41) Message boards : Number crunching : ATLAS Simulation. (Message 30955)
Posted 22 Jun 2017 by Profile adrianxw
Post:
I have an ATLAS simulation task running on my system. I was alerted to the issue I'm seeing because I could see a lot of waiting to run tasks, and eventually came across the ATLAS, which says it is running on 8 CPU's, ie. all of them, (multithreaded). I have removed this task class from my profile, as I am not happy with the way it works. Question is, is that the only task class which does this, or am I better off just dropping LHC from my projects list?
42) Message boards : Number crunching : Weird error message. (Message 28339)
Posted 4 Jan 2017 by Profile adrianxw
Post:
There is nothing wrong with my network connection, obviously, if it is some link to a required external resource that is missing, well, nothing I can do about that is there. My hiatus may resume.
43) Message boards : Number crunching : Weird error message. (Message 28337)
Posted 4 Jan 2017 by Profile adrianxw
Post:
I re-enabled crunching here after a hiatus, and downloaded a large work unit, my first for some years. Big download, but it went okay. The work unit started running and went merrily on its way for just over a minute when it crashed out with a computation error. The reason puzzled me...

Exit status -203 (0xFFFFFF35) ERR_NO_NETWORK_CONNECTION

... about one minute after it had just downloaded the thing?
44) Message boards : Number crunching : Invalid tasks (Message 26345)
Posted 12 Apr 2014 by Profile adrianxw
Post:
The lost time has resulted in my setting no new tasks. If they know the issue is there, why are they still sending jobs to Windows machines. My machines could have been doing useful work for someone else.
45) Message boards : Number crunching : Upload Errors (Message 25984)
Posted 23 Oct 2013 by Profile adrianxw
Post:
@henry

You've lost me now. My sig? My sig is from the lyrics of the Pink Floyd track "Sheep" on the Animals album.
46) Message boards : Number crunching : Upload Errors (Message 25972)
Posted 23 Oct 2013 by Profile adrianxw
Post:
To help vent frustration?
47) Message boards : Number crunching : Upload Errors (Message 25964)
Posted 23 Oct 2013 by Profile adrianxw
Post:
Usually though, a scientist sees an event and learns from it. The problem we are having now is an old "friend". Look at this thread for example, from a few months back.

If they need a webmaster, I'm free.
48) Message boards : Number crunching : Upload Errors (Message 25950)
Posted 23 Oct 2013 by Profile adrianxw
Post:
My question was: if they aren't available who else might help us?


Help us? The question surely should be, if they can't even cope with running a web server, why should any of us be helping them.
49) Message boards : Number crunching : Database Error (Message 23531)
Posted 16 Oct 2011 by Profile adrianxw
Post:
If I go to My Account page and then look at Tasks, the page with my current and recently completed tasks listed shows, looks normal.

However, immediately above the list, well the box that contains the list, under the LHC@Home logo "Database Error" is shown.

Probably not significant, but hey.
50) Message boards : Number crunching : In my team, but not in my team. (Message 23331)
Posted 3 Oct 2011 by Profile adrianxw
Post:
Indeed so. Cheers.
51) Message boards : Number crunching : In my team, but not in my team. (Message 23319)
Posted 2 Oct 2011 by Profile adrianxw
Post:
If I do that, will I still be the founder when I return?
52) Message boards : Number crunching : In my team, but not in my team. (Message 23316)
Posted 2 Oct 2011 by Profile adrianxw
Post:
I appear to be in a state of flux team wise. If I look at my account, I can see my credit etc. and that I am a member, indeed, the founder of my team. Yet I do not appear in the teams stats. On my personal side, I see I am...

>>> Founder but not member of

... somewhat odd. I can see my team mates there.
53) Message boards : Number crunching : Another shut down (Message 22030)
Posted 10 Mar 2010 by Profile adrianxw
Post:
http://news.bbc.co.uk/2/hi/science/nature/8556621.stm

I was really hoping to see something from this machine, but I\'ll probably be dead before we get there.
54) Message boards : Number crunching : Ok...who got lucky? (Message 21989)
Posted 7 Mar 2010 by Profile adrianxw
Post:
I got 2 wu's 6th March, one complete and credited, the other still pending. It will be good to see this old guy running again.
55) Message boards : Number crunching : Initial Replication (Message 19282)
Posted 20 Mar 2008 by Profile adrianxw
Post:
IOW, this situation is one that could "fall through the cracks" no matter what is done with the server-side aborts.

But if IR = Q it would never have been sent to 2 of the 5 machines. Thus 2 of the machines have crunched the wu for no good reason. If any of the original replication machines had failed, then send it out again. A lot of projects do that, often with a shorter then usual deadline. Some, SIMAP for example, send failed units out again to "known good hosts" that have a proven track record of returning good work promptly, (I receive a good number of those myself).

I do consider it to be important as I consider the squandering of a freely given resource which could have been doing something more productive to be, at best, arrogant, others would probably choose other terminology.

You and I are unlikely ever to agree on this point. You have behaved in a generally civil manner however and I thank you for the debate.

56) Message boards : Number crunching : Initial Replication (Message 19277)
Posted 19 Mar 2008 by Profile adrianxw
Post:
And now a forth machine has crunched the unit. 31,000+ wasted crunching.
57) Message boards : Number crunching : Initial Replication (Message 19275)
Posted 19 Mar 2008 by Profile adrianxw
Post:
it would seem that lobbying Alex and Neasan will not have a direct impact as all they're going to be able to do is mention it to their boss.

