Volunteering to fix F@H Viewer

Moderators: Site Moderators, FAHC Science Team

Volunteering to fix F@H Viewer

Postby Quantize » Mon Mar 16, 2020 2:32 pm

Dear Folding@Home,

- Being a programmer and not (formally) a researcher:
- For strong promotional reasons, especially during this time:

=> I'd like to volunteer to fix your F@H viewer

I have plenty of programming skill yet I have never looked at your particular codebase before.
I can also just look at the VCS myself but I suspect it may be way, way more efficiënt if someone could reach out and bring me up to speed on it so that Ican take over from there.

Could the person who has been working on it before send me such a message, please?
A quick summary is all I'll need.

Sincerely yours,
Meerten
Quantize
 
Posts: 26
Joined: Mon Mar 16, 2020 2:15 pm

Re: Volunteering to fix F@H Viewer

Postby rixon » Mon Mar 16, 2020 7:33 pm

Meerteen, if you do that, I'll send you some (detailed/constructive) bug reports! Not sure what the clients are written in (looks like C or C++), but FAHControl is written in Python/Qt w/ the source on the machine, so anyone can have a go at working on that. Where to send patches... your guess is as good as mine.
rixon
 
Posts: 10
Joined: Sun Mar 15, 2020 6:25 am
Location: Kansas City

Re: Volunteering to fix F@H Viewer

Postby Nathan_P » Mon Mar 16, 2020 7:39 pm

I've forwarded Quantize's request on to the project team, they are discussing it but will probably have to forward it to the project director. Lets see what happens
Image
Nathan_P
 
Posts: 1170
Joined: Wed Apr 01, 2009 10:22 pm
Location: Jersey, Channel islands

Re: Volunteering to fix F@H Viewer

Postby asm » Mon Mar 16, 2020 8:00 pm

rixon wrote:FAHControl is written in Python/Qt


FYI - on macOS at least - FAHControl 7.5.1 is not using Qt - it's using GTK. It includes libs such as libgdk-quartz-2.0.0 and libgtkmacintegration-gtk2.2 - no Qt libs to be found...
asm
 
Posts: 10
Joined: Fri Mar 06, 2020 9:12 pm

Re: Volunteering to fix F@H Viewer

Postby Quantize » Mon Mar 16, 2020 9:34 pm

rixon wrote:Meerteen, if you do that, I'll send you some (detailed/constructive) bug reports! Not sure what the clients are written in (looks like C or C++), but FAHControl is written in Python/Qt w/ the source on the machine, so anyone can have a go at working on that. Where to send patches... your guess is as good as mine.

Great! I might take you up on that offer.

Nathan_P wrote:I've forwarded Quantize's request on to the project team, they are discussing it but will probably have to forward it to the project director. Lets see what happens

Alright! Thank you, Nathan.

I'd've left my contact info but I fear spambots and crawlers picking up on it.
I'll just be watching this topic and PM box.
Quantize
 
Posts: 26
Joined: Mon Mar 16, 2020 2:15 pm

Re: Volunteering to fix F@H Viewer

Postby JimboPalmer » Mon Mar 16, 2020 9:54 pm

I am on the outside looking in, but my understanding is that it worked in Core_a4 and not in Core_a7 so having the data layout for both cores should provide guidance. I hear rumors it works in Core_22 but only in Linux, that might or might not be informative. A version that supported a7 and 22 would be great as Core_21 is on the way out.
Tsar of all the Rushers
I tried to remain childlike, all I achieved was childish.
A friend to those who want no friends
JimboPalmer
 
Posts: 2022
Joined: Mon Feb 16, 2009 5:12 am
Location: Greenwood MS USA

Re: Volunteering to fix F@H Viewer

Postby dfgirl12 » Mon Mar 16, 2020 10:35 pm

The viewer works in Linux (CPU: GRO_A7, and GPU: Core_21 and Core_22 all work on Linux), just not in Windows.

Searching the forum or looking at the issues in GitHub, you should be able to find more info on it. Like:
https://github.com/FoldingAtHome/fah-is ... pen+viewer
https://github.com/FoldingAtHome/fah-viewer/issues/5
https://github.com/FoldingAtHome/fah-control/issues/31
viewtopic.php?f=96&t=29258
Some recent general info on volunteering to help with the open source parts of FAH: viewtopic.php?f=15&t=32315

It would be awesome to finally get the viewer to work on Windows again. Any help would be appreciated :D
dfgirl12
 
Posts: 38
Joined: Fri Aug 21, 2009 9:34 am

Re: Volunteering to fix F@H Viewer

Postby dfgirl12 » Mon Mar 16, 2020 11:04 pm

There was a workaround mentioned recently: https://github.com/FoldingAtHome/fah-issues/issues/1283 by using the previous viewer executable (and a required dll) file with the current FAH v7.5.1 installation for viewing Core_22 on Windows
dfgirl12
 
Posts: 38
Joined: Fri Aug 21, 2009 9:34 am

Re: Volunteering to fix F@H Viewer

Postby Quantize » Thu Mar 19, 2020 9:05 am


Thank you dfgirl12!

You actually saved me a lot of time and pain because I assumed the handful of issues (7) noted at the 'fah-viewer' repo to be ALL.
I'm now much more up to speed and know where to resume this bughunt.
Quantize
 
Posts: 26
Joined: Mon Mar 16, 2020 2:15 pm

Re: Volunteering to fix F@H Viewer

Postby kostuek » Thu Mar 19, 2020 10:13 am

dfgirl12 wrote:The viewer works in Linux (CPU: GRO_A7, and GPU: Core_21 and Core_22 all work on Linux), just not in Windows.


