Folding movies

[Jesse_V]

I have been having difficulties trying to get it to connect to V7. I tried the -port 36330 option, and

Code: Select all

20:29:47:Server connection id=18 on 0.0.0.0:36330 from 127.0.0.1
20:29:58:Server connection id=18 ended

appears in my log. Viewer.exe says that it couldn't get data due to a timeout. Under Configure->Remote Access, 127.0.0.1 is listed as Allow under the Passwordless IP Address Restriction, so I don't think it's being blocked. FahCore A4. Any ideas? Or does this only work for v6?

EDIT: Playback is amazing! However, FAHViewer's rotation controls aren't as wild, so there its much easier to rotate its demo protein around. Is FAHViewer open-source? I thought it was.

[gwildperson]

The website for viewer.exe says "Folding@Home client's port. 61469 by default" which is (IIRC) is the port the V6 client used to connect to it's viewer. I don't believe V7 is supported by viewer.exe though I'd be happy to hear from the OP that I'm wrong.

If I''m correct, it's unfortunate that Mikhail M is building a utility for a (soon to be) obsolete client and almost everybody responding to him is talking about the Windows public beta client V7.1.52

[iceman1992]

On the requirements page http://www.refplane.com/folding/Requirements.html

software: installed Folding@Home GPU client v. 6.41 for writing movies. Will probably work with other v. 6 clients. This is not required for playing movies.

[Mikhail M]

EDIT: Playback is amazing! However, FAHViewer's rotation controls aren't as wild, so there its much easier to rotate its demo protein around. Is FAHViewer open-source? I thought it was.

Thank you, Jesse_V . FAHViewer is not open source, but I was able to find some v. 6 sources, they were not password-protected. There are several places, the sources in them are differ greatly and I had problems with building all of them. So there is a mess there. For v. 7 I was unable to find any sources. The protocol between the folding calculator and the viewer has changed. I don't know the new protocol, so I can't support it. The current viewer can connect to the port, but I was unable to understand what I have to send through the port now and what I will get.

[Jesse_V]

EDIT: Playback is amazing! However, FAHViewer's rotation controls aren't as wild, so there its much easier to rotate its demo protein around. Is FAHViewer open-source? I thought it was.

Thank you, Jesse_V . FAHViewer is not open source, but I was able to find some v. 6 sources, they were not password-protected. There are several places, the sources in them are differ greatly and I had problems with building all of them. So there is a mess there. For v. 7 I was unable to find any sources. The protocol between the folding calculator and the viewer has changed. I don't know the new protocol, so I can't support it. The current viewer can connect to the port, but I was unable to understand what I have to send through the port now and what I will get.

I found several links that might be useful to you:
https://fah-web.stanford.edu/svn/pub/trunk/viewer/ and look at the "src" folder.
Documentation about socket connections, etc: https://fah-web.stanford.edu/projects/F ... eInterface
Here is FAHViewer's documention: https://fah-web.stanford.edu/projects/F ... /FahViewer

[jcoffland]

I was unable to connect to the latest 7-th version nor find suitable sources. Is it possible to get them anywhere?

The v7 viewer source is available in subversion here:

Code: Select all

svn co https://fah-web.stanford.edu/svn/pub/trunk/viewer

FYI, it appears you used the the v6 source code for your viewer. I believe you are required to release your source code under the Mithral License A. Although the terms of this license are confusing to me. The v7 code is GPL 2+.

[Mikhail M]

Oh, that's wonderful, thank you! It's strange that I was unable to find them.

I have tried building the v. 7 Viewer and have even found the cbang library the sources depends on. But I didn't find MS Visual Studio project for it. As I understand it is possible to build them and all theirs dependencies on Windows. And it is probably easy with the help of Python and Scons. But I'm not familiar with them. I will probably try this later. Do you probably know a good instruction covering this (for example for another project with similar config)? Or is it enough to install latest versions of Python and Scons?

I will open my Viewer's source code. And I hope I will not even be sorry too much about that . I could open it now if somebody is being interested in integrating it into v. 7 Viewer. Although there is nothing special in it... It will be after I finish compression of the movies

[iceman1992]

I need help.. It only shows test protein, using GPU client version 6.41 on Win 7.. What do I do?
When I press W, it stays "Writing movie: 0 frames" and "Error on getting data: timeout"

[Mikhail M]

iceman1992, are you able to run the standard viewer (right-click on the F@h icon in the task bar / Display)? Most probably it's a firewall problem - something is blocking connection to the localhost:61469 port

[iceman1992]

The standard viewer also displays the test protein, as it has for a year or so.. I checked my firewall's events, it doesn't block the viewer..

[Mikhail M]

Then you are not lucky... You may try to reinstall the client as in the last but one post at viewtopic.php?f=50&t=18444. Or to ask at this forum...

[iceman1992]

Ah so it's just me? Unlucky The v6 viewer stopped displaying the current protein for me like 1 year ago.. Haven't seen one since, even after multiple times uninstalling and reinstalling.

[Mikhail M]

As I understand, it is rather typical situation that some version of F@h doesn't work for somebody. Releasing of well tested versions is not theirs top priority. Or they simply don't have enough resources for this.

[bruce]

The change has noting to do with the viewer. It was designed to display protein data from a file called a .xyz format which was produced by some versions of gromacs. Later versions of gromacs decided not to produce that format so the well-tested versions of the viewer stopped working ... gradually ... as more an more projects began doing better science using the newer versions of Gromacs while projects that still produced the .xyz files still worked.

The viewer developed for V7 works fine with any FahCore that still produces a .xyz file but for most people, it just displays the test protein. There's an open ticket to teach the viewer how to read .trr and .tpr files (see #409) but it requires data from both files and Gromacs produces more than one format of those files so it's a significant enhancement to make the viewer work again.

It turns out to be a rather complicated problem, especially since so many people choose NOT to run the viewer because it uses resources that could otherwise be used to do more science and earn higher PPD. (Personally, even if the viewer were 100% problem free, I'd still run it rather rarely but that doesn't mean it's not important to others.)

Oh, and the viewer also comes in more than one flavor, depending on which version of graphics support it's talking to. Programming resources are, in fact, limited, and for all of these reasons, it gets assigned a lower priority than quite a number of things that limit what can be done with the folding calculations, themselves.

[iceman1992]

Ah thanks bruce for the confirmation. That's what I knew, so I was surprised to see so many people with non-demo proteins
Yes I'd still run the viewer only occasionally if it were 100% working, but it's a good tool to invite people..
And it gets much better with the ability to record the viewer like Mikhail M developed.