... and thus it has been with QM running the show, and was before when it was run from Geneva. As I said, I have been around here for a long time, a lot longer then you possibly, and with a project like CERN, the politics and empires take precedence over actually doing anything. Hence my one liner...
58) Message boards : Number crunching : Initial Replication (Message 19272)
Posted 19 Mar 2008 by Profile adrianxw
Post:
Yes, I am aware of the features of the newer servers, and yes, I have perhaps not the absolute latest BOINC core on all machines, but within 1 or 2 roemmended minors, 5.10.28 on this one, 5.10.45 on my Vista system etc. It brings me back to the first sentence of the first post I added to this thread...

Since when has lobbying acheived anything around here?

... and I've been around here for a while.

I'd vote with my feet, but that isn't going to send a message to the admins. When they have as many CPU's out here as they need, and everyone fighting to get the work when it comes, a few dozen, hundred, maybe thousand leaving the project is un-noticed.

To answer your Rosetta question, I just hopped onto Rosetta and edited my preferences to see what I could and couldn't set for my run time. The shortest is 1 hour, the longest 1 day [sic] i.e. not 24 hours, (I wonder what happens at daylight savings start/end...).

As I said before, there is no such thing as a "complete" work unit if you mean to "fold protein x". The way the algorithm works is that an amino acid, (AA), sequence is sent with a random, (or maybe a guess based on other information), start configuration.

The configuration of the AA residues, (angles/rotations between individual AA residues), is then altered and the overall free energy of the resulting fold compared to the original, if it is better, another change is done, if worse, it is undone and a different random change is done. This process continues until the lowest energy configuration has been found for that AA sequence, that start point and a random first move. That constitutes a "decoy" in Rosetta terminology. The Rosetta server side then compares the decoys and issues new start points, or takes other actions depending on what ios found.

How long it takes depends on the length of the AA sequence, (a few tens, to many hundred AA's), and the idiosyncracies of the individual residues, (some residues are versatile, others really difficult to work with). Hence my comment before, it will always calculate at least one decoy.

What it then does is looks at how long that decoy took to find and makes a decision based on a number of criteria to see if it believes it has time to do another within the user chosen run time, if yes, it starts again and finds another decoy, if not, it says "finished" and uploads.

In that respect, a single decoy, (and some decoys can take over an hour to calculate), is as much a full wu as anywhere else, it depends on a projects definition of a wu. At Rosetta it is just that a wu may contain more then one decoy, this saves network bandwidth for those that have such constraints.

The choice of the original configuration depends a lot on what is already known about the protein or proteins of similar formula. Of course, if it is a totally unknown protein, the ab-initio fold start point is random - it can't be anything else.

If you need to know anything more about Rosetta just ask, (I'd say PM me, but of course, that would require an upgrade to the server software...).
59) Message boards : Number crunching : Initial Replication (Message 19268)
Posted 19 Mar 2008 by Profile adrianxw
Post:
Extending from the same argument,

You can't do that, at Rosetta, the smallest amount of work a wu can do is a single iteration. You can set your requested run time very low I suppose, but it would always overshoot it. You claim that this is not a complete wu, yet here, the wu's run for "n" trips around the LHC, where "n" is variable. What constitutes a "complete" wu here? The beam might collapse 2 circuits after the wu was set to run for. Same at Rosetta, a "complete" wu would be when every possible protein conformation had been checked, that would take forever and is why we use DC in the first place.
For the other tasks you listed, you were 4th or 5th in. This time, you're 2nd.

So what? In the other wu's I wasted my cycles, in that one, others will. If you look at it now, there are 3 results, all agree, no errors, credit granted. 2 machines out there are now crunching, for a long time, for a total waste of time, cycles and energy.

The other is not at issue. Sure, if I had a big resource share and the work was thick and fast, but neither is the case.

And yes, I do consider it a pretty poor show when a resource contributed by volunteers is wasted.
60) Message boards : Number crunching : Initial Replication (Message 19266)
Posted 18 Mar 2008 by Profile adrianxw
Post:
I see no insults being hurled by me in my post? However...
You could've run 2 Rosetta tasks that completed within 26 minutes? I've never run Rosetta, but I find this unlikely.

You start by calling me a liar.

At Rosetta, you can choose how long you want each work unit to run. What it does is downloads a start point and then runs as many iterations of the model as is possible to more or less complete and return the wu at your desired run time. In fact, the run time is "approximate" because the runtimes are rarely an exact multiple of the iteration for a particular protein, it evens out. That said, you can easily choose to run 1 hour wu's if you so wish.

Point 2, this wu is one of the ones that arrived yesterday. My machine completed the wu in 17,677 seconds, and is currently the fastest machine that has completed it. A little maths shows that to be 4.9+ hours. I could therefore have completed ~4.9 1 hour Rosetta wu's in the same time. I also note that one of the last wu's you crunched was longer then that I link to now.

Here is the wu summary in case it disappears.


I do, indeed, have the ability to micromanage the 2 super quads I have here at home, but I have other machines at remote locations that I visit infrequently. LHC is not attached there.

I said I was getting narked at LHC wasting cycles that could be being used for productive purposes. I stand by that. All of the wu's I have had recently have been crunched to completion without error by all of the machines they have been sent to. A IR = Q would have been totally adequate.

I am left to ponder whether to continue having a percentage allocated to LHC at all.

The first line of my last post states the situation, the project doesn't NEED to do anything.


Previous 20 · Next 20


©2024 CERN