GPU viewer does not work properly in Ubuntu 18.04 for me (graphic errors and disappearing molecule). CPU is fine.
kostuek
 
Posts: 32
Joined: Tue Mar 17, 2020 12:03 pm

Re: Volunteering to fix F@H Viewer

Postby Quantize » Sat Mar 21, 2020 3:25 pm

RESULTS so far:

Modern GPU WU's cannot be viewed as the simulation values that are relayed are vastly different than CPU WU's.

This behaviour is not yet supported and after a while will crash the viewer: https://meertronix.com/WhyGPUWasDisabled.png
It is possible that connectivity was intentionally severed to prevent this crash and to save time not having to write a method to ignore incompatible GPU slots while keeping the right ones.

Naturally, I've made a couple of changes and have compiled a new viewer.
This version should successfully connect and does not attempt to view GPU slots at all.

Kindly download it here and let me know if this works for you too:
http://meertronix.com/FAHViewer.exe

Should there still be any GPU cores that are actually compatible, I can also make it so that those are still viewed.
But we'd need to find out which cores will work because I have no idea how many variants there are and which ones should be selected.

Also I haven't yet looked at the GPU Trajectory wrapper thing.
Last edited by Quantize on Sat Mar 21, 2020 3:54 pm, edited 4 times in total.
Quantize
 
Posts: 26
Joined: Mon Mar 16, 2020 2:15 pm

Re: Volunteering to fix F@H Viewer

Postby Jesse_V » Sat Mar 21, 2020 3:49 pm

If it would help your codebase, you are welcome to take any OpenGL code from a tool I wrote long ago: https://github.com/Jesse-V/Folding-Atomata, maybe it would help.
F@h is now the top computing platform on the planet and nothing unites people like a dedicated fight against a common enemy. This virus affects all of us. Lets end it together.
Jesse_V
Site Moderator
 
Posts: 2851
Joined: Mon Jul 18, 2011 5:44 am
Location: Western Washington

Re: Volunteering to fix F@H Viewer

Postby Quantize » Sat Mar 21, 2020 4:20 pm

Well thank you that sounds very generous.

Thing is it's not my codebase and also I'm not that great with graphical thinking and math. I code, I bughunt, I fix. I help out where others can't or don't.
My primary objective here was to get the viewer connecting with the WU's again as it has only been showing me a demo proteïne ever since I started folding.

If I had the biophysiological knowhow you guys have then I'd be in the lab with you finding the cure and I wouldn't bother fixing the viewer either.
But alas hopefully I can contribute something to F@H & COVID research in this way, apart from my own cycles ofcourse.

Imo the viewer serves a significant promotional function and will remain in people's minds for a long time even after COVID-19.
Right now loads and loads of people are coming so it's F@H's time to shine and have a stunning first impression on those people. First impressions are everything.

Please keep up the good work and let me know whether my work on the viewer is appreciated or if I should go do something else against COVID-19.
Quantize
 
Posts: 26
Joined: Mon Mar 16, 2020 2:15 pm

Re: Volunteering to fix F@H Viewer

Postby str4d » Sun Mar 22, 2020 11:37 am

Quantize wrote:Imo the viewer serves a significant promotional function and will remain in people's minds for a long time even after COVID-19.
Right now loads and loads of people are coming so it's F@H's time to shine and have a stunning first impression on those people. First impressions are everything

I agree! It's why I started helping with FAHViewer in the last few days as well :) (though it turned out the bug I went in to fix had already been resolved!)

I think the best next step would be to open a PR on FoldingAtHome/fah-viewer with your current changes, and we can discuss them there and in FoldingAtHome/fah-viewer/issues/5. I've made several PRs myself in the last few days, and the maintainer has been very quick to merge (though they were more cursory changes, such as setting up a CI workflow to exercise the build process).
str4d
 
Posts: 2
Joined: Sun Mar 22, 2020 11:24 am

Re: Volunteering to fix F@H Viewer

Postby Quantize » Mon Mar 23, 2020 2:12 am

str4d wrote:I agree! It's why I started helping with FAHViewer in the last few days as well :) (though it turned out the bug I went in to fix had already been resolved!)

Hi! I'm not yet completely sure what you're referring to here. What did you go in to fix?
In the last release the viewer only works on demo mode.
It would seem something has gone horribly wrong in the process of compiling the latest version, winsock and all.
That, or it was intentionally hardcoded into demo mode awaiting an update for GPU trajectory support and was not communicated with git.
str4d wrote:I think the best next step would be to open a PR on FoldingAtHome/fah-viewer with your current changes, and we can discuss them there and in FoldingAtHome/fah-viewer/issues/5. I've made several PRs myself in the last few days, and the maintainer has been very quick to merge (though they were more cursory changes, such as setting up a CI workflow to exercise the build process).

For a while I thought those 6 issues were the only issues with the viewer and thus was not properly brought up to speed.
Good thing you're now looking after this.

May I ask why can't all repositories have their own issue page?
In what way exactly does throwing everything onto one huge heap create order & overview instead of .. well, one huge heap? :'D
Quantize
 
Posts: 26
Joined: Mon Mar 16, 2020 2:15 pm

Next

Return to V7.5.1 Public Release Windows/Linux/MacOS X

Who is online

Users browsing this forum: No registered users and 0 guests

cron