Message boards :
News :
New multicore app and WUs
Message board moderation
Previous · 1 · 2 · 3 · 4 · 5 · Next
Author | Message |
---|---|
Send message Joined: 28 Mar 09 Posts: 490 Credit: 11,731,645,728 RAC: 47,738 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
OK, we will start production mode next week. Unfortunately we will need more than 50x the current number of CPUs, but it is just the start now, so it is ok. You will need a windows app for this. |
Send message Joined: 16 May 13 Posts: 41 Credit: 145,731,947 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() |
Since I had 100% errors (Message 48156 - Posted: 12 Nov 2017 | 2:36:31 UTC) on my first batch of these CPU tasks, I created a symlink as instructed, then deleted the symlink as subsequently instructed, but I have never received a single task since my 12 Nov 2017 post. Same here ... |
Send message Joined: 2 Jul 16 Posts: 338 Credit: 7,987,341,558 RAC: 178,897 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() |
I received some yesterday on a new install of Ubuntu 17.10. No symlink or anything and they completed. |
Send message Joined: 21 Mar 16 Posts: 513 Credit: 4,673,458,277 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
If you need that many CPUs, you will definitely need a windows app |
Send message Joined: 23 Dec 09 Posts: 189 Credit: 4,798,881,008 RAC: 311 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Will the app name stay with "*QC309big*" or will it change for the real stuff? So we might make a app_config file or better still, might you propose an app_config file to limit cpu cores per work-unit to X cores. @PappaLitto: I quite happy with a Linux only app! It is time to make some Linux USB sticks (16GB USB3.0, 10 USD): I work with Lubuntu 17.10 on varios computers I do not use, and it works great! Or try BOINCOS v2.0 Beta Release, there all is pre-configured. |
Send message Joined: 9 Dec 08 Posts: 1006 Credit: 5,068,599 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() |
Making a windows app will probably need one of the following two solutions. Neither is perfect (by far). * The "Windows Subsystem for Linux" from Microsoft. It's unfortunately W10 only (as far as I can tell), and probably we'd be the first BOINC project to use it (=headaches). * A VirtualBox app. Its downsides are known I think. By the way, question for the gurus: when you run a vbox app, is virtualbox automatically installed on your system? |
Send message Joined: 11 Jul 09 Posts: 1639 Credit: 10,159,968,649 RAC: 295,172 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
By the way, question for the gurus: when you run a vbox app, is virtualbox automatically installed on your system? No. The user has to install it themselves, and usually some VBox extensions are recommended as well. There are two ways of installing VBox for Windows: 1) Via a combined single-click installer for both VBox and BOINC, available from BOINC. The simplicity is attractive, but there are downsides - there is no control over e.g. installation location, and the version of VBox included is usually several steps behind the current release. 2) Direct from the Oracle VBox site. BOINC will still recognise this - there's no special BOINC code in the combined VBox installer. Any VBox extensions desired will always have to be downloaded from Oracle. There may be other adjustments required to the host computer, such as enabling virtualisation in the BIOS, which might be unfamiliar to the casual user. |
Send message Joined: 9 May 13 Posts: 171 Credit: 4,594,296,466 RAC: 117,924 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
klepel asked:
<app_config> <app> <name>acemdlong</name> <max_concurrent>2</max_concurrent> <gpu_versions> <gpu_usage>1</gpu_usage> <cpu_usage>2</cpu_usage> </gpu_versions> </app> <app> <name>acemdshort</name> <max_concurrent>2</max_concurrent> <gpu_versions> <gpu_usage>1.0</gpu_usage> <cpu_usage>2</cpu_usage> </gpu_versions> </app> <app> <name>QC</name> <max_concurrent>1</max_concurrent> </app> <app_version> <app_name>QC</app_name> <plan_class>mt</plan_class> <avg_ncpus>4</avg_ncpus> <cmdline>--nthreads 4</cmdline> </app_version> </app_config> This will limit the QC (quantum chemistry) app to 4 threads per task and a maximum of 1 task at a time. You can adjust to your preferences. Hope that helps. |
Send message Joined: 9 Dec 08 Posts: 1006 Credit: 5,068,599 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() |
Thanks to both! |
Send message Joined: 2 Jul 16 Posts: 338 Credit: 7,987,341,558 RAC: 178,897 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() |
I think if the CPU app goes to VBox for Windows/Linux then there will be less user support than just Linux. More than one current task will need to be allowed for efficient CPU usage. |
Send message Joined: 28 Mar 09 Posts: 490 Credit: 11,731,645,728 RAC: 47,738 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Even when you get the windows app going, it looks like you're still going short on the number of crunchers by more than half, based on the server status page which shows currently 821 users crunching long units in the last 24 hours, while 34 are crunching quantum chemistry. (821/34 = 24.15) So, in order to meeting 50x, you will eventually have to create an app for multi CPU-GPU. Quantum chemistry has a long way to go. In the mean time, you can't make the windows app too difficult for the crunchers to set up, because most of us are not computer gurus and you will end up with only a few more crunchers. This is a big undertaking. Good luck guys!! |
Send message Joined: 28 Jul 12 Posts: 819 Credit: 1,591,285,971 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Even when you get the windows app going, it looks like you're still going short on the number of crunchers by more than half, based on the server status page which shows currently 821 users crunching long units in the last 24 hours, while 34 are crunching quantum chemistry. (821/34 = 24.15) I am all in favor of GPU, but as noted on many project forums, it doesn't work for most problems. But don't write off Linux on the CPU yet. It is just in the startup phase. I have even taken my machines off until the production version is released. Once the word gets around (be sure to post a note on the BOINC forum), you will get lots of help. And CPUs are getting more cores all the time. |
