1)
Message boards :
Cafe Rosetta :
how do I get back into other projects, and boincstats
(Message 79675)
Posted 1 Mar 2016 by River~~ Post: But it sounds like your primary avenue forward is to ask the admins of the various projects to modify the EMail address associated with the accounts, or to enjoy a fresh beginning, as you did when you stopped using the EMail address. Sounds like my best bet. @Snags: yep, thanks for the thought, but I've spent ages trying every password I ever remember using, like my first car, my first savings a/c, and some other things like that... I did hit paydirt on CPDN that way, but still no joy with the project I most wanted to re-join. I used to teach General Relativity before I retired, so no prizes for guessing which, especially with recent news ... Time to throw myself on the goodwill of the Einstein mods, I think! R~~ |
2)
Message boards :
Cafe Rosetta :
how do I get back into other projects, and boincstats
(Message 79646)
Posted 27 Feb 2016 by River~~ Post: I believe, if you know the project password, you could sign on to the project website, ... THAT IS EXACTLY THE PROBLEM. As I already said in the OP, "I would really like to reconnect with Einstein, LHC, but don't have the authenticators for those two projects. I no longer have access to the email address, so being emailed there is not going to help me." If I understand you correctly, by "project password" you mean the hex string tagged as <authenticator> ? Or if not where do I look for the project password? Sadly (or happily for Rosetta) the only computer I still have seems to have been connected only to Rosetta at the time my internet connection died permanently. FIRST QUESTION is there any OTHER way of connecting back to those accounts on those other projects? SECOND QUESTION how do I get going again with BoincStats? any ideas anyone? |
3)
Message boards :
Number crunching :
Help Rosetta with FREE 60 days on Google Cloud!
(Message 79634)
Posted 25 Feb 2016 by River~~ Post: Two comments on the original post. First about nesting VMs. You can't if you are using the special CPU flags to facilitate running in a virtual machine - these are either "on" or "off". There is no way the processer can understand that a level could be the virtual guest of one machine and still be the host of another VM. If you don't use hardware acceleration you can do it (but not on any of the commercially hosted VM servers. Qemu without the KQemu extensions will quite happily run itself running itdelf running itself, just getting slower at every layer of emulation. SEcond, if you are not using a GUI I don't think you need boinc-manager. As far as I know boinccmd comes with the client. So I am suggesting you adjust those apt-get's accordingly. I will try it myself in a month or so, when I have time. As you seem to be aallowed only one trial, there is no rush for me to get my free trial in. But I will do, if only for the stats... Thanks for this thread R~~ |
4)
Message boards :
Number crunching :
CERN Engineer Details AMD Zen Processor Confirming 32 Core Implementation, SMT
(Message 79633)
Posted 25 Feb 2016 by River~~ Post: Just noticed my last sentence got cut off somehow. My post was intended to end like this: I hope AMD do well with this,because I strongly agree with other posts which say that we the computer buyers need the market to have some strong competition for Intel. R~~ |
5)
Message boards :
Number crunching :
Android. How to limit tasks downloaded?
(Message 79632)
Posted 25 Feb 2016 by River~~ Post: apologies, I did not explain this bit very well, as I described an earlier plan when I said 66 Set the parameter for max disk space allowed to Boinc to 0.01 GB or anything less: with the current Rosetta WU nothing will ever be downloaded. If you get work, divide the space you allowed by the number of units you got, and then divide by two to make sure. 99 In fact after some experimentation I found what you have to set is the reverse figure to that - you have to play around with the amount of space you want BOINC to leave free for your other apps. So to get more WU you make the number smaller, to get fewer WU you make the number bigger. again, hope this helps - and if you or anyone else knows a better way to do this, do let us know as I am still experimenting myself. |
6)
Message boards :
Number crunching :
CERN Engineer Details AMD Zen Processor Confirming 32 Core Implementation, SMT
(Message 79629)
Posted 25 Feb 2016 by River~~ Post: Multi core It's good for BOINC people, off course it is, but remember that you need memory for each core that's enabled for crunching. The market driven radon for huge multi core CPUs is similar - it's for servers running many virtual machines, or a server running multiple containers. Each VM or each container is by definition able to rub It's own threads with no additional programming (once you've installed Xen or Docker oute your choice off the proprietary alternatives) Or, even more so, for a server hosting eight VMs, each one running seven containers: this allows one thread for Docker and one each for the containers. Almost everyone gets their own core, it's just missing one for the hyperbole. So 64 cores makes very good sense in the server market. Will I be able to afford them? Probably only two our three years later as second hand. For the home computing market, AMD is also looking at 64 bit ARM chips - Many guess id's that that is what most of its will be buying, if we go for AMD. That is a more radical departure for AMD - up to now it's been totally loyal to Intel's x86 architecture, to the point that when Intel tried to drop x86 with ia64, AMD was able to use market pressure forces to force then to bee loyal to their own offspring. As a Brit I'm also excited to see further scope for the wonderful UK based ARM design. I hope |
7)
Message boards :
Number crunching :
Android. How to limit tasks downloaded?
