How long can structures be held?

[eXXile]

I have a few machines nonetting, which can't get a connection to the net. When we pass the finish mark for this protein, if I upload the structures will it count and will I just need to download a new protein?

[KWSN_Millennium2001Guy]

I believe that once a protein is completed that the uploaded work units are discarded and the autoupdate function downloads the next protein for you. You then begin working on the next protein. I have heard that you might still get credit for that old work you completed, but I haven't experienced that myself. Because of the way that I am set up, I have been deleting the old installations and installing the client again during each update.

Ni!

[eXXile]

What if I have a single machine that has been updated with the new protein, and I just take the protein information from that machine? If the protein.trj file that holds the protein information, could this just be transferred between the machines on nonet?

[KWSN_Millennium2001Guy]

Excellent question, but I don't know the answer. I have been treating Howard's installation files as complete sets of files. I don't think it is a good idea to just move particular individual files around unless Howard explicitly gives us instructions on an acceptable methodology.

Ni!

[eXXile]

Originally posted by KWSN_Millennium2001Guy
Excellent question, but I don't know the answer. I have been treating Howard's installation files as complete sets of files. I don't think it is a good idea to just move particular individual files around unless Howard explicitly gives us instructions on an acceptable methodology.

Ni!
Thanks for helping out a noob.

[KWSN_Millennium2001Guy]

After this update, I now can say that once a protein is finished, you will not get credit for any leftover workunits when you try to upload them. It will just prompt you to get the next update and will discard the old work units.

Ni!

[FoBoT]

Originally posted by eXXile
If the protein.trj file that holds the protein information,

no, that is the source file, not the output, the output is all the .log.bz2 and .val.bz2 files

if you need to move stuff to upload, move the WHOLE folder , don't take a chance of messing it up

also, once a protein is done, the server WILL NOT accept the old work, you get no credit and the client won't clear it out, you have to manually delete the old .log.bz2 and .val.bz2 files yourself, if you want to reclaim the disk space, if you ignore them, no harm , just get cluttered

people nonetting MUST watch the progress and when the PROJECT approaches the time to start on a new protein, the person nonetting needs to upload cached work BEFORE the switch, otherwise its lost cycles/work

[Brian the Fist]

Please do not move around individual files to try to manually update machines not connected to the net. If you have such machines, please either update a machine that IS on the net and then copy the entire client directory onto your otehr machines, or download the complete package again from the web site, once, and put this fresh copy on all your unnetworked machines.

I cannot ever say for sure what files will be updated between proteins for certains, but at a MINIMUM, protein.trj AND native.val are both updated.

If you do not follow these instructions, you risk having the server discard all your new results.