Remaining issues with Experimental?

Case number:954892-2004857
Topic:Developer Preview
Opened by:jflat06
Status:Open
Type:Question
Opened on:Thursday, March 1, 2018 - 22:05
Last modified:Saturday, March 10, 2018 - 18:53

Hey everyone, I've gone over all the feedbacks here (thanks for the reports!), and I am wondering what problems people are still having after the most recent Experimental update.

Any testing help is greatly appreciated!

(Thu, 03/01/2018 - 22:05  |  10 comments)


LociOiling's picture
User offline. Last seen 15 min 40 sec ago. Offline
Joined: 12/27/2012
Groups: Beta Folders

Looking pretty good overall. I was able to run the symmetry puzzle successfully on both my previous save and a new start from initial pose. No problems noted in either.

I was able to get an "asymmetry introduced by ActionLocalMinimize" by transforming the initial symmetry puzzle pose to all sheet, idealizing the secondary structure, and using the tweak tool to straighten. I added a cutpoint and aligned the sheet in an unlikely fashion. I'm guessing the crash occurred on a wiggle.

I've attached debug.txt and log.txt.

LociOiling's picture
User offline. Last seen 15 min 40 sec ago. Offline
Joined: 12/27/2012
Groups: Beta Folders

The problem symmetry puzzle solution has been shared with scientists as "crashed for me". A wiggle on auto wiggle power should do the trick.

jflat06's picture
User offline. Last seen 1 day 20 hours ago. Offline
Joined: 09/29/2010
Groups: Window Group

Thanks for the feedback - I think that problem is independent from the (worse) crashes that started plaguing symmetry. We'll look into it!

bertro's picture
User offline. Last seen 15 hours 44 min ago. Offline
Joined: 05/02/2011
Groups: Beta Folders

On the symmetry HBond Networks puzzle, wiggle definitely seems faster. But I find that mutate(1) is very expensive, it takes over 3 minutes to complete one iteration in a script.

jflat06's picture
User offline. Last seen 1 day 20 hours ago. Offline
Joined: 09/29/2010
Groups: Window Group

Hey Bertro, thanks for the feedback.

Is mutate performing worse after this change?

bertro's picture
User offline. Last seen 15 hours 44 min ago. Offline
Joined: 05/02/2011
Groups: Beta Folders

I have one un-updated experimental client left. Let me see if I can compare...

bertro's picture
User offline. Last seen 15 hours 44 min ago. Offline
Joined: 05/02/2011
Groups: Beta Folders

Sorry,I had already converted it to devprev :(

Cannot compare anymore.

LociOiling's picture
User offline. Last seen 15 min 40 sec ago. Offline
Joined: 12/27/2012
Groups: Beta Folders

The "exp test" puzzle is a little goofy, see http://fold.it/portal/node/994559#comment-36370

Remix and idealize SS don't work, but rebuild does, so it's still playable.

LociOiling's picture
User offline. Last seen 15 min 40 sec ago. Offline
Joined: 12/27/2012
Groups: Beta Folders

Still on "exp test", seeing a large log file, 44 MB, over 250K lines with lots of lines like these:

core.kinematics.AtomTree: [ ERROR ] No proper DoF can be found for these four atoms: 65-2, 65-3, 66-1, 66-2!
core.optimization.LineMinimizer: [ ERROR ] Inaccurate G! step= 3.8147e-08 Deriv= -18.3154 Finite Diff= -1.64748
core.optimization.AtomTreeMinimizer: [ WARNING ] ***************************************************************
core.optimization.AtomTreeMinimizer: [ WARNING ] ** WARNING: Non-ideal minimization used with a ScoreFunction **
core.optimization.AtomTreeMinimizer: [ WARNING ] ** which isn't set up for non-ideal minimization **
core.optimization.AtomTreeMinimizer: [ WARNING ] ***************************************************************

I've tried attaching the log file, but apparently it's too big.

jflat06's picture
User offline. Last seen 1 day 20 hours ago. Offline
Joined: 09/29/2010
Groups: Window Group

Actually, it looks like that puzzle is left over from a LONG time ago. It uses some features which aren't supported anymore.

It looks like someone just forgot to take it down, I'll do that now.

Sitemap

Developed by: UW Center for Game Science, UW Institute for Protein Design, Northeastern University, Vanderbilt University Meiler Lab, UC Davis
Supported by: DARPA, NSF, NIH, HHMI, Microsoft, Adobe, RosettaCommons