BABYL OPTIONS: -*- rmail -*- Version: 5 Labels: Note: This is the header of an rmail file. Note: If you are seeing it in rmail, Note: it means the file has no messages in it.  1,, Mail-from: From pbsnyder@pppl.gov Sun Sep 12 18:18:34 1999 Received: from orion.pppl.gov (IDENT:pbsnyder@orion.pppl.gov [198.35.4.73]) by pppl.gov (8.9.2/8.9.2) with ESMTP id SAA04042; Sun, 12 Sep 1999 18:18:33 -0400 (EDT) Received: from localhost (pbsnyder@localhost) by orion.pppl.gov (8.9.2/8.9.2) with ESMTP id SAA29440; Sun, 12 Sep 1999 18:18:33 -0400 (EDT) X-Authentication-Warning: orion.pppl.gov: pbsnyder owned process doing -bs Date: Sun, 12 Sep 1999 18:18:32 -0400 (EDT) From: Phil Snyder To: Greg Hammett cc: mbeer@pppl.gov, bdorland@ipr.umd.edu Subject: Re: Phil's dissertation and code In-Reply-To: <199909112233.SAA02598@orion.pppl.gov> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Content-Length: 2916 *** EOOH *** X-Authentication-Warning: orion.pppl.gov: pbsnyder owned process doing -bs Date: Sun, 12 Sep 1999 18:18:32 -0400 (EDT) From: Phil Snyder To: Greg Hammett cc: mbeer@pppl.gov, bdorland@ipr.umd.edu Subject: Re: Phil's dissertation and code In-Reply-To: <199909112233.SAA02598@orion.pppl.gov> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Hi all, I added to this directory a sample linear input file which reproduces Fig 4.5 in my thesis (approximately, I lowered the resolution so it runs in a reasonable amount of time, the thesis uses ld=129, x0=xp=-3). The file is /work/pbsnyder/pgryffin/hhc1mod5.in (apologies for clunky naming). I also checked in a new itg.in file that includes the electromagnetic parameters with comments. The important things to note are that beta_e is half of the electron beta (and thus 1/4 of the total beta in the simple T_i=T_e single species cases). Also, nu_ei has a factor of the mass ratio in it and thus should always be a very small number. Numerical stability generally requires a much smaller timestep in the electromagnetic cases. Furthermore, in nonlinear cases, the courant routines do not properly reduce the timestep and maintain stability (despite some attempts to correct this problem). Thus, in nonlinear runs, the initial timestep must be set small enough to satisfy Courant throughout the entire run. If the timestep gets adjusted automatically the results are _garbage_. A very rough rule of thumb is to use a timestep which is a factor of 5-10 below its smallest value in an equivalent nonlinear electrostatic run. I would suggest for now sticking to T_i=T_e single species cases for EM runs. T_i .ne. T_e and multi-species have not been thoroughly tested. Also, please consult with me before publishing or presenting any EM runs, until I have time to fully comment the code, and give a more detailed account of things to be careful about. I intend to add more test cases and more comments to the code and to check in all of the recent changes after I get out to GA. We're leaving tomorrow. My home phone and address in SD as of 9/23 are: 7015 Charmant Dr., Apt #217 San Diego CA 92122 818 452-3050 -Phil On Sat, 11 Sep 1999, Greg Hammett wrote: > Phil's dissertation is now posted on his web site w3.ppl.gov/~pbsnyder. On > the pppl unix cluster, in /work/pbsnyder/pgryffin is a standard version of > the code Phil used towards the end of his thesis with his electromagnetic > extensions. This directory is shared with hecate.princeton.edu, and has > been compiled and run there. Phil says it usually compiles and runs easily > on the T3E as well, though the t3e doesn't support the ncar graphics so > postc doesn't do as much. There are a few short sample runs in pbstest. > There was a Ti/Te bug in earlier versions that is fixed in this version. > He thinks Ti/Te is now fixed, but it should be benchmarked against GS2 to > be sure. Phil will check this version into CVS soon, merging it with some > changes Bill had made to an earlier version of the electromagnetic code. > But for now, I thought it would be good to have a directory that contains a > standard version from the end of his thesis research (I've saved copies of > it in HPSS at NERSC also). > > Greg >