Re: [Gfs-users] Trouble with boxes entirely within a solid in MPI.
Brought to you by:
popinet
Menu
▾
▴
Re: [Gfs-users] Trouble with boxes entirely within a solid in MPI.
|
From: Stephane Z. (lmm) <za...@lm...> - 2012-01-24 16:21:15
|
HI I edited my .darcs/author file. I look forward to the new version of the patch. cheers to all Stéphane Zaleski Le 22 janv. 2012 à 03:07, Stephane Popinet a écrit : > Hi Stephane, > >> A partial fix to this, which in particular makes the example below work fine, is in the >> attached patch. > > Thanks for the patch. I have created a new version of it: > > Sun Jan 22 15:01:42 NZDT 2012 Stephane Zaleski <ste...@up...> > * Do not call boundary_tree on boundaries with no root cell > > with a slightly different formatting. Could you please also edit your > darcs author address (in ~/.darcs/author) and set it to: > > Stephane Zaleski <ste...@up...> > > to have a consistent formatting. > > cheers > > Stephane > > > > >> >> Best wishes >> >> Stéphane Zaleski >> >> >> Le 29 déc. 2011 à 16:16, Stephane Zaleski (gmailc) a écrit : >> >>> Hi all >>> >>> we think it may be useful to report a troublesome behavior. >>> >>> Impossibility of having a boundary with no communication. >>> >>> In this example box id 1, pid 0 has an mpi boundary with box id 2, pid 1 >>> but box id 1 is "empty" because it is entirely within the solid. We suppose that attempts to initiate communications >>> that are non-existent results in the error. >>> >>> if the example is run in serial, it works. >>> if the solid is moved so that box id 1 is not wholly in the solid >>> any more, it works >>> if box id 2, which currently is attributed to pid 1, is instead >>> attributed to pid 0, the example works again. >>> >>> We find this error on several systems. For example, on >>> MacOS X 10.7.2 >>> gerris: using 2D libgfs version 1.3.2 (111222-170027) >>> compiled with flags: -DBSD_SOURCE -D_DARWIN_C_SOURCE -D_DARWIN_C_SOURCE >>> MPI: yes >>> pkg-config: yes >>> m4: yes >>> >>> % mpirun -np 4 gerris2D -m bug1.gfs >>> *** Process received signal *** >>> Signal: Segmentation fault: 11 (11) >>> Signal code: Address not mapped (1) >>> Failing at address: 0x18 >>> [ 0] 2 libsystem_c.dylib 0x00007fff9345ecfa _sigtramp + 26 >>> [ 1] 3 ??? 0x0000000000000028 0x0 + 40 >>> [ 2] 4 libgfs2D-1.3.2.dylib 0x0000000101c4be46 periodic_match + 134 >>> [ 3] 5 libgfs2D-1.3.2.dylib 0x0000000101c62d07 box_match + 231 >>> [ 4] 6 libgfs2D-1.3.2.dylib 0x0000000101c63a76 domain_foreach + 278 >>> [ 5] 7 libgfs2D-1.3.2.dylib 0x0000000101c55f7b domain_match + 171 >>> [ 6] 8 libgfs2D-1.3.2.dylib 0x0000000101c55e8b gfs_domain_match + 107 >>> [ 7] 9 libgfs2D-1.3.2.dylib 0x0000000101c6f6fc gfs_simulation_refine + 604 >>> [ 8] 10 libgfs2D-1.3.2.dylib 0x0000000101c7599d simulation_run + 557 >>> [ 9] 11 libgfs2D-1.3.2.dylib 0x0000000101c71580 gfs_simulation_run + 176 >>> [10] 12 gerris2D 0x0000000101c07c13 main + 7331 >>> [11] 13 gerris2D 0x0000000101c05f64 start + 52 >>> *** End of error message *** >>> -------------------------------------------------------------------------- >>> mpirun noticed that process rank 1 with PID 8283 on node MacBook-Pro-de-Stephane.local exited on signal 11 (Segmentation fault: 11). >>> -------------------------------------------------------------------------- >>> >>> It is easy to work around the bug, for instance by changing the pid of the neighboring boxes. >>> >>> Nevertheless, we are trying to find a more elegant solution. >>> >>> Hope it helps >>> >>> Stéphane Zaleski, with Pascal Ray, Daniel Fuster, Arup Das, Gilles Agbaglah and Bertrand Lagrée >>> >>> <bug1.gfs> >>> >>> >>> >>> ------------------------------------------------------------------------------ >>> Ridiculously easy VDI. With Citrix VDI-in-a-Box, you don't need a complex >>> infrastructure or vast IT resources to deliver seamless, secure access to >>> virtual desktops. With this all-in-one solution, easily deploy virtual >>> desktops for less than the cost of PCs and save 60% on VDI infrastructure >>> costs. Try it free! http://p.sf.net/sfu/Citrix-VDIinabox_______________________________________________ >>> Gfs-users mailing list >>> Gfs...@li... >>> https://lists.sourceforge.net/lists/listinfo/gfs-users >> >> >> ------------------------------------------------------------------------------ >> Write once. Port to many. >> Get the SDK and tools to simplify cross-platform app development. Create >> new or port existing apps to sell to consumers worldwide. Explore the >> Intel AppUpSM program developer opportunity. appdeveloper.intel.com/join >> http://p.sf.net/sfu/intel-appdev >> _______________________________________________ >> Gfs-users mailing list >> Gfs...@li... >> https://lists.sourceforge.net/lists/listinfo/gfs-users >> > > ------------------------------------------------------------------------------ > Try before you buy = See our experts in action! > The most comprehensive online learning library for Microsoft developers > is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3, > Metro Style Apps, more. Free future releases when you subscribe now! > http://p.sf.net/sfu/learndevnow-dev2 > _______________________________________________ > Gfs-users mailing list > Gfs...@li... > https://lists.sourceforge.net/lists/listinfo/gfs-users |