(Message 79628)
Posted 25 Feb 2016 by River~~ Post: Greetings! hi Jay I assume you have discovered, as I seem to have, that web prefs don't propagate to the Android client in the same way as they do on the regular Linux client. I'm still working on that and will post again if I work out how to make that happen. In the meantime, here is a fudge that world's for me on two different Androids where I have tried it. I've only tried it once on each, like your Android both off mine are munching their very first WUs as I write this. Right now my two core Galaxy Note 1 has only downloaded one WU, despite being enabled for both cores in prefs. It's a kludge, and will need adjusting when the project upgrades their software. It will also give erratic results if you mix projects. It works by starving BOINC of disk space. By luck, I have only just discovered the same issue myself and this is my workaround. It's been tried exactly twice so far and worked both times. Set the parameter for max disk space allowed to Boinc to 0.01 GB or anything less: with the current Rosetta WU nothing will ever be downloaded. If you get work, divide the space you allowed by the number of units you got, and then divide by two to make sure. Once you have ahem "successfully failed" to get work, check the log file. In the Android client this is currently the last of the options in the main menu, and you access the main menu by tapping the three lines top left. Look for entries around the time you got the no work message. There will be one that tells you how much space id's needed for one WU. Set your limit to (N plus 0.5) times that value and you should get around N units. It's not perfect: sometimes the project might send out WU with mixed sizes, duo you may get some variation. If at some later time you are getting consistently the wrong number the repeat the slice calibration again. And out is not likely to work across projects which will each have their own disk space needs. Also Einstein for example sometimes used the same huge data file for several WU, so that the first downloaded WU puff a particular kind would demand more space than follow up work on the same data. At present, Feb 2016, a setting of 1GB per WU seems to be working for Rosetta on my Androids. The figure I got from from the logs was just over 950MB Hope that's useful - it's not elegant but it works for me - at least for me River~~ |
8)
Message boards :
Cafe Rosetta :
how do I get back into other projects, and boincstats
(Message 79571)
Posted 20 Feb 2016 by River~~ Post: hi everyone, Lots of questions, most of which are not to do with this project, so please feel free to post brief signposts to where I can get the info on other sites. I have just returned to BOINC after a nearly 8 year absence. I am feeling frustrated right now as I know that in 2007 I knew the answers to all these questions.... Luckily I tracked down my Rosetta authentication code on an old computer (which itself hadn't been powered up for half a decade) so I am back with my history here. I would really like to reconnect with Einstein, LHC, but don't have the authenticators for those two projects. I no longer have access to the email address, so being emailed there is not going to help me. Can I use my cross project id to get back in there? If not, any other ideas? And Boincstats. My stats gif no loner shows up in my sig, though the HTML looks like (say) that of Bill Michael who does sjho up nicely. Has Boincstats introduced registration? Has Willy changed the HTML and I've just missed the change? How do I llog in to Boincstats without remembering the password and without access to that ancient email address? And if I end up with a new authenticator on each of those other two projects, will Boincstats patch it together with my old account when those stats are linked to Rosetta, which creates the continuity? any help gratefully received River~~ (also known as gravywavy on Einstein@home) |
9)
Message boards :
Cafe Rosetta :
Personal Milestones
(Message 79565)
Posted 20 Feb 2016 by River~~ Post: A milestone with a difference... A gap between posts of 3000+ days - last post before this was 10 April 2007. And my recent daily average has dropped to 0.01 or something. I haven't had a net connection for all that time, and that is a bit of a downer for BOINC. I also never uploaded my final work, which is still on the hard drive of my Pentium 3 computers (running Win2k and Linux with kernel 2.6). I'm guessing David won't be too interested in getting it back now... I had to fire up one of the Linux boxes to authenticate my account, first time that box has had power for maybe six years. OOOOh look,my first WU since April 2007 has just started running, having downloaded in the background while I've been typing. bye for now ~~ |
10)
Message boards :
Number crunching :
Stop or pause BOINC at next checkpoint.
