September 2005 - Ifeffit - millenia.cars.aps.anl.gov

Tip of the Week: Features of the Guess, Def, Set page in Artemis
by Bruce Ravel 16 Sep '05

16 Sep '05

Hi, This week's tip of the week is "Features of the Guess, Def, Set page in Artemis" http://cars9.uchicago.edu/iffwiki/HoraeTipOfTheWeek Enjoy, B -- Bruce Ravel ---------------------------------------------- bravel(a)anl.gov Molecular Environmental Science Group, Building 203, Room E-165 MRCAT, Sector 10, Advance Photon Source, Building 433, Room B007 Argonne National Laboratory phone and voice mail: (1) 630 252 5033 Argonne IL 60439, USA fax: (1) 630 252 9793 My homepage: http://cars9.uchicago.edu/~ravel EXAFS software: http://cars9.uchicago.edu/~ravel/software/exafs/

1 0

compiling Ifeffit and friends on a 64 bit linux machine
by Bruce Ravel 15 Sep '05

15 Sep '05

Hi, Some months ago -- see the thread that starts here: http://millenia.cars.aps.anl.gov/pipermail/ifeffit/2004-December/001550.html Jake from U. Calgary posted a message about trouble getting Ifeffit et al to compile on his 64 bit machine. Over the last few days I have been upgrading to SUSE 9.3 on my home computer, which has an Intel 64 bit processor. (2 of them actually, lucky me!) 9.3 is much better for a 64 bit machine than 9.2 was. (It's also the first OpenSUSE release, which is a development I am very happy about.) Initially I ran into the exact same problem that Jake described in his original post. At the time, Carlo suggested that the -fPIC flag ("Position Independent Code", whatever that is) would be important. That turned out to be the trick. When installing Ifeffit, it was necessary to do the configure step like this: ./configure CC="cc -fPIC" FFLAGS="-g -O2 -fPIC" After that, the rest of installing Ifeffit and all of installing the horae tarball proceded normally. BUT, there is one other caveat. I had to make a small change to the pgplot makefile to get it compiled for my 64 bit machine. Everyplace in the makefile where it said "/usr/X11R6/lib", I had to change that to "/usr/X11R6/lib64". Compiling pgplot requires a bit of voodoo. I have always relied on Matt's PGPLOT_install script, which comes with the ifeffit tarball. In this case, I ran the script. It did some stuff, the failed. I cd-ed to /usr/local/pgplot, which is where it gets installed, and edited the makefile in that directory as I indicated in the last paragraph. Then in that directory I did a "make" and the pgplot compilation went to completion. Of course, I had to be root run the PGPLOT_install script and to do the compilation in /usr/local/pgplot. Hope that helps! (I'd be interested to hear how it goes with your opteron, should you try this, Carlo.) Bruce -- Bruce Ravel ----------------------------------- bravel(a)anl.gov -or- ravel(a)phys.washington.edu Environmental Research Division, Building 203, Room E-165 Argonne National Laboratory phone and voice mail: (1) 630 252 5033 Argonne IL 60439, USA fax: (1) 630 252 9793 My homepage: http://cars9.uchicago.edu/~ravel EXAFS software: http://cars9.uchicago.edu/~ravel/software/exafs/

2 2

Bug report for Artemis 0.8.000
by schmithals＠hmi.de 12 Sep '05

12 Sep '05

Hello Bruce, After a break I just started to use Artemis again and I am delighted about the new features, especially the ability to store and compare subsequent fits in one file. Below I want to submit the bugs that I came accross so far (on Win2000): - If one opens by mistake an input file that is unknown to Artemis (or Athena) this file will be destroyed and cut to about 4kB. Workaround: Just DON'T do this (or keep regular backups). - Funny things happen when one studies an atoms/feff calculation without having loaded an experiment file: I imported an atoms.inp, ran atoms and feff and did two summations (Sum | All included paths ..). For the first summation I included only the first path, for the second summation only the second path. The first summation seems fine, for the second summation Artemis obviously includes both paths. In subsequent summations this effect will continue: all paths used in the past are considered plus the paths currently included. Workaround: Load a 'dummy'-chi file even for just making summations. I was teaching the use of Artemis to a few colleagues. I recommend them to keep a copy of their atoms.inp-files at a separate place for re-use in later projects. If time permits maybe you could account for these issues: - If saving an atoms.inp file via Theory | Write special output | atoms the name 'artemis.stuff' will be provided. Is it possible to change that to 'atoms.inp'? - The link to TKAtoms is missing for newly installed systems. Cheers, Gerrit Schmithals

