Re: [Paraview] XDMF Version 2 and 3 Differences
Oops I forgot, +3 xdmf3 has support for DMP parallel writes in master (will appear in PV 5.5.) David E DeMarle Kitware, Inc. Principal Engineer 21 Corporate Drive Clifton Park, NY 12065-8662 Phone: 518-881-4909 On Mon, Oct 2, 2017 at 1:58 PM, David E DeMarlewrote: > Hi Mike. > > You mean select the XDMF "Reader", not "Writer" when you open the file > correct? > > In theory XDMF3 should read almost anything XDMF2 could. The main > practical differences in my recollection are: > > + xdmf3 doesn't leak memory and is faster and less of a memory hog, > especially with time varying data, in my experience. > + xdmf3's field data arrays work better and in particular support > arbitrary width arrays > + Thanks to Michal Habera's Google Summer of Code project xdmf3 master (in > nightly binaries and will appear in PV 5.5) has support for "other" > centered (more than 1 per cell) data in support of FEniCS higher order > cells. > + xdmf3 has a much better suite of tests and contribution path than xdmf2 > ever did > > = xdmf3 had a boost dependency up until about a week ago (will appear in > PV 5.5). > > - xdmf3 tends to complain on files that use default/implied values in the > xml entries. > - xdmf3 doesn't automatically parallelize structured datasets when read in > parallel > - xdmf3 doesn't support strides for structured datasets > - xdfm3 doesn't support "hyperslab, list or range" temporal specifications > - xdmf3 lacks support for function specifications > - xdmf3's notion of subset is slightly different than xdmf2 (2's were > relative to cell, 3's are relative to face/edge or something to that effect) > > If you can share one of your files I'll take a look. > > thanks > > > > David E DeMarle > Kitware, Inc. > Principal Engineer > 21 Corporate Drive > Clifton Park, NY 12065-8662 > Phone: 518-881-4909 <(518)%20881-4909> > > On Thu, Sep 28, 2017 at 8:57 AM, Michael Jackson < > mike.jack...@bluequartz.net> wrote: > >> Where might I find the differences between XDMF version 2 and 3? Our >> project writes a .xdmf file along side our HDF5 file so that ParaView can >> visualize the data. Ever since ParaView version 5.x was released with the >> XDMF 3 included we always have to select the "XDMF Writer" when opening the >> file. If we select the either of the XDMF 3 writers no data is loaded. I >> was wanting to figure out if there is something we need to adjust in our >> XDMF files so that the XDMF 3 readers work correctly with our data/files. >> >> Thanks >> -- >> Michael Jackson | Owner, President >> BlueQuartz Software >> [e] mike.jack...@bluequartz.net >> [w] www.bluequartz.net >> >> >> ___ >> Powered by www.kitware.com >> >> Visit other Kitware open-source projects at >> http://www.kitware.com/opensource/opensource.html >> >> Please keep messages on-topic and check the ParaView Wiki at: >> http://paraview.org/Wiki/ParaView >> >> Search the list archives at: http://markmail.org/search/?q=ParaView >> >> Follow this link to subscribe/unsubscribe: >> http://public.kitware.com/mailman/listinfo/paraview >> > > ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the ParaView Wiki at: http://paraview.org/Wiki/ParaView Search the list archives at: http://markmail.org/search/?q=ParaView Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/paraview
Re: [Paraview] XDMF Version 2 and 3 Differences
Hi Mike. You mean select the XDMF "Reader", not "Writer" when you open the file correct? In theory XDMF3 should read almost anything XDMF2 could. The main practical differences in my recollection are: + xdmf3 doesn't leak memory and is faster and less of a memory hog, especially with time varying data, in my experience. + xdmf3's field data arrays work better and in particular support arbitrary width arrays + Thanks to Michal Habera's Google Summer of Code project xdmf3 master (in nightly binaries and will appear in PV 5.5) has support for "other" centered (more than 1 per cell) data in support of FEniCS higher order cells. + xdmf3 has a much better suite of tests and contribution path than xdmf2 ever did = xdmf3 had a boost dependency up until about a week ago (will appear in PV 5.5). - xdmf3 tends to complain on files that use default/implied values in the xml entries. - xdmf3 doesn't automatically parallelize structured datasets when read in parallel - xdmf3 doesn't support strides for structured datasets - xdfm3 doesn't support "hyperslab, list or range" temporal specifications - xdmf3 lacks support for function specifications - xdmf3's notion of subset is slightly different than xdmf2 (2's were relative to cell, 3's are relative to face/edge or something to that effect) If you can share one of your files I'll take a look. thanks David E DeMarle Kitware, Inc. Principal Engineer 21 Corporate Drive Clifton Park, NY 12065-8662 Phone: 518-881-4909 On Thu, Sep 28, 2017 at 8:57 AM, Michael Jackson < mike.jack...