Message boards : Number crunching : no new tasks?
Author | Message |
---|---|
tomduly Send message Joined: 5 Apr 20 Posts: 3 Credit: 578,397 RAC: 0 |
Currently, rosetta@home does not provide new WUs to the volunteer community. Unfortunately, no information is given why. Gigaflops of computing power are running empty. Dear Rosetta team: pls. be aware, that providing distributed computing power for free, isn't for free for the most of us, since it costs a significant amount of electricity, that we have to pay for. Keeping the community uninformed, is... disrespectful. Tom |
yoerik Send message Joined: 24 Mar 20 Posts: 128 Credit: 169,525 RAC: 0 |
I hate to complain - but several volunteers have raised issue with the lack of communication from the project to us, for a while now. https://boinc.bakerlab.org/rosetta/forum_thread.php?id=13533&postid=96797 It's good that only work that helps the project(s) is ran, but it's not good to run out of WUs completely without notice. I hate continuing to compare them to WCG - their communication isn't much better, but shortages on subprojects are noted well in advance. |
mikey Send message Joined: 5 Jan 06 Posts: 1895 Credit: 9,188,532 RAC: 3,503 |
Neither one of you guys would do well in the days before Boinc, Seti the ONLY Project, went down for over a week because someone stole the copper phone lines, yes dialup days, and NO ONE from Seti said one word until they came back online after the phone company replaced the line. YES the website was still running but NO communications what so ever from them to us crunchers, they DID tell a few users individually and that's how we all found out. Even later during it's run when Seti reissued the exact same workunits over and over and over again Seti never said a thing until they got caught and after denying it for weeks finally admitted it was because 'they ran out of workunits and thought people would leave the project if they didn't have workunits to crunch'. NO the reissued workunits were NOT reprocessed by the Project, the results were thrown away and the workunits were sent out again and again. They did eventually apologize but it wasseen asweakand waaaay too late. Rosetta is dependant on outside vendors to supply tasks to them, when that doesn't happen they run out of tasks. One thing to know though is within a few days they always get new workunits again. So give it time guys and crunch for some other Boinc project right now. This is NOT a once in a lifetime thing for them, no they have no clue it's about to happen and no they have no clue when the units will flow again. World Community Grid has LOTS of their Open Pandemic COVID-19 workunits available to crunch in the meantime. |
yoerik Send message Joined: 24 Mar 20 Posts: 128 Credit: 169,525 RAC: 0 |
Neither one of you guys would do well in the days before Boinc, Seti the ONLY Project, went down for over a week because someone stole the copper phone lines, yes dialup days, and NO ONE from Seti said one word until they came back online after the phone company replaced the line. YES the website was still running but NO communications what so ever from them to us crunchers, they DID tell a few users individually and that's how we all found out. Even later during it's run when Seti reissued the exact same workunits over and over and over again Seti never said a thing until they got caught and after denying it for weeks finally admitted it was because 'they ran out of workunits and thought people would leave the project if they didn't have workunits to crunch'. NO the reissued workunits were NOT reprocessed by the Project, the results were thrown away and the workunits were sent out again and again. They did eventually apologize but it wasseen asweakand waaaay too late. and that's fair - but we're not in the age of dialup anymore. It's more than reasonable to expect some level of communication from the project that we are spending time, money and resources to support. I'm not whining, or demanding for more tasks. I specifically said I'm glad that they're not wasting resources by giving us bad WUs, multiple times across the forums. But - there hasn't been a news posting since early May - it's nearly July. That seems to be par for the course for Rosetta, at a glimpse of past news threads - but it is more than reasonable to ask for a few