<html>
<head>
<style><!--
.hmmessage P
{
margin:0px;
padding:0px
}
body.hmmessage
{
font-size: 10pt;
font-family:Verdana
}
--></style>
</head>
<body class='hmmessage'>
I've tried many tests, changing min max coords, CFL number, NXB vs NYB/NZB. Initial conditions (rho, temp, etc.) seem ok (using visit), but in the first timestep I always receive, with gdb (and -fpe0 enabled), this error in avisco.F90:<BR>
<BR>
Program received signal SIGFPE, Arithmetic exception.<BR>avisco (j=5, k=5, avis=..., dirgeom=..., xyzswp=2, nzni=5, nznf=13, nsdim=3, xtop=8.364632500000001e+23, xbot=-2.3767975000000004e+23,<BR> ytop=4.4465908125712189e-323, ybot=4.4465908125712189e-323, ylft=4.4465908125712189e-323, yrgt=0, zlft=-0.19634954084936207, zrgt=0.58904862254808621,<BR> x=..., xl=..., xzn=..., yzn=..., zzn=..., u=..., uttp=..., utbt=..., utrt=..., utlt=..., cvisc=0.10000000000000001) at avisco.F90:417<BR>417 avis(i) = (sin(x(i)) * u(i) - sin(x(i-1)) * u(i-1)) / &<BR><BR>
I've print:<BR>
(gdb) print dytb<BR>$2 = 7.9050503334599447e-323<BR>(gdb) print dxtb<BR>$3 = 2.5055296642745135e-72<BR>(gdb) print dztb<BR>$4 = 6.9533558064452965e-310<BR>
<BR>
I think around the polar region, coords differences (dytb, etc..) may go to zero (and so denom), but I still don't know how to avoid or fix this. As I said I tried changing min rad (and also max theta to 180 and phi to 360), but this error continues to appear with low or high refinement (40 blocks or 1000 blocks). <BR>
Hints?<BR>
<BR> <BR>> Date: Thu, 25 Mar 2010 14:41:26 -0500<BR>> From: klaus@flash.uchicago.edu<BR>> To: gaspmax@hotmail.com<BR>> CC: flash-users@flash.uchicago.edu<BR>> Subject: Re: [FLASH-USERS] xflash3 and 3D spherical coordinates<BR>> <BR>> On Thu, 25 Mar 2010, Massimo Gaspari wrote:<BR>> <BR>> > <BR>> > Hi all,<BR>> > I'm trying to plot with xflash3 a chk output created by <BR>> > a simulation in spherical coordinates (r, theta, phi). Does it work?<BR>> > (User's guide says that it is supported)<BR>> <BR>> Sorry, I've never tried this.<BR>> <BR>> It should be possible to trick xflash3 into ignoring the geometry<BR>> and treat the data as Cartesian (by making some simple change in<BR>> xflash3.pro), so you get at least *some* visualization with xflash3.<BR>> But again, I have never tried this for 3D spherical.<BR>> <BR>> <BR>> > Another question: have you tested 3D spherical coordinates with any hydro problem?<BR>> > Should I avoid them? <BR>> <BR>> Hydro with 3D spherical coordinates "should" work. However, we have not <BR>> tested it ourselves and are not using it. We would welcome any reports on<BR>> your experiences.<BR>> <BR>> I say it "should" work because 3D Cartesian as well as 1D and 2D spherical<BR>> are know to work, and the 3D spherical case does not add any essentially<BR>> new complications to the code.<BR>> <BR>> That was an answer from the code perspective. If it makes physical sense <BR>> to use spherical coordinates is another question. Also whether the <BR>> behavior at the poles and especially at the origin are represented well <BR>> enough in those coordinates (without excessive refinement) is open (and <BR>> should depend on your problem). 3D Cartesian may be preferable for such <BR>> reasons.<BR>> <BR>> Klaus<BR> <br /><hr />Chiacchiera con i tuoi amici via Webcam su Messenger. <a href='http://www.messenger.it/videoconversazioni.aspx' target='_new'>Videochiamali!</a></body>
</html>