(Message 39244)
Posted 10 Apr 2007 by River~~ Post: I've written a little tool, called stopatcheckpoint :P... Nice one - thanks! Thanks for picking up on my idea and running with it. I'll be using your implementation on my winboxes :) I've posted a cross link from the Linux thread to here. R~~ |
11)
Message boards :
Number crunching :
linux script - stopatcheckpoint
(Message 39243)
Posted 10 Apr 2007 by River~~ Post: ...The script is in Bash (Unix command line) so will not port to windows without more effort than I am going to put in... but there is always someone willing to take an idea further. Cristophe has posted an .exe file for windows users that does roughly the same thing. Nice one C! See this post R~~ |
12)
Message boards :
Rosetta@home Science :
How estimate time for rosetta completion?
(Message 38518)
Posted 28 Mar 2007 by River~~ Post: Rosetta is a project to improve the ability to accurately model protein (and now also RNA) folded structures, and their interactivity with other molecules. Accurate prediction for all proteins is almost certainly many years away yet, although the predictions for many protiens are now good enough to be used, there's still plenty of work to be done on this front. However much we know about protein structures there will always be more to find out, and therefore better ways to write computer programs. In that sense the Rosetta project is adding to something that will never be finished. In another sense, Rosetta itself is only one part of the action. At some point Bakerlab, or/and David himself, might end the Rosetta project but similar work will go on elsewhere; probably for as long as there are humans alive. In the same way as when David started out he could build on other people's work then other people will be building on the current work. So the question does make sense, in the sense of 'how far does the Rosetta project intend to run with its own contribution'. I think the truth here is that David et al can see so much room for improvement on the current models that, in practical terms, there is no end point in sight. |
13)
Message boards :
Number crunching :
Ps3 - linux app?
(Message 38181)
Posted 23 Mar 2007 by River~~ Post: Given that the PS3 already runs Linux (see http://www.eurogamer.net/article.php?article_id=70611 for example), what chance of compiling the boinc & rosetta code for that platform? Why wait for Sony to offer? What am I missing, it seems such an obvious idea there must be some reason its not already been done? River~~ |
14)
Message boards :
Number crunching :
turn off hyperthreading?
(Message 38019)
Posted 19 Mar 2007 by River~~ Post: I can't believe that people still think that they can produce more with HT turned off. It is possible, at least in theory, and the fact that there are reports where others claim to have ssen this, and the fact that WCG recommend to turn it off, all make the theory plausible. One example, if a single WU uses a lot more than half the cache, at any level, then turning HT off will reduce cache misses. The increase in speed due to better use of the cache will offset the increased productivity of HT in all cases, and in extremecases may lose more than HT has saved. Another example, is if a single WU can run in main memory without using the swapfile, but the addition of a second WU brings the swapfile into use. In either of these cases, and in several other examples, turning HT off would increase speed. For that reason, my advice is to try it for a few days, or at most for a week, and see what the actual throughput is with HT on and HT off, and with your typical mix of projects and your typical use of the box for non-BOINC work. Interestingly, the new memory management features of the client will step the number of tasks down when the swapfile comes into play, meaning that if it was swapfile issues that stopped you using HT before, you should try it again and let BOINC choose whether to use one thread or two. River~~ |
15)
Message boards :
Number crunching :
80 cores, not science fiction ... it is real! 1.2 TERA!
