compute-0-0 4 compute-0-1 4 compute-0-2 4 I am confused about 3 things 1. ADF uses its own HP-MPI library ($ADFHOME/bin/hpmpi) how do I configure SGE so that it uses that HP-MPI library and not the system MPI ? 2. When I run ADF in interactive mode from ADF-GUI, I know I can change the .machinefile to control behaviour of ADF in a prallel run. How do I manage that using SGE? To put it in another way, how do I use SGE to determine which nodes are free and ADF should start as and when it gets, say 8 processors (i.e. 2 nodes) free? How to tell ADF to run only on those specific nodes? 3. How to create a parallel run shell script (which would interact with "qsub" command) from an ADf ".job" file? Thank you :) -Surasjajwej ~$udhi "Success is getting what you want. Happiness is wanting what you get." - Dale Carnegie "There is only one success - to be able to spend your life in your own way." - Christopher Morley --001636e0b9c65418a804772d5c82 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Dear CCL subscribers,
I have few queries regarding Grid Engine's que= ue system and ADF (2008.01e).
First a little background :
I have a 4= node mini-cluster(4 x Core 2 Quad) using Rocks 5.1/CentOS 5, where I have = managed to install and run ADF2008.01e in parallel. The important file for = parallel execution, ".machine" or the "machinefile" is = working properly. It's contents are --

compute-0-0 4
compute-0-1 4
compute-0-2 4

I = am confused about 3 things
1. ADF uses its own HP-MPI library ($ADFHOME/= bin/hpmpi) how do I configure SGE so that it uses that HP-MPI library and n= ot the system MPI ?
2. When I run ADF in interactive mode from ADF-GUI, I know I can change the= .machinefile to control behaviour of ADF in a prallel run. How do I manage= that using SGE? To put it in another way, how do I use SGE to determine wh= ich nodes are free and ADF should start as and when it gets, say 8 processo= rs (i.e. 2 nodes) free? How to tell ADF to run only on those specific nodes= ?
3. How to create a parallel run shell script (which would interact with &qu= ot;qsub" command) from an ADf ".job" file?

Thank you = :)


-Surasjajwej
~$udhi
"Success is getting what you w= ant. Happiness is wanting what you get."
- Dale Carnegie
"There is only one success - to be able to spend y= our life in your own way."
- Christopher Morley
--001636e0b9c65418a804772d5c82-- From owner-chemistry@ccl.net Sat Oct 31 13:26:01 2009 From: "John McKelvey jmmckel-,-gmail.com" To: CCL Subject: CCL: Intel with default compiler options are incorrect Message-Id: <-40585-091031120839-12328-1J0ATu9zCQCMGWp+EdaQ+g-.-server.ccl.net> X-Original-From: John McKelvey Content-Type: multipart/alternative; boundary=001636c5c165c9a8ad04773d5a4d Date: Sat, 31 Oct 2009 12:08:26 -0400 MIME-Version: 1.0 Sent to CCL by: John McKelvey [jmmckel||gmail.com] --001636c5c165c9a8ad04773d5a4d Content-Type: text/plain; charset=ISO-8859-1 Went to Google with the entry "ifort" Second result was about different results from gfortran and ifort: it's "heap" vs "stack" and large arrays. gnu uses heap and ifort uses stack. Suggested solution there was to use the ifort following compiler options ifort -g -traceback -heap-arrays -mcmodel=medium Then ifort agreed with gfortran John -- John McKelvey 10819 Middleford Pl Ft Wayne, IN 46818 260-489-2160 jmmckel++gmail.com --001636c5c165c9a8ad04773d5a4d Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Went to Google with the entry "ifort"

Second result was ab= out different results from gfortran and ifort: it's "heap" vs= "stack" and large arrays.=A0 gnu uses heap and ifort uses stack.= =A0 Suggested solution there was to use the ifort following compiler option= s

ifort -g -traceback -heap-arrays -mcmodel=3Dmedium

Then ifort ag= reed with gfortran

John

--
John McKelvey10819 Middleford Pl
Ft Wayne, IN 46818
260-489-2160
jmmckel++gmail.com
--001636c5c165c9a8ad04773d5a4d--