Re: [Paraview] pvserver always shows 100% cpu usage
Hi John, I compared a)MPI_Send/MPI_Recv with b)MPI_ISend/MPI_IRecv/MPI_Wait and it depends on the situation which style is the fastest - style b is not simply worse than style a. If you have high latency because of your network-hardware you are better of with style b. But it is difficult to tell, because there is not much difference. Anyway, is there one single part in the paraview-code, which handles the MPI-Stuff, or is MPI all over the place? If first is the case, it would be easy to implement some wait-packet+constant polling of process 0 to get rid of MPI-lib-dependens, but processor and/or memory affinity is probably a problem then ... (http://www.open-mpi.org/faq/?category=tuning#using-paffinity) Greetings Jens John M. Patchett schrieb: Hi Jens, I would think that each pvserver process would have to be able to detect that it was in a lengthy wait state because it sure would suck to have an MPI_Wait in the middle of each send/recv pair during compositing ... And it would probably be difficult to handle the tiled display communication. The design of a solution might not be so straight forward I think, you would have to first decide that it really is a problem versus cost of doing business. -John. On Dec 5, 2008, at 9:53 AM, Jens wrote: Hi John, thought about this problem again... A solution could be a)to use MPI_IRecv/MPI_Wait instead of MPI_Recv. If that results in the same 100% cpu for the MPI_Wait-command, b) it could be a solution to add a wait()/sleep() just beweeen MPI_IRecv and MPI_Wait. How long this wait/sleep would be could depend a) on the time the function MPI_Wait takes to return. b) or on some value send by process 0 (process 0 could send other processes to sleep for a while) What do you think? Greetings Jens John M. Patchett schrieb: Hi Jens, Your pvserver is probably waiting on an MPI_Recv and your MPI implementation is spinning. You will note that process 0 probably isn't doing this, as the other nodes are waiting on process 0 to send. I have searched this problem all the way to the MPI developers as it's easy to replicate without paraview and the MPI guys assure me the alternatives are worse. -John. On Dec 5, 2008, at 8:42 AM, Jens wrote: Hi, if I run mpirun -np 4 ./pvserver on our cluster-node and connect from my client, this pvserver always shows 100% cpu usage - even if I do nothing at the client. Seems to me as if there is a loop waiting for the client to ask for action - but this loop is calling no wait/sleep function. Greetings Jens ___ ParaView mailing list ParaView@paraview.org http://www.paraview.org/mailman/listinfo/paraview ___ ParaView mailing list ParaView@paraview.org http://www.paraview.org/mailman/listinfo/paraview
Re: [Paraview] pvserver always shows 100% cpu usage
Hi Paul, this is a very good hint. Thanks a lot. But is sadly does not work :( I started pvdataserver like this: mpirun --mca mpi_yield_when_idle 1 -np 4 ./pvdataserver and still see three processes with 100% cpu, while no-one is connected. This happens on a Dell R805 (2xAMD QuadCore-Opteron) with CentOS 5.2, OpenMPI 1.2.8 and Paraview 3.4.0 - all compiled with Intel 11.0. Greetings Jens Paul Edwards schrieb: Hi all, It just depends on how it is implemented with MPI. There will be lower latency if it spins in a loop waiting for a message. Here is a link to the FAQ which shows how you can stop this for OpenMPI: http://www.open-mpi.org/faq/?category=running#oversubscribing http://www.open-mpi.org/faq/?category=running#force-aggressive-degraded Regards, Paul 2008/12/5 John M. Patchett [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] Hi Jens, If I recall correctly, your observation lies in the type of hardware you are using. I think some hardware allows a developer to leverage an interrupt while some requires polling for a received message. Design requirements for MPI to be fast with low latency usually (I presume) outweigh lowering heat production. I think I first noticed this issue