sentences a month to keep us up to date on the project's work - put out an announcement to the clients so that non-forum users (which there are clearly plenty) can be aware of the shortage without panicking when they suddenly stop getting tasks. Not everyone here is a computer expert - I'm certainly not. But I also know that stringing a few sentences together isn't difficult. I was ridiculed here before during the last potential shortage in the beginning of May. It is certainly not my intention to whine or complain - and I'm sorry if I'm coming off that way, but what OP asks for - and what other users asked for in the news post I linked prior - is not much to ask. It takes a minute of one person's time to keep us informed - and the project is either unaware or unwilling of how important communication is, especially when it requires such little effort. |
[VENETO] boboviz Send message Joined: 1 Dec 05 Posts: 1995 Credit: 9,653,470 RAC: 7,312 |
Neither one of you guys would do well in the days before Boinc, Seti the ONLY Project, went down for over a week because someone stole the copper phone lines, yes dialup days, and NO ONE from Seti said one word until they came back online after the phone company replaced the line. YES the website was still running but NO communications what so ever from them to us crunchers, they DID tell a few users individually and that's how we all found out. I remember.....but this is not a justification. Rosetta is dependant on outside vendors to supply tasks to them, when that doesn't happen they run out of tasks. Only Robetta works is from outside. Others simulations are inside Bakerlab. Maybe they have hw problems Maybe they want to "clean up" the queues to start new simulation campaigns Maybe they are analizing results before new works Maybe And volunteer goes to other projects and, maybe, they will not return |
[VENETO] boboviz Send message Joined: 1 Dec 05 Posts: 1995 Credit: 9,653,470 RAC: 7,312 |
It takes a minute of one person's time to keep us informed - and the project is either unaware or unwilling of how important communication is, especially when it requires such little effort. +1 |
Grant (SSSF) Send message Joined: 28 Mar 20 Posts: 1687 Credit: 18,015,673 RAC: 23,804 |
It takes a minute of one person's time to keep us informed - and the project is either unaware or unwilling of how important communication is, especially when it requires such little effort.Or maybe it's night time over there & everyone is still asleep? Grant Darwin NT |
gmercato Send message Joined: 2 May 20 Posts: 9 Credit: 1,097,720 RAC: 0 |
Right, it's 0057h in Seatle, WA now. For the next 7h probably nothing is gonna happen. |
[VENETO] boboviz Send message Joined: 1 Dec 05 Posts: 1995 Credit: 9,653,470 RAC: 7,312 |
Or maybe it's night time over there & everyone is still asleep? The queue is empty since yesterday. Over 24h of interruption... |
Don_Herres Send message Joined: 24 May 20 Posts: 1 Credit: 701,054 RAC: 1 |
Depends on what part of the system is down. The usual about not sending a message because the messaging system does not work. |
Daedalus Send message Joined: 1 Aug 08 Posts: 39 Credit: 10,107,163 RAC: 301 |
i don't know if it is related but BOINC on my android smartphone has been unable to communicate with r@h for days. Soon after Einstein@home. i doubt it is related but no warning whatsoever. Edit: That said i crunch for folding@home which is far from being useless. I used to alternate between projects. |
CIA Send message Joined: 3 May 07 Posts: 100 Credit: 21,059,812 RAC: 0 |
Depending on the day I have ~90 cores spread across several machines running Rosetta. Even with the queue mostly empty currently I have ~90 tasks total. Most machines are set to 24 hour crunch times vs the default 8, and they all have bare min (.1, .1) or zero cache. So even with the WU drought things are still coming through. My primary work machine I'm typing this on only had 6 of 24 cores crunching this morning when I came into work, but about 20 minutes ago it randomly grabbed 18 more WU's on it's own all with the same naming convention. Checking that computers tasks here on the site all the WU's I just got aren't re-issued ones, so new work is trickling out. Just slowly. /edit. I will add that in the last few days I've seen a lot of "error while downloading" messages spread across all my machines. 21 errors and counting. |