@bluequartz.net> wrote: > Where might I find the differences between XDMF version 2 and 3? Our > project writes a .xdmf file along side our HDF5 file so that ParaView can > visualize the data. Ever since ParaView version 5.x was released with the > XDMF 3 included we always have to select the "XDMF Writer" when opening the > file. If we select the either of the XDMF 3 writers no data is loaded. I > was wanting to figure out if there is something we need to adjust in our > XDMF files so that the XDMF 3 readers work correctly with our data/files. > > Thanks > -- > Michael Jackson | Owner, President > BlueQuartz Software > [e] mike.jack...@bluequartz.net > [w] www.bluequartz.net > > > ___ > Powered by www.kitware.com > > Visit other Kitware open-source projects at http://www.kitware.com/ > opensource/opensource.html > > Please keep messages on-topic and check the ParaView Wiki at: > http://paraview.org/Wiki/ParaView > > Search the list archives at: http://markmail.org/search/?q=ParaView > > Follow this link to subscribe/unsubscribe: > http://public.kitware.com/mailman/listinfo/paraview > ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the ParaView Wiki at: http://paraview.org/Wiki/ParaView Search the list archives at: http://markmail.org/search/?q=ParaView Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/paraview
Re: [Paraview] ParaViewWeb Visualizer black screen
Hi Louise, You are right, there is indeed more changes which might not be strait forward for you to fix. Even if in reality, it is only a bunch of import path to fix. Just out of curiosity, did you try to run the command line that the launcher is running by hand and see if you have the same behavior as well as maybe other error output? Otherwise, you right the screen size should be more than enough. Seb PS: I've added back the mailing list so others could benefit from our discussion. On Mon, Oct 2, 2017 at 10:02 AM,wrote: > Hi Seb, > > > > The screen size is 1920 x 1080, which seems like it should be sufficiently > large enough. It does seem as though that’s the problematic line from the > logs though… > > > > I’m currently running 5.4.0 and not the nightly/master so testing out > pvbatch over pvpython is difficult. I tried getting the nightly version > working but it seems as though there are more changes involved than just > that – half the paths I reference are wrong (probably due to stuff changing > paths upstream, e.g. launcher.py is in a different location now) and it > seems more trouble than it’s worth to test it. > > > > Louise > > > > *From:* Sebastien Jourdain [mailto:sebastien.jourd...@kitware.com] > *Sent:* 02 October 2017 14:47 > *To:* Davies, Louise (STFC,RAL,SC) > *Cc:* paraview@paraview.org > *Subject:* Re: [Paraview] ParaViewWeb Visualizer black screen > > > > Hi Louise, > > > > I'm wondering how big is the display on the remote server? Even if no > screen is connected, you do need to define a display size which in your > case may be tiny by default? > > > > Otherwise, if you use the nightly binaries or paraview/master, you can > tweak your launcher config to use pvbatch instead of pvpython which > leverage a true offscreen rendering which will be then independent of the > screen resolution. But you will need to set the following environment > variable. PV_ALLOW_BATCH_INTERACTION=1 > > > > It might also be something else, but the message "window did not resize" > looked suspicious. > > > > Seb > > > > > > On Mon, Oct 2, 2017 at 6:41 AM, wrote: > > Hello, > > > > I have been trying to set up a ParaViewWeb server on my CentOS 6 machine > to run Visualizer. I have gotten to a point where the UI loads correctly > but remote rendering only displays a black screen. Not even the axes in the > bottom left are displayed. However, if I switch to local rendering mode it > works as normal. There are no errors at all in the Apache logs and the > session log starts fine but after the twisted server startup and the lines > filled with plus signs, I get the lines: > > > > 2017-10-02 11:14:48+0100 [_GenericHTTPChannelProtocol,0,127.0.0.1] Client > has reconnected, cancelling reaper > > 2017-10-02 11:14:48+0100 [_GenericHTTPChannelProtocol,0,127.0.0.1] > on_connect: connection count = 1 > > 2017-10-02 11:14:48+0100 [-] Property BeginValueCapture has no GetData() > method, skipping > > 2017-10-02 11:14:48+0100 [-] Property