with mvapich when infiniband was new. I don't think we had it with Myrinet 2000 or Gig-E. I also think ParaView is a normal mpi application, we just don't notice machines spinning while waiting during batch processing, we just assume they're working hard. -John. On Dec 5, 2008, at 9:42 AM, Jens wrote: Hi John, thanks for your answer. That makes sense. Normal mpi-apps are probably not written to wait for more things to do - they are simply always busy. It is just a pity that the cluster has to run 100% producing a lot of heat for nothing. So the MPI-lib will probably not change this behavior :( ? (I am using open-mpi 1.2.8) Greetings Jens John M. Patchett schrieb: Hi Jens, Your pvserver is probably waiting on an MPI_Recv and your MPI implementation is spinning. You will note that process 0 probably isn't doing this, as the other nodes are waiting on process 0 to send. I have searched this problem all the way to the MPI developers as it's easy to replicate without paraview and the MPI guys assure me the alternatives are worse. -John. On Dec 5, 2008, at 8:42 AM, Jens wrote: Hi, if I run mpirun -np 4 ./pvserver on our cluster-node and connect from my client, this pvserver always shows 100% cpu usage - even if I do nothing at the client. Seems to me as if there is a loop waiting for the client to ask for action - but this loop is calling no wait/sleep function. Greetings Jens ___ ParaView mailing list ParaView@paraview.org mailto:ParaView@paraview.org http://www.paraview.org/mailman/listinfo/paraview ___ ParaView mailing list ParaView@paraview.org mailto:ParaView@paraview.org http://www.paraview.org/mailman/listinfo/paraview ___ ParaView mailing list ParaView@paraview.org http://www.paraview.org/mailman/listinfo/paraview
Re: [Paraview] pvserver always shows 100% cpu usage
Hi Paul, I read the following on open-mpi-mailinglist: One issue is that yield might mean yield if there is someone else ready to run. Like a traffic sign: if someone else is there, you yield. If no one else is there, there's no way to tell that someone is yielding. Next, even if someone else is trying to run, yield doesn't give up the CPU 100%. It's still rather pesky. so, to make is short: this is no solution :( Greetings Jens Paul Edwards schrieb: Hi all, It just depends on how it is implemented with MPI. There will be lower latency if it spins in a loop waiting for a message. Here is a link to the FAQ which shows how you can stop this for OpenMPI: http://www.open-mpi.org/faq/?category=running#oversubscribing http://www.open-mpi.org/faq/?category=running#force-aggressive-degraded Regards, Paul 2008/12/5 John M. Patchett [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] Hi Jens, If I recall correctly, your observation lies in the type of hardware you are using. I think some hardware allows a developer to leverage an interrupt while some requires polling for a received message. Design requirements for MPI to be fast with low latency usually (I presume) outweigh lowering heat production. I think I first noticed this issue with mvapich when infiniband was new. I don't think we had it with Myrinet 2000 or Gig-E. I also think ParaView is a normal mpi application, we just don't notice machines spinning while waiting during batch processing, we just assume they're working hard. -John. On Dec 5, 2008, at 9:42 AM, Jens wrote: Hi John, thanks for your answer. That makes sense. Normal mpi-apps are probably not written to wait for more things to do - they are simply always busy. It is just a pity that the cluster has to run 100% producing a lot of heat for nothing. So the MPI-lib will probably not change this behavior :( ? (I am using open-mpi 1.2.8) Greetings Jens John M. Patchett schrieb: Hi Jens, Your pvserver is probably waiting on an MPI_Recv and your MPI implementation is spinning. You will note that process 0 probably isn't doing this, as the other nodes are waiting on process 0 to send. I have searched this problem all the way to the MPI developers as it's easy to