Screenshot instructions:
Windows
Mac
Red Hat Linux
Ubuntu
Click URL instructions:
Rightclick on ad, choose "Copy Link", then paste here →
(This may not be possible with some types of ads)
From: Vladimir Kolobov <vik@cf...>  20120521 01:43:57
Attachments:
Message as HTML

Stephane, Thanks a lot ! Your patch has fixed this problem. Best Regards, Vladimir On 5/20/2012 4:44 PM, Stephane Popinet wrote: > Hi Vladimir, > > Thanks for the bug report. Please confirm that the attached patch > fixes the problem. The LocateArray structure was not updated when > boxes entirely within a solid were destroyed. > > cheers > > Stephane > > >  > Live Security Virtual Conference > Exclusive live event will cover all the ways today's security and > threat landscape has changed and how IT managers can respond. Discussions > will include endpoint security, mobile security and the latest in malware > threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ > > > _______________________________________________ > Gfsdevel mailing list > Gfsdevel@... > https://lists.sourceforge.net/lists/listinfo/gfsdevel 
From: Vladimir Kolobov <vik@cf...>  20120519 14:16:11
Attachments:
Message as HTML

Stephane and Daniel, The following problem may also be related to the recent problems we observed with VOF/solids. When running the attached step.gfs file, we are getting gts_object_is_from_class: assertion `c != NULL' failed and then gerris segfaults. It happens when trying to locate a target inside the solid using OutputLocation { istep = 1 } stdout { 1.531250e+00 2.812500e01 0.} (Same problem exists when using a gts file). Thanks you, Vladimir step.gfs 12 17 GfsSimulation GfsBox GfsGEdge {} { # Stop the simulation at t = 15 Time { end = 15 } # Use an initial refinement of 6 levels (i.e. 2^6=64x64 for each box) Refine 4 RefineSolid 4 Solid (x> 1.499&& y< 0.505 ? 1 : 1 ) # Add a passive tracer called T VariableTracer {} T # Set the initial xcomponent of the velocity to 1 Init {} { U = 1 } # Adapt the mesh using the vorticity criterion at every timestep # down to a maximum level of 6 and with a maximum tolerance of 1e2 #AdaptVorticity { istep = 1 } { maxlevel = 6 cmax = 1e2 } # Adapt the mesh using the gradient criterion on variable T at # every timestep, down to a maximum level of 6 and with a maximum tolerance of 1e2 #AdaptGradient { istep = 1 } { maxlevel = 6 cmax = 1e2 } T # Set a viscosity source term on the velocity vector with xcomponent U # The Reynolds number is Re = D*U/Nu = 0.125*1/0.00078125 = 160 # where D is the cylinder diameter (as defined in cylinder.gts) SourceDiffusion {} U 0.00078125 SourceDiffusion {} V 0.00078125 # Writes the time and timestep every 10 timesteps on standard error OutputTime { istep = 10 } stderr # Writes the simulation size every 10 timesteps on standard error OutputBalance { istep = 10 } stderr # Writes info about the convergence of the Poisson solver on standard error OutputProjectionStats { istep = 10 } stderr # Outputs profiling information at the end of the simulation to standard error OutputTiming { start = end } stderr OutputSimulation { istep = 1 } step%06ld.gfs {} OutputLocation { istep = 1 } stdout { 1.531250e+00 2.812500e01 0.} } GfsBox { # Left boundary on the leftmost box is: # Dirichlet U=1 for the xcomponent of the velocity # Dirichlet T = 1 if y< 0, 0 otherwise left = Boundary { BcDirichlet U 1 BcDirichlet T { return y< 0. ? 1. : 0.; } } } Box {} Box {} Box {} Box {} Box {} Box {} Box {} Box {} Box {} Box {} Box {} 1 2 right 1 8 top 2 3 right 2 7 top 3 4 right 3 6 top 4 5 top 8 7 right 8 9 top 7 6 right 7 10 top 6 5 right 6 11 top 5 12 top 9 10 right 10 11 right 11 12 right 
From: Daniel Fuster <dfuster@gm...>  20120519 17:52:49
Attachments:
Message as HTML