CaptureValuesFloat has no GetData() > method, skipping > > 2017-10-02 11:14:48+0100 [-] Property CaptureZBuffer has no GetData() > method, skipping > > 2017-10-02 11:14:48+0100 [-] Property EndValueCapture has no GetData() > method, skipping > > 2017-10-02 11:14:48+0100 [-] Property Send to OpenVR has no GetData() > method, skipping > > 2017-10-02 11:14:48+0100 [-] Property StartCaptureLuminance has no > GetData() method, skipping > > 2017-10-02 11:14:48+0100 [-] Property StopCaptureLuminance has no > GetData() method, skipping > > Warning: In /home/buildslave/dashboards/buildbot/paraview-pvbinsdash- > linux-shared-release_superbuild/build/superbuild/ > paraview/src/VTK/Rendering/OpenGL2/vtkXOpenGLRenderWindow.cxx, line 1093 > > vtkXOpenGLRenderWindow (0x3e184d0): warning window did not resize in the > allotted time > > > > Since it works with local rendering and not remote rendering I feel as > though it must be a problem with OpenGL, but running glxinfo shows that the > user running the launcher has access to direct rendering. > > > > Any suggestions on what could possibly be going wrong would be greatly > appreciated. > > > > Kind Regards, > > > > Louise > > > ___ > Powered by www.kitware.com > > Visit other Kitware open-source projects at http://www.kitware.com/ > opensource/opensource.html > > Please keep messages on-topic and check the ParaView Wiki at: > http://paraview.org/Wiki/ParaView > > Search the list archives at: http://markmail.org/search/?q=ParaView > > Follow this link to subscribe/unsubscribe: > http://public.kitware.com/mailman/listinfo/paraview > > > ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the ParaView Wiki at: http://paraview.org/Wiki/ParaView Search the list
Re: [Paraview] ParaView session state
Hi Eduardo, That message is indeed the state but only for the proxy object and the VTK instance underneath. Unfortunately (for you) the state of the vtkSMSessionProxyManager is not the whole state of the application. Each state is specific to its instance. In order to get the full state of the application you will have to traverse the tree of objects that compose the application and gather all the states on the way. Unfortunately I couldn't find the time to walk through the code to validate what I'm saying. Hope this help and sorry I could not get you a better answer or better pointers on where to look at. Seb On Wed, Sep 27, 2017 at 3:26 PM,wrote: > Hi Aron and all, > > Following this discussion > > For what I can see, all the session state is stored in a variable called > "State" inside Internals of vtkSMSessionProxyManager ( > vtkSMSessionProxyManagerInternals). This State is of type vtkSMMessage, > which is a typedef of *paraview_protobuf::Message*. > > In other words, the state of the whole session is stored as a generated > protobuf class (or package), which is defined in vtkPVMessage.proto. I > guess this make sense as -- I image -- this makes read/write from a flat > XML much easier (as well as transferring from client-to-server and vice > versa). It would be great if someone with more knowledge on these details > can clarify that I am in fact correct... > > Cheers, > Eduardo. > > September 26, 2017 10:33 AM, "Aron Helser" <%22aron%20helser%22%20%3caron.hel...@kitware.com%3E>> wrote: > > Hi Eduardo, > I think you've basically got it, but the state is more complicated than > you think. Basically everything has a vtkSMProxy, and those proxies contain > the state of the pipeline, representation, and views. Anything that is > modified from the defaults gets saved in the state file. The > vtkSMSessionProxyManager > is the root of a session, as you've found. > The input data is saved in the state file as absolute paths, and you do > need the input data. When you read state, it has that intelligent option to > search for the data files in another directory... > I've just been learning this stuff trying to add dynamic lights to > ParaView, so hopefully I'm close :) > HTH, Regards, > Aron > On Tue, Sep 26, 2017 at 12:33 PM, wrote: > > Hello list, > > Where is the data structure(s) in the code where all the objects for *the > state* of a particular ParaView session are kept while it is running?. I > have been reading the code, and I think it should be in the "internals" of > vtkSMSessionProxyManager, but I am unsure. As far as I understand, the > state (which can be saved to XML) is all it takes to recreate a session (Am > I right here? Probably you need the input data as well, right?). > > Thanks for the help!, > Eduardo. > > P.S.: Are my emails going to SPAM? Can someone respond ("hi") to this > email to confirm otherwise? Thanks !. > > ___ > Powered by www.kitware.com > > Visit other Kitware open-source projects at http://www.kitware.com/opensou > rce/opensource.html > > Please keep messages on-topic and check the ParaView Wiki at: > http://paraview.org/Wiki/ParaView > > Search the list archives at: http://markmail.org/search/?q=ParaView > > Follow this link to subscribe/unsubscribe: > http://public.kitware.com/mailman/listinfo/paraview > > > > ___ > Powered by www.kitware.com > > Visit other Kitware open-source projects at http://www.kitware.com/ > opensource/opensource.html > > Please keep messages on-topic and check the ParaView Wiki at: > http://paraview.org/Wiki/ParaView > > Search the list archives at: http://markmail.org/search/?q=ParaView > > Follow this link to subscribe/unsubscribe: > http://public.kitware.com/mailman/listinfo/paraview > > ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the ParaView Wiki at: http://paraview.org/Wiki/ParaView Search the list archives at: http://markmail.org/search/?q=ParaView Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/paraview
Re: [Paraview] Remote rendering through reverse SSH breaks with a GLX error
Robert, A few suggestions: 1. let's see what GL you're getting. If you run the following, what it the output? pgc104% DISPLAY=:0.0 glxinfo 2. What GPU do you have and which drivers? If it's NVidia with new enough drivers that support EGL, an EGL-enabled build may be a good way to circumvent x-server issues. Utkarsh On Thu, Sep 28, 2017 at 11:43 AM, Robert Sawkowrote: > Dear Paraviewers, > > I am trying to run Paraview in a client-server mode on a remote compute node > using a reverse SSH through the login node. > > I have successfully worked out my port forwarding skulduggery and was able to > connect to a host which does NOT run X server. Inside my local client. I do > get > an error > > "Display is not accessible on the server side. Remote rendering > will be disabled" > > This is to be expected on the node without X. > > On the node with X running my client gets connected but then breaks with: > > Waiting for client... > Connection URL: cs://pgc104:6 > Accepting connection(s): pgc104:6 > Client connected. > Xlib: extension "GLX" missing on display ":0.0". > Xlib: extension "GLX" missing on display ":0.0". > ... > (core dumped) > > > Finally on the login node client-server works fine and rendering (I guess) > is done on the remote side as I don't get any errors. We have made sure that > xorg.conf on login and X-enabled compute node are the same. > > > My connection procedure: > > local% ssh -L 6:localhost:6 login_node > login% bsub -R "select[hname=pgc104]" -Is /bin/bash > pgc104% ssh -N -R 6:localhost:6 login & > pgc104% DISPLAY=:0.0 pvserver -sp=6 --use-offscreen-rendering > local% paraview->connect to localhost:6 > > Same thing works on pgc304 which doesn't run X. Please let us know if you have > any suggestions. > > Best wishes, > Robert > -- > Ironically, it was Christopher Columbus who was wrong > http://en.wikipedia.org/wiki/Myth_of_the_Flat_Earth > ___ > Powered by www.kitware.com > > Visit other Kitware open-source projects at > http://www.kitware.com/opensource/opensource.html > > Please keep messages on-topic and check the ParaView Wiki at: > http://paraview.org/Wiki/ParaView > > Search the list archives at: http://markmail.org/search/?q=ParaView > > Follow this link to subscribe/unsubscribe: > http://public.kitware.com/mailman/listinfo/paraview ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the ParaView Wiki at: http://paraview.org/Wiki/ParaView Search the list archives at: http://markmail.org/search/?q=ParaView Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/paraview
Re: [Paraview] ParaViewWeb Visualizer black screen
Hi Louise, I'm wondering how big is the display on the remote server? Even if no screen is connected, you do need to define a display size which in your case may be tiny by default? Otherwise, if you use the nightly binaries or paraview/master, you can tweak your launcher config to use pvbatch instead of pvpython which leverage a true offscreen rendering which will be then independent of the screen resolution. But you will need to set the following environment variable. PV_ALLOW_BATCH_INTERACTION=1 It might also be something else, but the message "window did not resize" looked suspicious. Seb On Mon, Oct 2, 2017 at 6:41 AM,wrote: > Hello, > > > > I have been trying to set up a ParaViewWeb server on my CentOS 6 machine > to run Visualizer. I have gotten to a point where the UI loads correctly > but remote rendering only displays a black screen. Not even the axes in the > bottom left are