replicate without paraview and the MPI guys assure me the alternatives are worse. -John. On Dec 5, 2008, at 8:42 AM, Jens wrote: Hi, if I run mpirun -np 4 ./pvserver on our cluster-node and connect from my client, this pvserver always shows 100% cpu usage - even if I do nothing at the client. Seems to me as if there is a loop waiting for the client to ask for action - but this loop is calling no wait/sleep function. Greetings Jens ___ ParaView mailing list ParaView@paraview.org mailto:ParaView@paraview.org http://www.paraview.org/mailman/listinfo/paraview ___ ParaView mailing list ParaView@paraview.org mailto:ParaView@paraview.org http://www.paraview.org/mailman/listinfo/paraview ___ ParaView mailing list ParaView@paraview.org http://www.paraview.org/mailman/listinfo/paraview
Re: [Paraview] [Ensight] 2D gold forme
Josue: Would you please provide the dataset and procedure used to re-produce the bug? Thanks. -Zhanping -- Zhanping Liu, PhD Kitware, Inc. 28 Corporate Drive Clifton Park, NY 12065-8662 Phone: 518-371-3971 x 138 http://www.zhanpingliu.org On Fri, Dec 5, 2008 at 9:29 AM, Josué Barboza [EMAIL PROTECTED]wrote: Hi all, I'm trying to read a 2D ensight gold results. The .case file looks like: FORMAT type: ensight gold GEOMETRY model: NS_Inlet.ens.geo VARIABLE scalar per node: pressure NS_Inlet.ens.pressure vector per node: velocity NS_Inlet.ens.velocity scalar per node: temperature NS_Inlet.ens.temperature The variable files are in binary format. Paraview shows the mesh but nothing else (no variables). All is ok when Ensight is used. Thx in advance for you help. -- Dr. Josué BARBOZA Research Scientist CENAERO, Virtual Manufacturing Group PLEASE NOTICE OUR NEW OFFICE LOCATION Bâtiment EOLE, 1er étage Rue des Frères Wright, 29 B-6041 Gosselies Belgium Tel: + 32 (0)71 91.93.67 Fax: + 32 (0)71 91.93.31 Visit our web site at www.cenaero.be and our dedicated Friction Stir Welding website www.fsw-cenaero.org CENAERO participates to the VIVACE European project (www.vivaceproject.com ) CENAERO bénéficie du soutien de la Région Wallonne et des fonds structurels FEDER et FSE de l'Union Européenne +-+-+- Disclaimer +-+-+- Privileged/Confidential Information may be contained in this message. If you are not the addressee indicated in this message (or responsible for delivery of the message to such person), you may not print, retain, copy nor disseminate this message or any part of it to anyone and you should notify the sender by reply email and destroy this message. Neglecting this clause could be a breach of confidence. Please advise immediately if you or your employer does not consent to Internet email for messages of this kind. Opinions, conclusions and other information in this message that are not related to the official business of my firm shall be understood as neither given nor endorsed by it. ___ ParaView mailing list ParaView@paraview.org http://www.paraview.org/mailman/listinfo/paraview ___ ParaView mailing list ParaView@paraview.org http://www.paraview.org/mailman/listinfo/paraview
Re: [Paraview] default_server.pvsc in PV-3.4.0
Ramesh, Try putting the file in under lib/paraview-3-4 (same location as paraview-real executable). Utkarsh On Fri, Nov 21, 2008 at 10:04 AM, Rakesh Hammond [EMAIL PROTECTED] wrote: Dear all, I am trying to use the default_server.pvsc file to load up profiles, which are site specifi rather than user specific - as explained in the wiki page, this can be done by adding default_server.pvsc file to the bin directory. I got this to work without any problem. However, duing compilation, I specified an INSTALL_PREFIX, if I now copy the above file to this bin directory (this gets created when doing make install), paraview loads up ok, but it shows nothing in the server dialog box. It looks like, I need to copy some more files across. Does anyone know which files its looking for? Can some one please give me some pointers on how to solve this problem? Thanks, Rakesh ps : I am running PV-3.4.0 on Linux ___ ParaView mailing list ParaView@paraview.org http://www.paraview.org/mailman/listinfo/paraview ___ ParaView mailing list ParaView@paraview.org http://www.paraview.org/mailman/listinfo/paraview