gmercato Send message Joined: 2 May 20 Posts: 9 Credit: 1,097,720 RAC: 0 |
My Intel cores are working for folding at home now too but Rosetta seems to have the only ARM64 Linux client. So my Pi4s will have to wait for the rosetta queue. |
shanen Send message Joined: 16 Apr 14 Posts: 195 Credit: 12,662,308 RAC: 0 |
Might be hard to believe, but some of the other BOINC projects have even worse project management than this one. The management bottleneck on this one seems to be focused on one or two people. I've run a number of projects over the decades,,, Actually started with seta@home before BOINC was created. Then again this one does seem to be entering its death throes. I tested out another project with one of my machines, but didn't like it much, so I haven't switched yet. Anyone have any recommendations? WCG keeps bugging me to come back, so maybe they've cleaned up their act. #1 Freedom = (Meaningful - Constrained) Choice{5} != (Beer^3 | Speech) |
Jim1348 Send message Joined: 19 Jan 06 Posts: 881 Credit: 52,257,545 RAC: 0 |
Then again this one does seem to be entering its death throes. I tested out another project with one of my machines, but didn't like it much, so I haven't switched yet. Anyone have any recommendations? WCG keeps bugging me to come back, so maybe they've cleaned up their act. They are definitely NOT in their death throes, just the usual poor communications. It is probably because they have so many researchers submitting work from various locations that they have no idea what is coming in. At least that is the charitable interpretation. As for WCG, it is not a bad idea. I am doing Africa Rainfall Project (ARP), which run well on my Ryzen 3000 series; around 12 to 16 hours. You could do COVID-19 (and beyond) research with OPN also, but they are developing a GPU version which should be out before long. I would wait for that. |
manalog Send message Joined: 8 Apr 15 Posts: 24 Credit: 233,155 RAC: 0 |
Put them on tn-grid https://gene.disi.unitn.it/test it works very well on arm even if 32 bit (if I remember correctly). If you want, you could compile the source for your platform and share the binary with the others. It's extremely easy. Other projects for arm are: wcg opn , universe, lhc sixtrack, einstein EDIT: I've just checked and there is a 64 bit arm version of tn-grid actually |
Sid Celery Send message Joined: 11 Feb 08 Posts: 2127 Credit: 41,266,340 RAC: 8,573 |
Then again this one does seem to be entering its death throes. I tested out another project with one of my machines, but didn't like it much, so I haven't switched yet. Agree with pretty much all that - that's how things are here. Usually Rosetta runs pretty well. Task shortages come once or twice per year. Maybe 3 at most. One of those times is either Christmas or shortly after returning from the holidays in January. In March, everyone and their dog turned up here - number of hosts increased x10 and the additions were hefty in CPU power, so all tasks got wiped out and it took 7-10 days to sort it out. Is this the first shortage since? I dug up a message from late March, some of which ( poimt 3) doesn't apply any more, but most does. Worth re-posting it as a reminder what's involved in bring tasks to us Yes indeed. We burned through all the work units. |
manalog Send message Joined: 8 Apr 15 Posts: 24 Credit: 233,155 RAC: 0 |
Just got some new tasks with this name: DL_TrR_PSSM_plus_rosetta1998881_0001_0001_fragments_abinitio_SAVE_ALL_OUT_950049_595_0 DeepLearning TrRosetta. https://boinc.bakerlab.org/rosetta/forum_thread.php?id=14042 https://boinc.bakerlab.org/rosetta/forum_thread.php?id=13921 PSSM: https://en.wikipedia.org/wiki/Position_weight_matrix IMHO prepare for a breakthrough in R@h after this pseudo-break ;) |
Jim1348 Send message Joined: 19 Jan 06 Posts: 881 Credit: 52,257,545 RAC: 0 |
Just got some new tasks with this name: Yes, I got some too. They are back. |
Endgame124 Send message Joined: 19 Mar 20 Posts: 63 Credit: 20,367,622 RAC: 2,215 |
Looks like about 30000 tasks ready to download. Some of my hosts that were out of work have started to pick some up, and I’m guessing I’ll have work across the board by morning. |
Message boards :
Number crunching :
no new tasks?
©2024 University of Washington
https://www.bakerlab.org