Volunteering to fix F@H Viewer

Moderators: Site Moderators, FAHC Science Team

ChristianVirtual
Posts: 1576
Joined: Tue May 28, 2013 12:14 pm
Location: Tokyo

Re: Volunteering to fix F@H Viewer

Post by ChristianVirtual »

Jorgeminator wrote:The water molecules are not the only problem when viewing GPU projects. Only carbon, oxygen and sulphur atoms are shown properly at the moment. Hydrogen and nitrogen atoms in the protein are not recognized. We must wait for an updated core for these to work.
Tell me about waiting ... I mentioned it few years back but as noted earlier the limited resources prevented a fix. I miss the beauty of proteins
ImageImage
Please contribute your logs to http://ppd.fahmm.net
rplate
Posts: 8
Joined: Mon Jan 07, 2013 4:22 am
Hardware configuration: Windows 10 || CPU: Intel i5-6500 3.20 GHz || GPU: AMD Radeon R9 290

Re: Volunteering to fix F@H Viewer

Post by rplate »

Quantize wrote:I'm happy to report some nice progress on FAHViewer.

Download it here: http://www.meertronix.com/FAHViewer_beta0.exe

Despite thorough testing know that this release is coming fresh from the factory and may contain a bug or two.
In its current state it's working flawlessly for me whatever I throw at it. YMMV
Working well now. Downloaded it and it crashed after a few seconds and was then unable to reopen with an error that I don't fully recall. If it occurs again I'll copy it in.

Fixed the glitch by downloading your previous version from a few replies ago and it worked fine, and now both do.

Thanks to all for taking this issue up again and hopefully fixing the coolest screensaver I've seen!
Windows 10 || CPU: Intel i5-6500 3.20 GHz || GPU: AMD Radeon R9 290
Quantize
Posts: 26
Joined: Mon Mar 16, 2020 1:15 pm

Re: Volunteering to fix F@H Viewer

Post by Quantize »

I'm happy to report some nice progress on FAHViewer.

The bad:
Still no OPENMM_22 support

The bandaid:
- Auto switch to compatible slot on startup
- If no compatible slot is available while a 0x22 core is running then:
show its progress info while stating "INCOMPAT" at the protein label
- Remove empty slots from ] and [ slot cycling ("READY"-state slots)

The good:
- Increased base fps from 8 to 12
- Turbo rendering mode: press 't' to increase your fps to 50-60 for a nice demo
for optimal results you'll need to pause your CPU work for a bit
- Adjusted initial rotation speed for optimal folding visibility
- You can now toggle atom wiggling with 'w'
- You can now toggle rotation on/off with 'r'
(ofcourse this is in addition to the regular arrow keys for increasing/decreasing)
- Help and about windows can now be closed with the same key

To-do:
- Make F@H logo clickable so that it opens up the home page and/or forum

Download it here: http://www.meertronix.com/FAHViewer_beta0.exe

Despite thorough testing know that this release is coming fresh from the factory and may contain a bug or two.
In its current state it's working flawlessly for me whatever I throw at it. YMMV
Last edited by Quantize on Tue Mar 31, 2020 12:05 pm, edited 2 times in total.
Quantize
Posts: 26
Joined: Mon Mar 16, 2020 1:15 pm

Re: Volunteering to fix F@H Viewer

Post by Quantize »

rplate wrote: Working well now. Downloaded it and it crashed after a few seconds and was then unable to reopen with an error that I don't fully recall. If it occurs again I'll copy it in.

Fixed the glitch by downloading your previous version from a few replies ago and it worked fine, and now both do.
Those two binaries are 100% exact copies of the same. Must've been your AV.
Modern virusscanners will interfere with the first time running a foreign .exe, e.g. Avast CyberCapture.

You're welcome.
RawwrBag
Posts: 1
Joined: Wed Apr 01, 2020 7:18 am

Re: Volunteering to fix F@H Viewer

Post by RawwrBag »

Hi Quantize, I was hoping to follow along on your fork, but I can't seem to find it. Can you link me to your github? Thx.
LaPeche35
Posts: 4
Joined: Fri Mar 27, 2020 3:33 pm

Re: Volunteering to fix F@H Viewer

Post by LaPeche35 »

Quantize wrote:I'm happy to report some nice progress on FAHViewer.