2 1

new tip of the week
by Bruce Ravel 09 Sep '05

09 Sep '05

Hi! I just posted this week's Tip of the Week at http://cars9.uchicago.edu/iffwiki/HoraeTipOfTheWeek This weeks topic: Easily marking arbitrary groups in Athena Enjoy, B -- Bruce Ravel ----------------------------------- bravel(a)anl.gov -or- ravel(a)phys.washington.edu Environmental Research Division, Building 203, Room E-165 Argonne National Laboratory phone and voice mail: (1) 630 252 5033 Argonne IL 60439, USA fax: (1) 630 252 9793 My homepage: http://cars9.uchicago.edu/~ravel EXAFS software: http://cars9.uchicago.edu/~ravel/software/exafs/

1 0

Re: [Ifeffit] Re: ifeffit installation on Linux
by Matt Newville 09 Sep '05

09 Sep '05

Hi Peter, > Here's the output: > > src/cmdline/Makefile:TERMCAP_LIB = -ltermcap > src/cmdline/Makefile:# readline_LIB = -L$(top_srcdir)/readline > -lreadline $(TERMCAP_LIB) > src/cmdline/Makefile:readline_LIB = > $(top_srcdir)/readline/libreadline.a $(TERMCAP_LIB) > readline/Makefile:TERMCAP_LIB = -ltermcap > readline/Makefile: $(CC) $(CCFLAGS) -o $@ ./examples/rl.c > ../libreadline.a ${TERMCAP_LIB} > # locate termcap: > /usr/lib/libtermcap.so.2 > /usr/lib/libtermcap.so.2.0.8 > /usr/lib/termcap/libtermcap.a > /usr/lib/termcap/libtermcap.so > # locate curses: > /usr/lib/curses/libcurses.a > /usr/lib/curses/libcurses.so > /usr/lib/libcurses.so.1 > /usr/lib/libcurses.so.1.0.0 > Thanks for the report. That looks like the change to configure did what it was supposed to do: instead of having './lib/termcap/libtermcap.a', it now has '-ltermcap'. The odd things to me are that a) libtermcap.a seems to be in an odd location (/usr/lib/termcap), which makes me wonder how any readline applications work. b) you still end up with a './lib/termcap/libtermcap.a' somewhere in your Makefile, but *not* as TERMCAP_LIB=./lib/termcap/libtermcap.a I have no idea where that would be coming from. I don't think that an extra '-L./lib/termcap' is causing a problem (it's ok to have '-L' arguments that point nowhere) but that the linking can't find the right libtermcap.a with '-ltermcap' alone. If the Makefiles said TERMCAP_LIB= '-L/usr/lib/termcap -ltermcap' that would work for sure. I've been relying on a 'BASH_CHECK_LIB_TERMCAP' macro in autoconf, which apparantly doesn't know to look in /usr/lib/termcap. At this point, it may be easier to just write a function to look for the right termcap library. I think readline can use termcap, curses, or ncurses, and to be honest, I find the differences confusing. Anyway, thanks for the help! --Matt

1 0

ifeffit installation on Linux
by marco antoni alsina corvalan 08 Sep '05

08 Sep '05

Hi everyone, I'm currently making my way into Linux, so I decided to change most of my programs to this plattform, which obviously include ifeffit ;) I downloaded the ".tar" file, ran with success the PGPlot script for the graphics library installation (at least the demos run!!), and executed the "configure" script, getting the message "ready for the next step"...however I also get the following info: .... checking termcap.h usability... no checking termcap.h presence... no checking for termcap.h... no .... checking for tgetent... no checking for tgetent in -ltermcap... no .... Anyway then I executed the "run" command, getting the following error: .... cc: ./lib/termcap/libtermcap.a: No such file or directory make[3]: *** [ifeffit] Error 1 make[3]: Leaving directory `/home/user/Programas/ifeffit-1.2.7/src/cmdline' make[2]: *** [all-recursive] Error 1 make[2]: Leaving directory `/home/user/Programas/ifeffit-1.2.7/src' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/home/user/Programas/ifeffit-1.2.7' make: *** [all-recursive-am] Error 2 If anyone can help me overcoming this impasse I would be very greatful. Also if you need more info please let me know Regards, Marco ______________ Marco A. Alsina Estudiante de Magíster Departamento de Ingeniería Hidráulica y Ambiental Pontificia Universidad Católica de Chile Vicuña Mackenna 4860, Santiago, Chile (56 2) 3544873