Re: [Paraview] reg file loading
Is it possible that your csv file has a blank line at the end? Anyways, currently there's currently no direct way of visualizing the data read in by CSV except in the spreadsheet view to simply look at the raw data (or plots in CVS paraview). You'll have to write a python programmable filter or source. Look at http://www.paraview.org/Wiki/Here_are_some_more_examples_of_simple_ParaView_3_python_filters. The first example illustrates reading a CSV as a set of points. Utkarsh On Sun, Nov 23, 2008 at 5:31 AM, L Susithra [EMAIL PROTECTED] wrote: Hi, Thanks for your response. I am getting this error: ERROR: In ..\..\..\ParaView3\VTK\Filtering\vtkTable.cxx, line 194 vtkTable (0B71C660): Incorrect number of tuples in SetRow The version is 3.4.0. i have first three columns are node positions and the next three columns are corresponding x,y,z velocity vectors. I am not sure which format i should use it to load the data for flow visualization. Thanks Date: Fri, 21 Nov 2008 14:22:18 -0500 To: [EMAIL PROTECTED] CC: paraview@paraview.org Subject: Re: [Paraview] reg file loading From: [EMAIL PROTECTED] What is the error you are getting? What version of ParaView are you using? Utkarsh L Susithra wrote: Hello, I've a data ( velocity fields, x,y,z for each node )from the simulation and i need to load it in paraview to visulaise the flow field. i converted the data as .csv from excel and when i read it from paraview, i am not able to load the data , getting an error. It just reads the first column only (xvelocity). I am not sure is there a specific way to load the data. It would be a great help if any one could suggest me. Thanks Free up your senses. Experience reality up close on MSN video Try it! http://video.msn.com/?mkt=en-in ___ ParaView mailing list ParaView@paraview.org http://www.paraview.org/mailman/listinfo/paraview Calling TV buffs! Get TV listings, gossip on your fave stars and updates on hot new shows Try it now! ___ ParaView mailing list ParaView@paraview.org http://www.paraview.org/mailman/listinfo/paraview
[Paraview] Can I rotate camera in animation?
Hallo! Can I use different filters in animation or rotate camera? -- Andrey ___ ParaView mailing list ParaView@paraview.org http://www.paraview.org/mailman/listinfo/paraview
Re: [Paraview] Fluent / Ensight data on cluster
I tracked this down and you are right. When reading EnSight files (not sos), all processes read the whole dataset and then each one extracts a piece. If you submit a request at http://paraview.org/Bug, we can address this issue in the future. -berk On Fri, Dec 5, 2008 at 3:48 AM, Jacques Papper [EMAIL PROTECTED] wrote: The grid is unstructured. It comes up as a multiblock dataset in ParaView, but there is only one fluid volume. The other blocks are just surfaces. Jacques 2008/12/4 Berk Geveci [EMAIL PROTECTED] The answer depends on the data type. Is it structured or unstructured? On Dec 4, 2008, at 11:49 AM, Jacques Papper [EMAIL PROTECTED] wrote: Hi, I compiled Paraview for use in parallel. I have an Ensight data set that has been output from fluent. I also have the fluent .cas and .dat files (saved from a cluster run) I would like to load the data using the best of paraview in parallel. If I launch pvserver on 4 nodes and load the ensight data it seems that it loads the data 4 times. What would be the best / efficient way of using the parallel functionality ? Is there a way of partitioning the ensight file ? Or setting up the reader so that it partitions the data at the reading stage ? Or maybe is there a Fluent reader that will load in the partitions from the .cas file ? Jacques ___ ParaView mailing list ParaView@paraview.org http://www.paraview.org/mailman/listinfo/paraview ___ ParaView mailing list ParaView@paraview.org http://www.paraview.org/mailman/listinfo/paraview