(Message 37699)
Posted 12 Mar 2007 by River~~ Post: Can t wait to see the productization of all those great discoveries. That is why we need AMD. And why we also need Intel. If there were only one company in the marketplace, it could hold stuff back in exactly the way you suggest. In a multi-supplier environment (and two counts as multi, just) then there is a race to see who can get to market first. Intel will want to get this out as soon as it can, in case AMD gets something similar out first. It will be a long time coming, but more because a lot of work is still needed - new ways of compiling code for example (see this week's New Scientist) as well as the more hardware-oriented aspects - how to support this on a motherboard, what to use as a BIOS, wheter we will still offer DOS 16-bit backward compatibility, etc etc. R~~ |
16)
Message boards :
Number crunching :
UPDATE BOINC............
(Message 37697)
Posted 12 Mar 2007 by River~~ Post: ...Could someone explain to me the down side to F@H teaming up with BOINC? Seems to me this is a great case of win/win! OK, but before I explain, I am explaining *other people's* objection, OK, not my own. Personally I am with you all the way, I think it would be good for BOINC and good for F@H for them to team up. So if you disagree with what follows, pkease don't shoot the messenger. BOINC is a multi-project platform, by design. This means that it has to offer a common denominator for the needs of different projects. This raises the following issues 1. It never fits a specific project as well as an interface specifically designed for that project. Stand-alone project designers can tune redundancy, error checking, credit scoring, whether to penalise for non-return of WU, whether to penalise for late return of WU, etc etc to the real needs of their scientists and the perceived needs of their users. 2. The BOINC structure acts as a brake on developing new ideas about how to do a Distributed Computing (DC) project. It may be good for the application, but bad for the long term development of DC. An example is that on the non-BOINC project, pi-segment, you get 'charged' 100 credits to download a work unit, and then get more than that back when you return it, and get even more back if you return it on time. This discourages folk from downloading work and simply cancelling it. An idea like that would be hard to introduce to BOINC because of the BOINC policy of trying to get credits worth the same across all projects. BOINC stops the need to re-invent the wheel, but in so doing, is it preventing the discovery of the pneumatic tyre? 3. In the case of an existing project, there is user inertia. Look at the furore caused by the transfer of SETI-classic to SETI-BOINC amongst some of their users. People get used to particular ways of using bulletin boards. People have tuned their DC activities to getting the best credit under one system, and it feels to them like moving the goal-posts if the project moves to another way of doing things. CPDN got round this by continuing to offer the old interface for as long as anyone wants to use it - last time I looked there were still folk crunching and (I think) downloading new work on the CPDN-Classic system. WCG got round this by keeping BOINC as a second interface. They don't run the BOINC website, only the BOINC servers. Each task completed is scored on the WCG system and also on the BOINC system, so the accumulated credits you get on their site are different to those shown by BOINCstats, etc. But this puts off the very BOINC people the project want to attract - I rarely visit their site as I don't know where to find anything, and that means I crunch little of their work. Hope that helps you grok the countervailing views a little. Coming back to my own position, I would not say that any of the countervailing argumants are wrong, but that they are, IMO, outweighed by the advantages of BOINC - a cross project uniform interface, a cross project community (there are folk like John Keck, Fuzzy Hollynoodles, Misfit who turn up on several projects), and so on. And above all by the fact that it is so easy to join a new project. River~~ |
17)
Message boards :
Number crunching :
Delays on downlaod - file server struggling - or is it me?
