(unresolved) Compound window says 'running' for many hours and does not produce result.

Started by Bletchley Park

Bletchley Park Lv 1

Compound windows says 'running' for many hours and does not produce result.
V21-20230206-cb39f4bed1-win_x64-devprev
puzzle 2264

Bletchley Park Lv 1

The initial pose has eventually after several days resolved.
I now have three uploaded solutions that have been running for hours and still show 'running'
Maybe show a queue length ('you're the 15th in the list', waiting time is … ) for example.

BootsMcGraw Lv 1

This is for puzzle 2267, and not just for BP. Other players, including me, are experiencing it.

Please extend the deadline for 2267 by another day. It is unfair to expect us to play with broken tools.

Sandrix72 Lv 1

I have no clue what is going on. :(
The new puzzle do not appears and the compound library is only "running". In this way it is impossible to improve the score.

rmoretti Staff Lv 1

We're using the ZINC server (at docking.org) to do the compound library search. It's currently down at the moment, but hopefully should be back up when the sysadmins there get a chance to look at it. Once it's back up, the pending queue should be taken care of relatively quickly.

We'll keep monitoring the situation, and adjust the closing date as appropriate.

Bletchley Park Lv 1

Why not copy the repository and run this locally ? Credit the original authors and make sure you are not dependant on someone else's server. It is not working again right now.

rmoretti Staff Lv 1

It's actually a rather complex problem to do compound similarity searching over 30+ billion compounds, and to keep things up-to-date with the newest releases of the ultra large libraries (which are continually being added to and updated). The Irwin and Shoichet labs do a rather remarkable job of hosting the servers we're using.

If we become a burden to them, or if we do have sustained issues with uptime, we may consider looking into self-hosting the search. However, there's a non-trivial amount of effort to get that up and running and to maintain it. The sysadmins at ZINC have a lot more experience with this than we do. We'd probably have more downtime for a self-hosted solution rather than the ZINC server. (E.g. think about how many teething pains we've had with the AlphaFold server.) It sucks that it's down, but I'd encourage a bit of patience here.

jeff101 Lv 1

In ligand design puzzle 2301 yesterday, I sent several compounds called J5 to J9 to the compound library at least once each. The first time I sent them, each ran for many hours and eventually gave "error". The next time I sent them, each gave "error" right away, as if Foldit recognized they were tried before, so it just repeated the result it had given before. Does "error" mean compounds J5-J9 are not in the compound library? How can I get the library to tell me the right answer instead of "error" every time? Do I have to restart my Foldit client so it will really send compounds J5-J9 to the library again rather than finding its previous result and just repeating it to me? Do I have to remove or edit some local file so my Foldit client can't find its previous results for J5-J9? Please help.

If you're on the Foldit staff, can you find my 2301 J5-J9 shares w/scientists, send them to the library yourself, and then e-mail me a list of which ones are in the library, which ones are not, and how many library hits each one gives?

Thanks!