1336 - My Design Crash Happy [Submitted to Scientists]

Case number:845829-2003417
Opened by:Formula350
Opened on:Saturday, February 11, 2017 - 04:15
Last modified:Tuesday, February 21, 2017 - 21:46

Partway through 1336's run I changed up my design, introducing sheets, and seems that the two on the end of the protein would sometimes -- if all the mysterious conditions were just right -- cause FoldIt to crash if I messed with it. I never was able to nail down the cause to anything that would reproduce it with any regularity; however, after it expired while I was getting it ready to submit, I managed to crash my client 6 or 7 times...

I left notes on it for the way I had everything setup and what I would generally be doing when it would crash, and named it ! CAN CAUSE CRASHES !, then submitted it to Scientists. (As per Susume's suggestion when I had mentioned it a couple days earlier.)

After the last crash, I had at least gotten it submitted beforehand, but I just gave up and never bothered to submit my best afterwards (which was only 10-15pnts higher), so feel free to also consider that my actual submission too, if it is intriguing enough! :P


(Sat, 02/11/2017 - 04:15  |  7 comments)

Joined: 09/29/2016
Groups: Gargleblasters
Type: Suggestion » Bug

Oh I didn't notice that I had to also select "Bug" for type. Definitely wasn't meant to be a "suggestion"... Whoops! heh

Joined: 09/29/2016
Groups: Gargleblasters

Possible update...

While working in another BP puzzle (1342), again doing sheet work, and if I recall correctly... again with Parallel Bonds present, I have crashed twice in the last couple hours.

Once more it seems to be primarily when I'm rotating the solution whilst a tool is running (wiggle), but this time it seems to happen when the Parallel Bonds are forming. Whether or not these are coincidences occurring at the same time as the underlying problem causes the crash, I can't say. It does seem slightly too coincidental though.

I suppose I could always try turning off the bond animation and see if it'll happen (at the increased complexity of folding heh)

Joined: 09/29/2016
Groups: Gargleblasters

:| All bond visuals disabled and I crashed in under a minute...

Joined: 09/29/2016
Groups: Gargleblasters

UPDATE... I have created a consistent and repeatable means of crashing the client with my 1343 design which I have shared under the name "CLIENT CRASHING B-HELIX". I have added the steps to the Share as well, but here they are as well:

1) Whether this has anything to do with it, I play in Classic Interface, once you load my Share switch to Selection Interface.
2) Select Segment 65
3) Hold Shift and Select Segment 58
4) Click on the GUI option to "Delete the segments(s)!" and... *CRASH*

When this happened I immediately loaded the puzzle back up and those were my only steps I performed to crash the client, I did _not_ rotate the design at all or run any other tools first.

My computer:
HP ENVY x360 laptop (touchsceen... might make a diff)
OS: Windows 10 Home x64 Version 1511 (Build 10586.218) [Obtained via running 'winver.exe']
CPU: AMD FX-9800P Quad-core with integrated Radeon R7 graphics. (Driver: Radeon Crimson 16.7.3 WHQL)
RAM: 7.5GB (512MB dedicated to GPU) - ~4.45GB in use (Foldit running, 4.1GB not running)
HDD: 1TB HGST (Hitachi)
FoldIt: 20170210-55c03a4ae1-win_x86 [FoldIt isn't 64bit?]

Any other details needed please let me know.

Susume's picture
User offline. Last seen 21 hours 21 min ago. Offline
Joined: 10/02/2011

There is a known bug that if the rama map is open, deleting the last segment crashes the client. Does yours still crash when the rama map is closed?

Joined: 09/29/2016
Groups: Gargleblasters

I don't play with the Rama Map open. You can start up client, log in, load puzzle 1342, load my share, and follow the above steps to crash it.

In short, yes it crashes with Rama closed.

The test would actually be, in my case, whether it still crashes with Rama open! heh

I simply can't even play with my design on 1342 because it WILL crash me within roughly 15 mins unless I keep the Sheets frozen and banded together, which it'll last a little longer. Or if I refrain from doing a global Wiggle.

actiasluna's picture
User offline. Last seen 5 hours 15 min ago. Offline
Joined: 03/05/2015
Groups: Gargleblasters

I duplicated that issue on my Windows 10 (running on external HD on iMac hardware). FWIW I've always had problems cutting more than one seg at the end, particularly in these monomers (but that was previous mac client so YMMV as far as that's concerned.)


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, Amazon, Microsoft, Adobe, RosettaCommons