(Message 37696)
Posted 12 Mar 2007 by River~~ Post: I get the impression that the file server is struggling at the moment File downloads are often taking more than one attempt to succeeed, and it has taken over an hour to *start* downloading files in the last few days. I am not talking about transfer time once the downloads start - I mean that I am getting messages like the following in the message window Have I got a network error at my end that is causing this, or are other people also seeing this? Has the Rosetta fileserver got any known network issues at present? (Remind me - is it in fact the same box as the scheduler server, or different hardware?) Is Rosetta once again becoming a victim of its own success, and the fileservers becoming close to overload? This counts as a minor irritant for me, but must be very annoying to any dial-up users who are getting similar delays downloading files. River~~ ------------------------------------------------------------------ rosetta@home 12/03/2007 07:35:52 Sending scheduler request: To fetch work rosetta@home 12/03/2007 07:35:52 Requesting 75600 seconds of new work rosetta@home 12/03/2007 07:35:57 Scheduler RPC succeeded [server version 509] rosetta@home 12/03/2007 07:35:57 Deferring communication for 4 min 2 sec rosetta@home 12/03/2007 07:35:57 Reason: requested by project rosetta@home 12/03/2007 07:35:59 [file_xfer] Started download of file frags83_1bq9A.fasta.gz rosetta@home 12/03/2007 07:35:59 [file_xfer] Started download of file frags83_1bq9A.psipred_ss2.gz --- 12/03/2007 07:38:00 Project communication failed: attempting access to reference site rosetta@home 12/03/2007 07:38:00 [file_xfer] Temporarily failed download of frags83_1bq9A.fasta.gz: http error rosetta@home 12/03/2007 07:38:00 [file_xfer] Temporarily failed download of frags83_1bq9A.psipred_ss2.gz: http error rosetta@home 12/03/2007 07:38:00 [file_xfer] Started download of file boinc_frags83_aa1bq9A03_05.200_v1_3.gz rosetta@home 12/03/2007 07:38:00 [file_xfer] Started download of file boinc_frags83_aa1bq9A09_05.200_v1_3.gz --- 12/03/2007 07:38:01 Access to reference site succeeded - project servers may be temporarily down. --- 12/03/2007 07:40:01 Project communication failed: attempting access to reference site rosetta@home 12/03/2007 07:40:01 [file_xfer] Temporarily failed download of boinc_frags83_aa1bq9A03_05.200_v1_3.gz: http error rosetta@home 12/03/2007 07:40:01 [file_xfer] Temporarily failed download of boinc_frags83_aa1bq9A09_05.200_v1_3.gz: http error rosetta@home 12/03/2007 07:40:01 [file_xfer] Started download of file frags83_1bq9.pdb.gz rosetta@home 12/03/2007 07:40:01 [file_xfer] Started download of file rhhsm6splthh6d.ljatr.etable.gz --- 12/03/2007 07:40:04 Access to reference site succeeded - project servers may be temporarily down. --- 12/03/2007 07:42:03 Project communication failed: attempting access to reference site rosetta@home 12/03/2007 07:42:03 [file_xfer] Temporarily failed download of frags83_1bq9.pdb.gz: http error |
18)
Message boards :
Number crunching :
Job Sizes
(Message 37622)
Posted 8 Mar 2007 by River~~ Post: Thanks. Well, maybe I don't understand how things REALLY work, but my brain tells me that if the DCF and your preferred run time are to work as intended then every estimated task runtime (provided by the project) must remain constant no matter what... otherwise the DCF will never become adjusted to your run preference, it will always be a moving target.Anyone else observe this behavior? hmmm, In the normal case (ie non-Rosetta), the estimated run length needs to be proportional to the number of flops in a task, so that machines can learn the constant of proportionality. In the case of Rosetta, the wall time is the time used to calibrate the DCF, as it is the wall time that determines when a run stops, for any given setting. So to the first approximation I agree that the server-provided run length wants to be constant for any given machine. Ideally, so that new users start from a plausible approximation, it should depend inversely on the benchmark, so that the initial approx is 3hrs. Even if it does not depend on the benchmark, if it is constant then the machine only has to learn the DCF once. But then, different tasks run at different cpu efficiencies. A task that over-fills memory and needs a lot of swapping will run less cpu in the same wall time. This will mean that the DCF rises, as the average useful cpu cycle takes longer. This means that keeping a constant asserted value is still not going to keep the predicted run length stable when there is a varied mix of kinds of workunit. The whole DCF approach is based on the premise that work from a given project has (for example) the same mix of adds and multiplies, and the same proportion of page faults, etc etc over a huge number of WU. This assumption works well on Einstein, SETI, and LHC, which effectively run the same program for months or years. However, as these effects cannot be predicted easily, and as they vary considerably from one machine to another, I still agree that constant is the best approach -- just don't be surprised if the effects fluctuate rather. R~~ |