Hi Vladimir The "gts_object_is_from_class: assertion `c != NULL' failed" comes from GFS_IS_BOX (b>data) in gfs_domain_locate I think that the problem is that indeed this function GSList * b = locate_array_locate (domain>array, &target); seems to obtain the box where this point should be without considering if the box has been deleted because of the presence of the solid. I do not know if Stephane can propose an easy fix for this. To me it looks like if locate_array_locate function would be incomplete (it does not return NULL when it should), but it is a little bit obscure for me what exactly locate_linear_index means.... best Daniel 2012/5/19 Vladimir Kolobov <vik@...> > Stephane and Daniel, > > The following problem may also be related to the recent problems we > observed with VOF/solids. > > When running the attached step.gfs file, we are getting > > gts_object_is_from_class: assertion `c != NULL' failed > > and then gerris segfaults. > > It happens when trying to locate a target inside the solid using > OutputLocation { istep = 1 } stdout { 1.531250e+00 2.812500e01 0.} > > (Same problem exists when using a gts file). > > Thanks you, > > Vladimir > > step.gfs > > 12 17 GfsSimulation GfsBox GfsGEdge {} { > > # Stop the simulation at t = 15 > Time { end = 15 } > > # Use an initial refinement of 6 levels (i.e. 2^6=64x64 for each box) > Refine 4 > RefineSolid 4 > > Solid (x > 1.499 && y < 0.505 ? 1 : 1 ) > > # Add a passive tracer called T > VariableTracer {} T > > # Set the initial xcomponent of the velocity to 1 > Init {} { U = 1 } > > # Adapt the mesh using the vorticity criterion at every timestep > # down to a maximum level of 6 and with a maximum tolerance of 1e2 > #AdaptVorticity { istep = 1 } { maxlevel = 6 cmax = 1e2 } > > # Adapt the mesh using the gradient criterion on variable T at > # every timestep, down to a maximum level of 6 and with a maximum tolerance of 1e2 > #AdaptGradient { istep = 1 } { maxlevel = 6 cmax = 1e2 } T > > # Set a viscosity source term on the velocity vector with xcomponent U > # The Reynolds number is Re = D*U/Nu = 0.125*1/0.00078125 = 160 > # where D is the cylinder diameter (as defined in cylinder.gts) > SourceDiffusion {} U 0.00078125 > SourceDiffusion {} V 0.00078125 > > # Writes the time and timestep every 10 timesteps on standard error > OutputTime { istep = 10 } stderr > > # Writes the simulation size every 10 timesteps on standard error > OutputBalance { istep = 10 } stderr > > # Writes info about the convergence of the Poisson solver on standard error > OutputProjectionStats { istep = 10 } stderr > > # Outputs profiling information at the end of the simulation to standard error > OutputTiming { start = end } stderr > > OutputSimulation { istep = 1 } step%06ld.gfs {} > OutputLocation { istep = 1 } stdout { 1.531250e+00 2.812500e01 0.} > } > > GfsBox { > # Left boundary on the leftmost box is: > # Dirichlet U=1 for the xcomponent of the velocity > # Dirichlet T = 1 if y < 0, 0 otherwise > left = Boundary { > BcDirichlet U 1 > BcDirichlet T { return y < 0. ? 1. : 0.; } > } > } > > Box {} > Box {} > Box {} > Box {} > Box {} > Box {} > Box {} > Box {} > Box {} > Box {} > Box {} > > 1 2 right > 1 8 top > 2 3 right > 2 7 top > 3 4 right > 3 6 top > 4 5 top > 8 7 right > 8 9 top > 7 6 right > 7 10 top > 6 5 right > 6 11 top > 5 12 top > 9 10 right > 10 11 right > 11 12 right > > > >  > Live Security Virtual Conference > Exclusive live event will cover all the ways today's security and > threat landscape has changed and how IT managers can respond. Discussions > will include endpoint security, mobile security and the latest in malware > threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ > _______________________________________________ > Gfsdevel mailing list > Gfsdevel@... > https://lists.sourceforge.net/lists/listinfo/gfsdevel > >  http://www.danielfuster.com 
From: Vladimir Kolobov <vik@cf...>  20120520 20:12:34
Attachments:
Message as HTML

