[Paraview] Paraview v5.1.2: SurfaceLIC & software rendering issues

Thu Sep 29 12:25:25 EDT 2016

Ken,

I have included as attachment to this email the script I modified to reproduce the issue reported earlier with our generic F1 car. You will find the following structure once you have unpacked the zip archive:

Benchmark/
|-- Binaries
|   `-- ParaView-5.1.2
|-- Data
|   `-- shuttle-surf
|       |-- shuttle-surf_0_0.vtp
|       |-- shuttle-surf_0_0.vtp.md5-stamp
|       |-- shuttle-surf_1_0.vtp
|       |-- shuttle-surf_1_0.vtp.md5-stamp
|       |-- shuttle-surf_2_0.vtp
|       |-- shuttle-surf_2_0.vtp.md5-stamp
|       |-- shuttle-surf_3_0.vtp
|       |-- shuttle-surf_3_0.vtp.md5-stamp
|       |-- shuttle-surf_4_0.vtp
|       |-- shuttle-surf_4_0.vtp.md5-stamp
|       |-- shuttle-surf_5_0.vtp
|       |-- shuttle-surf_5_0.vtp.md5-stamp
|       |-- shuttle-surf_6_0.vtp
|       |-- shuttle-surf_6_0.vtp.md5-stamp
|       |-- shuttle-surf_7_0.vtp
|       `-- shuttle-surf_7_0.vtp.md5-stamp
`-- ParaView-5.1.2
    |-- LoResImages
    |-- runShuttle.sh
    `-- shuttleBenchmarkSurfaceLIC.py

The Data directory contains all surface files used for performing rendering. Binaries/ParaView-5.1.2 contains the installation of ParaView. This is where I have extracted the ParaView binaries I downloaded last week from the paraview.org site (ParaView-5.1.2-Qt4-OpenGL2-MPI-Linux-64bit.tar.gz).

The benchmark is started using script “runShuttle.sh”. It will use pvbatch to run the python script shuttleBenchmarkSurfaceLIC.py using OSMesa with llvm-pipe and swr-avx2. Both options are run using 1 core, 2 cores and 3 cores. Generated images are stored respectively in folder LoResImages.

With this script, I got the same issues as reported earlier, i.e. the --mesa-llvm on 1 core works well, on 2 cores the background is black instead of white, the --mesa-swr-avx2 option delivers a corrupted image:

mesa-llvm 1 core:
[cid:image001.jpg at 01D21A7E.C927DCC0]
mesa-llvm 2 cores:
[cid:image002.jpg at 01D21A7E.C927DCC0]

mesa-swr-avx2 1 core:
[cid:image003.jpg at 01D21A7E.C927DCC0]
The plaftorm I used to generate these images is:

OpenSuSE 12.3
Linux 3.7.10-1.45-desktop #1 SMP PREEMPT x86_64 GNU/Linux
2 x 10 core Intel Xeon ES-2650 @ 2.30GHz

Thank you very much in advance for your help.

Best regards,

Frank.

From: ParaView [mailto:paraview-bounces at paraview.org] On Behalf Of Albina, Frank
Sent: Mittwoch, 28. September 2016 17:36
To: Ken Martin
Cc: Del Citto, Francesco; paraview at paraview.org
Subject: Re: [Paraview] Paraview v5.1.2: SurfaceLIC & software rendering issues

Hi Ken!

Sure will do. I’ll come back to you asap with a self-contained test loop using your shuttle data.

Regards,

Frank.

From: Ken Martin [mailto:ken.martin at kitware.com]
Sent: Mittwoch, 28. September 2016 17:20
To: Albina, Frank
Cc: paraview at paraview.org<mailto:paraview at paraview.org>; Del Citto, Francesco
Subject: Re: [Paraview] Paraview v5.1.2: SurfaceLIC & software rendering issues

Running from the GUI with mpi and 2 nodes seems to maintain the background on my quick tests. Could you update your script to use the shuttle dataset? I included a copy of that data on this email. That would make it easier to debug.

Thanks
Ken

On Mon, Sep 26, 2016 at 4:55 AM, Albina, Frank <frank.albina at sauber-motorsport.com<mailto:frank.albina at sauber-motorsport.com>> wrote:
Dear all,

I have downloaded from the paraview.org<http://paraview.org> the latest binaries for Linux of ParaView v5.1.2. My main purpose is to make use of ParaView with software rendering in a clustered environment without GPU and, as such, I have been going through my small benchmark on a generic F1 car.

This what I have found so far when generating a SurfaceLIC representation on the whole car surface in batch mode.

Using pvbatch –mesa-llvm works fine. This is one of the produced images for reference:
[cid:image022.jpg at 01D21A7C.A741F430]
Using 4 cores to perform the same task (with the same script) with mpiexec –np 4 pvbatch --mesa-llvm shows a good speed-up but the background is black instead of white (although view.Backgound = [1,1,1] is set explicitely), as shown hereafter:

[cid:image023.jpg at 01D21A7C.A741F430]

On the same server (Intel Xeon ES-2650 v3), the CPU is AVX2 capable, so I tried also the same script with pvbatch --mesa-swr-avx2. The image background is corrupted with leftover frames apparently, as shown hereafter:

[cid:image024.jpg at 01D21A7C.A741F430]
Was this behavior to be expected?

Please let me know if there is any interest in providing with my benchmark.

Frank Albina

_______________________________________________
Powered by www.kitware.com<http://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

--
Ken Martin PhD
Chairman & CFO
Kitware Inc.
28 Corporate Drive
Clifton Park NY 12065
518 371 3971

This communication, including all attachments, contains confidential and legally privileged information, and it is intended only for the use of the addressee.  Access to this email by anyone else is unauthorized. If you are not the intended recipient, any disclosure, copying, distribution or any action taken in reliance on it is prohibited and may be unlawful. If you received this communication in error please notify us immediately and destroy the original message.  Thank you.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/paraview/attachments/20160929/4ca573af/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image022.jpg
Type: image/jpeg
Size: 15611 bytes
Desc: image022.jpg
URL: <http://public.kitware.com/pipermail/paraview/attachments/20160929/4ca573af/attachment-0006.jpg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image023.jpg
Type: image/jpeg
Size: 14417 bytes
Desc: image023.jpg
URL: <http://public.kitware.com/pipermail/paraview/attachments/20160929/4ca573af/attachment-0007.jpg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image024.jpg
Type: image/jpeg
Size: 26191 bytes
Desc: image024.jpg
URL: <http://public.kitware.com/pipermail/paraview/attachments/20160929/4ca573af/attachment-0008.jpg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.jpg
Type: image/jpeg
Size: 10930 bytes
Desc: image001.jpg
URL: <http://public.kitware.com/pipermail/paraview/attachments/20160929/4ca573af/attachment-0009.jpg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image002.jpg
Type: image/jpeg
Size: 22018 bytes
Desc: image002.jpg
URL: <http://public.kitware.com/pipermail/paraview/attachments/20160929/4ca573af/attachment-0010.jpg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image003.jpg
Type: image/jpeg
Size: 19925 bytes
Desc: image003.jpg
URL: <http://public.kitware.com/pipermail/paraview/attachments/20160929/4ca573af/attachment-0011.jpg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Benchmark.zip
Type: application/x-zip-compressed
Size: 2421512 bytes
Desc: Benchmark.zip
URL: <http://public.kitware.com/pipermail/paraview/attachments/20160929/4ca573af/attachment-0001.bin>