19)
Message boards :
Number crunching :
linux script - stopatcheckpoint
(Message 37618)
Posted 8 Mar 2007 by River~~ Post: An additional limitation to this script is that it does not wait for a task that has never checkpointed at all, that is it only waits for tasks that have passed their first checkpoint when the script is called. R~~ |
20)
Message boards :
Number crunching :
linux script - stopatcheckpoint
(Message 37612)
Posted 8 Mar 2007 by River~~ Post: hi, This is a bash script to stop BOINC soon after the running task checkpoints. One reason for writing it was to demonstrate the usefulness of the boinc_cmd utility that comes with the linux implementation of BOINC. The other motivation is described under Background, below. Background You will know that in contrast to many other BOINC projects, Rosetta checkpoints its results rarely. The reason this is a nuisance is that when you stop BOINC for any reason, work done since the last checkpoint is lost. BOINC and Rosetta between them recover from the situation when BOINC is re-started, but the lost work needs to be re-calulated. Rosetta and CPDN are the two projects I run where this is an issue. On CPDN it is a particular issue as the project encourage users to back up their work every week or so, and to do that involves stopping BOINC. On both Rosetta and CPDN, it may be wise to stop BOINC before doing computationally excessive work as this might cause theBOINC application. You may want to shut down the machine to go on holiday, install a new DVD drive, etc etc. This script waits for the next checkpoint and exits. One use would be to reboot or shutdown after the next checkpoint: $ ~boinc/stopatcheckpoint; reboot $ ~boinc/stopatcheckpoint; halt Another use would be to run some other work, then restart BOINC, as in the following command line ~boinc/stopatcheckpoint; /pi/timpi 22 22; /etc/init.d/boinc start (note that I've got a Debian-style start command for BOINC here, other systems can work out your own version of this) The script Copy and past the following into file stopatcheckpoint in the BOINC directory. The script needs to be run from there to make the boinc_cmd utility pick up its passwords. If you need to run it from elsewhere, either copy the password file across, or build the passwords into the script. After pasting the script into your file system, you need to make it runnable, ie some variation of chmod 755 stopatcheckpoint #!/bin/sh # # stopatcheckpoint # # Author River 2007 # # Copyright but may be distributed under the GPL # - see http://www.gnu.org/licenses/gpl.txt # # River asserts his moral right to be identified as the Author echo "wait for BOINC checkpoint..." cd ~boinc prv=`./boinc_cmd --get_results|grep -v 0.0000|grep checkpoint` while (./boinc_cmd --get_results|grep "$prv">>/dev/null) ; do clear ./boinc_cmd --get_results|grep -v 0.0000|grep "time|----" echo "waiting for change to:$prv ..." sleep 3 done echo echo "*** new checkpoint ***" echo ./boinc_cmd --get_results|grep -v 0.0000|grep "time|----" echo echo "stopping BOINC..." echo ./boinc_cmd --quit Sample output While waiting the screen displays some info on every WU on the machine: 1) ----------- checkpoint CPU time: 25081.809985 current CPU time: 26803.275281 estimated CPU time remaining: 57799.145096 2) ----------- estimated CPU time remaining: 78446.605958 waiting for change to: checkpoint CPU time: 25081.809985 ... here, WU 1 is running and WU 2 is waiting to run. If a WU is finished then the final CPU time is shown. If there is more than one checkpoint time in the list, there may be problems, as described below. The current CPU time and time remaining should be changing every few seconds. After the checkpoint is reached the screen is left with a display like: 1) ----------- checkpoint CPU time: 25081.809985 current CPU time: 26878.500846 estimated CPU time remaining: 57852.131426 2) ----------- estimated CPU time remaining: 78446.605958 waiting for change to: checkpoint CPU time: 25081.809985 ... *** new checkpoint *** 1) ----------- checkpoint CPU time: 26879.487696 current CPU time: 26881.467395 estimated CPU time remaining: 55478.468342 2) ----------- estimated CPU time remaining: 78446.605958 stopping BOINC... Where this script does and doesn't work This script works well on a single cpu machine with only one project, and when there is onlu one task that has started. There can be any number of completed tasks waiting for upload, and any number of unstarted tasks waiting to run. This script gets confused if there is more than one running / started task, for example: - on a multi-cpu box; and - if BOINC starts one task without finishing another, as is normal on a multi-project box, and as may happen on a single project box since BOINC v 5.8 introduced memory management. Hope this is useful. The script is in Bash (Unix command line) so will not port to windows without more effort than I am going to put in. River~~ |
©2024 University of Washington
https://www.bakerlab.org