Daniel, Thanks for your explanation. We remember that you already proposed a fix for a similar problem for MPI calculations. Here, gerris complains about NULL box and then segfaults: GfsCRITICAL **: gts_object_is_from_class: assertion `c != NULL' failed [maxwell03:06607] *** Process received signal *** [maxwell03:06607] Signal: Segmentation fault (11) [maxwell03:06607] Signal code: (128) [maxwell03:06607] Failing at address: (nil) [maxwell03:06607] [ 0] /lib64/libpthread.so.0 [0x38df80eb10] [maxwell03:06607] [ 1] /local/lib/libgfs2D1.3.so.2(gfs_domain_locate+0x60) [0x2b99ab64d410] Thanks, Vladimir On 5/19/2012 12:52 PM, Daniel Fuster wrote: > Hi Vladimir > > The "gts_object_is_from_class: assertion `c != NULL' failed" comes > from GFS_IS_BOX (b>data) in gfs_domain_locate > > I think that the problem is that indeed this function > GSList * b = locate_array_locate (domain>array, &target); > seems to obtain the box where this point should be without considering > if the box has been deleted because of the presence of the solid. > > I do not know if Stephane can propose an easy fix for this. To me it > looks like if locate_array_locate function would be incomplete (it > does not return NULL when it should), but it is a little bit obscure > for me what exactly locate_linear_index means.... > > best > Daniel > > > 2012/5/19 Vladimir Kolobov <vik@... <mailto:vik@...>> > > Stephane and Daniel, > > The following problem may also be related to the recent problems > we observed with VOF/solids. > > When running the attached step.gfs file, we are getting > > gts_object_is_from_class: assertion `c != NULL' failed > > and then gerris segfaults. > > It happens when trying to locate a target inside the solid using > OutputLocation { istep = 1 } stdout { 1.531250e+00 2.812500e01 0.} > > (Same problem exists when using a gts file). > > Thanks you, > > Vladimir > > step.gfs > > > 12 17 GfsSimulation GfsBox GfsGEdge {} { > > # Stop the simulation at t = 15 > Time { end = 15 } > > # Use an initial refinement of 6 levels (i.e. 2^^6 =64x64 for each box) > Refine 4 > RefineSolid 4 > > Solid (x> 1.499&& y< 0.505 ? 1 : 1 ) > > # Add a passive tracer called T > VariableTracer {} T > > # Set the initial xcomponent of the velocity to 1 > Init {} { U = 1 } > > # Adapt the mesh using the vorticity criterion at every timestep > # down to a maximum level of 6 and with a maximum tolerance of 1e2 > #AdaptVorticity { istep = 1 } { maxlevel = 6 cmax = 1e2 } > > # Adapt the mesh using the gradient criterion on variable T at > # every timestep, down to a maximum level of 6 and with a maximum tolerance of 1e2 > #AdaptGradient { istep = 1 } { maxlevel = 6 cmax = 1e2 } T > > # Set a viscosity source term on the velocity vector with xcomponent U > # The Reynolds number is Re = D*U/Nu = 0.125*1/0.00078125 = 160 > # where D is the cylinder diameter (as defined in cylinder.gts) > SourceDiffusion {} U 0.00078125 > SourceDiffusion {} V 0.00078125 > > # Writes the time and timestep every 10 timesteps on standard error > OutputTime { istep = 10 } stderr > > # Writes the simulation size every 10 timesteps on standard error > OutputBalance { istep = 10 } stderr > > # Writes info about the convergence of the Poisson solver on standard error > OutputProjectionStats { istep = 10 } stderr > > # Outputs profiling information at the end of the simulation to standard error > OutputTiming { start = end } stderr > > OutputSimulation { istep = 1 } step%06ld.gfs {} > OutputLocation { istep = 1 } stdout { 1.531250e+00 2.812500e01 0.} > } > > GfsBox { > # Left boundary on the leftmost box is: > # Dirichlet U=1 for the xcomponent of the velocity > # Dirichlet T = 1 if y< 0, 0 otherwise > left = Boundary { > BcDirichlet U 1 > BcDirichlet T { return y< 0. ? 1. : 0.; } > } > } > > Box {} > Box {} > Box {} > Box {} > Box {} > Box {} > Box {} > Box {} > Box {} > Box {} > Box {} > > 1 2 right > 1 8 top > 2 3 right > 2 7 top > 3 4 right > 3 6 top > 4 5 top > 8 7 right > 8 9 top > 7 6 right > 7 10 top > 6 5 right > 6 11 top > 5 12 top > 9 10 right > 10 11 right > 11 12 right > > >  > Live Security Virtual Conference > Exclusive live event will cover all the ways today's security and > threat landscape has changed and how IT managers can respond. > Discussions > will include endpoint security, mobile security and the latest in > malware > threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ > _______________________________________________ > Gfsdevel mailing list > Gfsdevel@... > <mailto:Gfsdevel@...> > https://lists.sourceforge.net/lists/listinfo/gfsdevel > > > > >  > http://www.danielfuster.com <http://www.danielfuster.com>; > > >  > Live Security Virtual Conference > Exclusive live event will cover all the ways today's security and > threat landscape has changed and how IT managers can respond. Discussions > will include endpoint security, mobile security and the latest in malware > threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ > > > _______________________________________________ > Gfsdevel mailing list > Gfsdevel@... > https://lists.sourceforge.net/lists/listinfo/gfsdevel 
From: Stephane Popinet <s.popinet@gm...>  20120520 21:44:51
Attachments:
locate.patch

Hi Vladimir, Thanks for the bug report. Please confirm that the attached patch fixes the problem. The LocateArray structure was not updated when boxes entirely within a solid were destroyed. cheers Stephane 
From: Vladimir Kolobov <vik@cf...>  20120521 01:43:57
Attachments:
Message as HTML

Stephane, Thanks a lot ! Your patch has fixed this problem. Best Regards, Vladimir On 5/20/2012 4:44 PM, Stephane Popinet wrote: > Hi Vladimir, > > Thanks for the bug report. Please confirm that the attached patch > fixes the problem. The LocateArray structure was not updated when > boxes entirely within a solid were destroyed. > > cheers > > Stephane > > >  > Live Security Virtual Conference > Exclusive live event will cover all the ways today's security and > threat landscape has changed and how IT managers can respond. Discussions > will include endpoint security, mobile security and the latest in malware > threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ > > > _______________________________________________ > Gfsdevel mailing list > Gfsdevel@... > https://lists.sourceforge.net/lists/listinfo/gfsdevel 
Sign up for the SourceForge newsletter:
No, thanks