The bad:
Still no OPENMM_22 support
...
I stopped read at this line. Did you check at str4d's work ? It supports OPENMM22
Please read https://github.com/FoldingAtHome/fah-is ... -602024280
and https://github.com/FoldingAtHome/fah-is ... -602115211
Its issue is that displays the whole water. Just need to fix that point.
bruce
Posts: 20824
Joined: Thu Nov 29, 2007 10:13 pm
Location: So. Cal.

Re: Volunteering to fix F@H Viewer

Post by bruce »

I have not looked at the code or tried the test versions.

I do know that there has been some confusion about how and where the FAHCore writes the images. Since FAHCore_21 and FAHCore 22 were developed semi-independently and probably not tested against the viewer, these inconsistencies will pose a challenge to display both FAHCore_22 snapshots and FAHCore_21/FAHCore_a7 snapshots. Obviously if the viewer tries to find images in each location, it presents a challenge to trying to fix the problem in the viewer.

IMHO, the problem should be fixed in FAHCore_22. having it write (and upload) the same sort of data that the other FAHCores write.

Then, too, I understand the Linux/OS-X fiewers are different than the Windows viewer. the executives need to impose a consistent set of standards on all the applicable software so that it all becomes interchangeable.

Item 2: The water filter does make the water transparent so you can see the protein ... but the purple haze is a modified water and the filter needs to recognize it as ordinary water.
Jorgeminator
Posts: 49
Joined: Tue Mar 24, 2020 11:24 am
Location: Finland

Re: Volunteering to fix F@H Viewer

Post by Jorgeminator »

LaPeche35 wrote:
Quantize wrote:I'm happy to report some nice progress on FAHViewer.

The bad:
Still no OPENMM_22 support
...
I stopped read at this line. Did you check at str4d's work ? It supports OPENMM22
Please read https://github.com/FoldingAtHome/fah-is ... -602024280
and https://github.com/FoldingAtHome/fah-is ... -602115211
Its issue is that displays the whole water. Just need to fix that point.
It's not just the water that is the problem. Part of the protein is also misrecognized.
Image
confused.brit
Posts: 9
Joined: Wed Apr 01, 2020 12:43 pm

Re: Volunteering to fix F@H Viewer

Post by confused.brit »

Great work Meerteen, i just got back into folding after stopping when they closed the ps3 project, and I missed seeing my protein sims wiggle. :)
Image
Quantize
Posts: 26
Joined: Mon Mar 16, 2020 1:15 pm

Re: Volunteering to fix F@H Viewer

Post by Quantize »

LaPeche35 wrote: I stopped read at this line.
Good to know.
Next time I'll start with the good. :lol:
Quantize
Posts: 26
Joined: Mon Mar 16, 2020 1:15 pm

Re: Volunteering to fix F@H Viewer

Post by Quantize »

confused.brit wrote:Great work Meerteen, i just got back into folding after stopping when they closed the ps3 project, and I missed seeing my protein sims wiggle. :)
Thank you!!
Xavierthe15
Posts: 4
Joined: Wed Jul 18, 2018 7:42 am

Re: Volunteering to fix F@H Viewer

Post by Xavierthe15 »

You may want to contact Joseph Coffland at Cauldron Development LLC to collaborate with him. That company is the originator for not just the visualizer but the F@H software as a whole. I'm about ready to put together a fundraising effort for that company just to get this long-requested issue over and done with. lol
Quantize
Posts: 26
Joined: Mon Mar 16, 2020 1:15 pm

Re: Volunteering to fix F@H Viewer

Post by Quantize »

RawwrBag wrote:Hi Quantize, I was hoping to follow along on your fork, but I can't seem to find it.
Ahoy brother. Good to have you here with us.
I'll be shooting you a quick PM first thing when I push the code.
LaPeche35 wrote:Did you check at str4d's work ? It supports OPENMM22
Yes ofcourse I have. Str4d has managed to compile the ol' codebase just like I did. And he put up a nice CI system on GitHub.
And now he's carefully documented it so that noone else has to spend the wee hours that we did to get the thing to actually compile.
These things are very commendable of him as it enables others to more easily join in on the project.