Send message Joined: 2 Jul 16 Posts: 338 Credit: 7,987,341,558 RAC: 178,897 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() |
Even when you get the windows app going, it looks like you're still going short on the number of crunchers by more than half, based on the server status page which shows currently 821 users crunching long units in the last 24 hours, while 34 are crunching quantum chemistry. (821/34 = 24.15) Oh there will be many more if there is consistent work. The work inconsistency of GPU work pushes many away. Compare the support for pogs vs duchamp. Similar projects but duchamp requires vbox |
![]() Send message Joined: 24 Jul 08 Posts: 36 Credit: 363,857,679 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
I would say don't mess with a virtualbox application if it would replace the Linux application. Too many headaches. If someone is running Windows, they could easily set up their own virtualbox VM and run it under Linux with the standard app. Win-win for everyone that way. It also give the user more control over the VM. Just my thoughts on it. Also, more and more people are migrating to Windows 10 and it is the direction all new machines are following. So, might as well prepare for the future. ![]() |
Send message Joined: 28 Jul 12 Posts: 819 Credit: 1,591,285,971 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
I would say don't mess with a virtualbox application if it would replace the Linux application. Too many headaches. If someone is running Windows, they could easily set up their own virtualbox VM and run it under Linux with the standard app. Win-win for everyone that way. It also give the user more control over the VM. Just my thoughts on it. Also, more and more people are migrating to Windows 10 and it is the direction all new machines are following. So, might as well prepare for the future. My guess is that they could leave the Linux application as it is, and just add a VirtualBox application for Windows. I have had no particular problems with VBox on either Windows or Linux machines recently. I run LHC, Cosmology and sometimes others on it. I would prefer that they set it up so that I don' have to configure my own machine. All you really have to do is to first ensure that running a Virtual Machine is enabled in your BIOS. A good primer is on the Cosmology site: http://www.cosmologyathome.org/faq.php#vtx There is a much more elaborate checklist (if you need it) by Yeti on LHC: https://lhcathome.cern.ch/lhcathome/forum_thread.php?id=4161&postid=29359#29359 After that, you just install VirtualBox and attach to the project. It is all set up from there. |
![]() Send message Joined: 24 Jul 08 Posts: 36 Credit: 363,857,679 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
There is actually more to it than that. I have ran every VM project as well and have done so with a whole slew of different hardware and software setups. There is a huge reason why those projects do not get much support. LHC only has a large user base now because it merged the projects with the original Six Track project. Event still, the virtualbox applications remain less popular. Keep also in mind that these projects have problems with new releases of Virtualbox as they have right now. If I make my own VM using the latest release, it does not suffer the same. The only advantage to a vbox application is to allow the scientist to have an easier time compiling a single application. This may sound great to them, but the amount of lost time on the end user far exceeds their lost time. Also, for reference if it helps, GPUGrid attempted vbox applications back in 2014. Discussion started in 2013 http://www.gpugrid.net/forum_thread.php?id=3542#33874 ![]() |
Send message Joined: 28 Jul 12 Posts: 819 Credit: 1,591,285,971 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
If you have big problems, don't do it. That is not a reason they should not offer it. LHC would not exist in its present form without VirtualBox; neither would Cosmology and others. And if you prefer to set up your own VirtualBox machine, you can still do that in order to run the Linux version from a Windows machine in any case. I think you are arguing the wrong point. |
![]() Send message Joined: 24 Jul 08 Posts: 36 Credit: 363,857,679 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
I'm actually arguing for keeping the Linux version rather than replace it. Telling me to not bother because you like them isn't acceptable to me. You play it off like they run great because you had little issue with them. You can scour their forums over and over to find the average user does not agree. You are right about LHC not being in its present form as it would still be Six Track running traditional work and the others doing it in house or eventually adapting things different. Cosmology would just be down one application as well. I don't see how that is relevant. Either way. My vote is to not embrace virtualbox if it means pulling non-virtualbox work. ![]() |
Send message Joined: 28 Jul 12 Posts: 819 Credit: 1,591,285,971 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
I'm actually arguing for keeping the Linux version rather than replace it. Telling me to not bother because you like them isn't acceptable to me. You play it off like they run great because you had little issue with them. You can scour their forums over and over to find the average user does not agree. You are right about LHC not being in its present form as it would still be Six Track running traditional work and the others doing it in house or eventually adapting things different. Cosmology would just be down one application as well. I don't see how that is relevant. Either way. My vote is to not embrace virtualbox if it means pulling non-virtualbox work. I won't bother responding to fiction. |
![]() Send message Joined: 24 Jul 08 Posts: 36 Credit: 363,857,679 RAC: 0 Level ![]() Scientific publications ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
|
©2025 Universitat Pompeu Fabra