displayed. However, if I switch to local rendering mode it > works as normal. There are no errors at all in the Apache logs and the > session log starts fine but after the twisted server startup and the lines > filled with plus signs, I get the lines: > > > > 2017-10-02 11:14:48+0100 [_GenericHTTPChannelProtocol,0,127.0.0.1] Client > has reconnected, cancelling reaper > > 2017-10-02 11:14:48+0100 [_GenericHTTPChannelProtocol,0,127.0.0.1] > on_connect: connection count = 1 > > 2017-10-02 11:14:48+0100 [-] Property BeginValueCapture has no GetData() > method, skipping > > 2017-10-02 11:14:48+0100 [-] Property CaptureValuesFloat has no GetData() > method, skipping > > 2017-10-02 11:14:48+0100 [-] Property CaptureZBuffer has no GetData() > method, skipping > > 2017-10-02 11:14:48+0100 [-] Property EndValueCapture has no GetData() > method, skipping > > 2017-10-02 11:14:48+0100 [-] Property Send to OpenVR has no GetData() > method, skipping > > 2017-10-02 11:14:48+0100 [-] Property StartCaptureLuminance has no > GetData() method, skipping > > 2017-10-02 11:14:48+0100 [-] Property StopCaptureLuminance has no > GetData() method, skipping > > Warning: In /home/buildslave/dashboards/buildbot/paraview-pvbinsdash- > linux-shared-release_superbuild/build/superbuild/ > paraview/src/VTK/Rendering/OpenGL2/vtkXOpenGLRenderWindow.cxx, line 1093 > > vtkXOpenGLRenderWindow (0x3e184d0): warning window did not resize in the > allotted time > > > > Since it works with local rendering and not remote rendering I feel as > though it must be a problem with OpenGL, but running glxinfo shows that the > user running the launcher has access to direct rendering. > > > > Any suggestions on what could possibly be going wrong would be greatly > appreciated. > > > > Kind Regards, > > > > Louise > > ___ > Powered by www.kitware.com > > Visit other Kitware open-source projects at http://www.kitware.com/ > opensource/opensource.html > > Please keep messages on-topic and check the ParaView Wiki at: > http://paraview.org/Wiki/ParaView > > Search the list archives at: http://markmail.org/search/?q=ParaView > > Follow this link to subscribe/unsubscribe: > http://public.kitware.com/mailman/listinfo/paraview > > ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the ParaView Wiki at: http://paraview.org/Wiki/ParaView Search the list archives at: http://markmail.org/search/?q=ParaView Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/paraview
[Paraview] ParaViewWeb Visualizer black screen
Hello, I have been trying to set up a ParaViewWeb server on my CentOS 6 machine to run Visualizer. I have gotten to a point where the UI loads correctly but remote rendering only displays a black screen. Not even the axes in the bottom left are displayed. However, if I switch to local rendering mode it works as normal. There are no errors at all in the Apache logs and the session log starts fine but after the twisted server startup and the lines filled with plus signs, I get the lines: 2017-10-02 11:14:48+0100 [_GenericHTTPChannelProtocol,0,127.0.0.1] Client has reconnected, cancelling reaper 2017-10-02 11:14:48+0100 [_GenericHTTPChannelProtocol,0,127.0.0.1] on_connect: connection count = 1 2017-10-02 11:14:48+0100 [-] Property BeginValueCapture has no GetData() method, skipping 2017-10-02 11:14:48+0100 [-] Property CaptureValuesFloat has no GetData() method, skipping 2017-10-02 11:14:48+0100 [-] Property CaptureZBuffer has no GetData() method, skipping 2017-10-02 11:14:48+0100 [-] Property EndValueCapture has no GetData() method, skipping 2017-10-02 11:14:48+0100 [-] Property Send to OpenVR has no GetData() method, skipping 2017-10-02 11:14:48+0100 [-] Property StartCaptureLuminance has no GetData() method, skipping 2017-10-02 11:14:48+0100 [-] Property StopCaptureLuminance has no GetData() method, skipping Warning: In /home/buildslave/dashboards/buildbot/paraview-pvbinsdash-linux-shared-release_superbuild/build/superbuild/paraview/src/VTK/Rendering/OpenGL2/vtkXOpenGLRenderWindow.cxx, line 1093 vtkXOpenGLRenderWindow (0x3e184d0): warning window did not resize in the allotted time Since it works with local rendering and not remote rendering I feel as though it must be a problem with OpenGL, but running glxinfo shows that the user running the launcher has access to direct rendering. Any suggestions on what could possibly be going wrong would be greatly appreciated. Kind Regards, Louise ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the ParaView Wiki at: http://paraview.org/Wiki/ParaView Search the list archives at: http://markmail.org/search/?q=ParaView Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/paraview