Other than that str4d hasn't been making any changes to the code itself recently that I know of, and thus 'his' version does not "support" OPENMM_22 any more than 'mine' does.
I expressly disabled 22 on 'mine' because the malformed data commonly crashes the app and tbh the purple cube looks quite meaningless to me. (it's also missing all nitrogen atoms)
And I bet that the product owner would prefer scientific accuracy over random abstract art aswell.

Really, we gotta get that core fixed / interpreted, and then we can talk perfect viewer.
Which is the very next step as far as I'm concerned.
ChristianVirtual wrote:I miss the beauty of proteins
Since CPU and older GPU cores are now totally working idk what exactly you're trying to add to the discussion with this.
Are you in the arena or not. I've seen your FAH_WrapperGPUTrajectory and it looks quite impressive tbh, y'have skill.

So now: https://github.com/FoldingAtHome/openmm
Where tf do we start lol.
Xavierthe15 wrote:You may want to contact Joseph Coffland at Cauldron Development LLC to collaborate with him.
Hi there! Yes I've seen mr. Coffland's name all over the place and I'm about to push this update to his GitHub.
Hoping that it will be received well, as he and I haven't worked with each other before.

Regarding your fundraising idea: While there may not be much of an excuse for years and years of delay, right now his support may be fully needed to keep F@H running smoothly.
I'm not exactly sure what throwing money in his direction at this exact COVID-19 time would achieve and if that would help or hurt the greater cause.

On the other hand I do think it may effect a whopping -7.5% churn rate and +2.5% of new members whenever the viewer is working.
So that's like.. another 10% directly: ~+150 pflops. That's equal to one supercomputer, literally.

Coming on top of some other long-term branding effects that are somewhat harder to put into numbers.

Thing is: given the nature of software debugging, releasing a core update risks destabilizing the machinery.
And if such were to happen while we're all folding at 1.5 exaflops then everyone's gonna regret that decision.

Thus we may want to brace for a "don't touch that!!" attitude right now, perhaps rightfully so.
By all means correct me if I'm wrong ofcourse, this is just the way I see it right now.
Last edited by Quantize on Sun Apr 05, 2020 2:22 pm, edited 1 time in total.
ChristianVirtual
Posts: 1576
Joined: Tue May 28, 2013 12:14 pm
Location: Tokyo

Re: Volunteering to fix F@H Viewer

Post by ChristianVirtual »

Quantize wrote:
ChristianVirtual wrote:I miss the beauty of proteins
Since CPU and older GPU cores are now totally working idk what exactly you're trying to add to the discussion with this.
Are you in the arena or not. I've seen your FAH_WrapperGPUTrajectory and it looks quite impressive tbh, y'have skill
Thank you,

I also made a viewer on iPhone / iPad ( https://vimeo.com/185187797 ) but lost my drive once the data delivered from newer Core/client combinations via remote interface broke and never got fixed. For me the visualization was always important and pushing motivation. And easy entry to explain to people what folding (somehow) does. Just looking over the fence into the arena and see if there is some improvement in that domain. But as it looks some fundamentals I would need to proceed are still on low priority bench. :?
ImageImage
Please contribute your logs to http://ppd.fahmm.net
Tohya
Posts: 48
Joined: Thu Feb 07, 2008 12:41 am

Re: Volunteering to fix F@H Viewer

Post by Tohya »

https://twitter.com/foldingathome/statu ... 3412453378
Interested in helping us develop and enhance the F@h architecture? Come join the development team at our F@h Fireside Chat on Thursday, April 9th, 4-5pm EDT! To receive a Discord invite link, fill out this form: https://tinyurl.com/firesidedev
Post Reply