5 14

Questions and answers
by Bruce Ravel 06 Sep '05

06 Sep '05

Good morning, Every week I receive a large amount of email regarding my software and XAS in general. I am sure that's no surprise to anyone reading this. While I certainly want to encourage people to feel free to contact me about XAS, the software, and any other topic, the time I have to address the needs and concerns of all the people contacting me is limited. In recent weeks I have received an unusual number of emails containing vague, ill-posed questions or requests for inordinate amounts of help. The other day I was frustrated by one such correspondent, so today I am indulging myself by ranting in a public forum ;-) Consider this message that I received the other day: > I have a sample with lead adsorbed onto manganese oxide and I have > its EXAFS. I want to study the Pb-Mn shell (including r(Pb-Mn) and > the number of Mn atom around Pb atom). How should I proceed? > > Awaiting your reply. Contrast that with an email I received from a user who had some questions along with a bug report regarding how e0 is determined by Athena: > I'm attaching an Athena project with a sample labeled Libethenite > and a Cu standard, along with an step by step explanation included > in the project journal. I trust the difference between these two is clear. I do not know either of these individuals except by their emails. Both are from places far away from Chicago. The first received a 5-line response while I spent two hours on a plane flight two weeks ago addressing the second one. I encourage people to contact me at any time with any question about XAS or my software. Indeed, if Athena and Artemis are useful data analysis tools today it is only because of the kind and generous feedback I have received from its users. But if a genie were to grant me a wish, it would be that more folks were like the second correspondent than the first. Thanks, B -- Bruce Ravel ----------------------------------- bravel(a)anl.gov -or- ravel(a)phys.washington.edu Environmental Research Division, Building 203, Room E-165 Argonne National Laboratory phone and voice mail: (1) 630 252 5033 Argonne IL 60439, USA fax: (1) 630 252 9793 My homepage: http://feff.phys.washington.edu/~ravel EXAFS software: http://feff.phys.washington.edu/~ravel/software/exafs/

1 0

RE: [Ifeffit] Ifeffit Wiki
by Barton, David (DG) 06 Sep '05

06 Sep '05

Bruce, >It makes me kind of sad to think that some of these neat-o features > are not widely used. If it makes you feel any better, I just used your linear combination module in Athena to fit 165 XANES spectra. Dave > -----Original Message----- > From: ifeffit-bounces(a)millenia.cars.aps.anl.gov > [mailto:ifeffit-bounces@millenia.cars.aps.anl.gov] On Behalf > Of Bruce Ravel > Sent: Friday, September 02, 2005 4:10 PM > To: XAFS Analysis using Ifeffit > Subject: Re: [Ifeffit] Ifeffit Wiki > > > On Wednesday 31 August 2005 17:22, Matt Newville wrote: > > Hi Folks, > > > > I've created and started a wiki for Ifeffit and related topics. > > Currently, this is at > > http://cars9.uchicago.edu/iffwiki > > I, for one, think the new Wiki is quite cool. Yet another > thing that we have to thank Matt for. > > My first use of the Wiki is to start a "Tip of the Week" web > page. Athena and Artemis have become sufficiently robust and > feature-full that there are a lot of interesting aspects of > the codes that are not widely known about or used. It makes > me kind of sad to think that some of these neat-o features > are not widely used. To keep a smile on my face, I've > decided to use the wiki to highlight some cool tidbit every > Friday afternoon. The first one is up at > > http://cars9.uchicago.edu/iffwiki/HoraeTipOfTheWeek > > Check it out. Feel free to send me suggestions for features > of A&A that you think would make a good Tip of the Week. My > plan is more or less to alternate between Athena and Artemis > and occassionally to throw in a Hephaestus tip as well. > > B > > -- > Bruce Ravel ----------------------------------- bravel(a)anl.gov -or- > ravel(a)phys.washington.edu > Environmental Research Division, Building 203, Room E-165 > Argonne National Laboratory phone and voice mail: > (1) 630 252 5033 > Argonne IL 60439, USA fax: > (1) 630 252 9793 > > My homepage: http://feff.phys.washington.edu/~ravel > EXAFS software: > http://feff.phys.washington.edu/~ravel/software/exafs/ > > > _______________________________________________ > Ifeffit mailing list > Ifeffit(a)millenia.cars.aps.anl.gov > http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit >

4 6

RE: [Ifeffit] Ifeffit Wiki
by Barton, David (DG) 06 Sep '05

06 Sep '05

Bruce, > So would it be useful to have some more automation in the LCF > utility. Automation is always appreciated, not required, but much appreciated. The data set that I was fitting was indeed the result of multiple fast scans performed during an in-situ catalysis experiment, similar to what Anatoly is planning for NSLS. My biggest challenge with the current Athena code. 0.8.041 in fitting many XANES spectra was not having a single report where all fitting results could be exported to. I would imagine if you automated the fitting routine you would produce some report that lists all fitting results including fitting statistics in a single file. In addition, in the current fitting utility there was a few wasted clicks and edits because the fitting range kept reverting back to -15 to 45 eV... My fitting range was different than the default and every time I choose another data set to fit the fitting range reverted back to the default which meant I had to choose each box and make the appropriate changes for each new data set. Regardless of these issues, the LCF module is fantastic. I truly appreciate the ease of use, the ability to view the fit along with components of the fit, and the ability to plot the residual. Thanks for another useful tool to make my life easier. Regards, Dave > -----Original Message----- > From: ifeffit-bounces(a)millenia.cars.aps.anl.gov > [mailto:ifeffit-bounces@millenia.cars.aps.anl.gov] On Behalf > Of Bruce Ravel > Sent: Friday, September 02, 2005 6:02 PM > To: XAFS Analysis using Ifeffit > Subject: Re: [Ifeffit] Ifeffit Wiki > > > On Friday 02 September 2005 16:12, Barton, David (DG) wrote: > > >It makes me kind of sad to think that some of these neat-o > features > > >are not widely used. > > > > If it makes you feel any better, I just used your linear > combination > > module in Athena to fit 165 XANES spectra. > > Spectacularly so! Thanks for the big smiley at the end of the week. > > So would it be useful to have some more automation in the LCF > utility. What occured to me as I read this was to have a way > to fit a fixed set of standards to all the marked groups and > right a big report at the end. That is, play around with > spectrum #1 until you have decided which standards to use and > how to do the fit. Then mark spectra 2-165, press the > button, and go get a cup of coffee. That seems like a good > idea to me. > > B > > -- > Bruce Ravel ----------------------------------- bravel(a)anl.gov -or- > > ravel(a)phys.washington.edu > > Environmental Research Division, Building 203, Room E-165 > Argonne National Laboratory phone and voice mail: > (1) 630 252 5033 > Argonne IL 60439, USA fax: > (1) 630 252 9793 > > My homepage: http://feff.phys.washington.edu/~ravel > EXAFS software: > http://feff.phys.washington.edu/~ravel/software/exafs/ > > > > _______________________________________________ > Ifeffit mailing list > Ifeffit(a)millenia.cars.aps.anl.gov > http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit >

1 0

new source code release
by Bruce Ravel 04 Sep '05

04 Sep '05

Hi folks, I have released the latest source code tarball for my codes. Tarball 054 is now available on my web site and at Source Forge. I am continuing Matt's experiment providing an Ifeffit Wiki by posting the change log on my wiki page: http://cars9.uchicago.edu/iffwiki/BruceRavel Scroll to bottom of the page (after smiling at the photo of me and my adorable niece) to find the page documenting the new features and bug fixes in this release. Matt and I are getting close to having a completely new Windows installer prepared. The new package has a more sensible file structure, a much improved launching tool for the programs, and new executables of my codes made with PerlApp. This version of my codes will be in the first release of the new Windows installer. Hopefully that will be ready soon. Regards, B -- Bruce Ravel ----------------------------------- bravel(a)anl.gov -or- ravel(a)phys.washington.edu Environmental Research Division, Building 203, Room E-165 Argonne National Laboratory phone and voice mail: (1) 630 252 5033 Argonne IL 60439, USA fax: (1) 630 252 9793 My homepage: http://feff.phys.washington.edu/~ravel EXAFS software: http://feff.phys.washington.edu/~ravel/software/exafs/

1 0