Message boards :
Number crunching :
The Great LHC Hold-Up
Message board moderation
Author | Message |
---|---|
Send message Joined: 24 Oct 04 Posts: 79 Credit: 257,762 RAC: 0 |
I believe we have been waiting these 2 weeks for new work when sooooo much was promised because of the "0" problem on different platforms.... but maybe I was wrong.... Twould be a nice gesture from the physicists to explain to us laymen what happened in a few short words even. As being a general partner in this collabaration a bit of info is always appreciated :) [Edit] And oh thank-you much Markku for the update on the when. |
Send message Joined: 1 Oct 04 Posts: 1 Credit: 2,941 RAC: 0 |
I found this interesting to read on what has been going on at CERN concerning LHC@home. explains what was going on,and where things are going...... kinda long, just scroll through the sections that aren't of interest., but readable. Christian Søttrup and Jakob Pedersen worked furiously all spring and summer to get SixTrack and BOINC to function together. You can read their thesis, which describes the opportunities for combining public resource computing, such as LHC@home, with Grid computing like the LHC Computing Grid. Here's the link to the paper: http://www.fatbat.dk/thesis/ click on the THesis PDF download link to read it. |
Send message Joined: 2 Sep 04 Posts: 545 Credit: 148,912 RAC: 0 |
Thanks, added it to the Wiki :) It is in the section of the site on LHC@Home, though, that listing is under the Beta Projects as they have nt yet admitted they are doing production ... at least that is what the first page always tells me ... |
Send message Joined: 2 Sep 04 Posts: 19 Credit: 26,799 RAC: 0 |
> Here's the link to the paper: > > http://www.fatbat.dk/thesis/ > > click on the THesis PDF download link to read it. > Yes lots of interesting reading there. This caught my eye. No need for optimised apps for LHC I think. "The LHC simulation that was chosen to be the basis for LHC@home was SixTrack. SixTrack has been developed over the last 20 years by Frank Schmidt, and is used by CERN and other synchrotrons3 around the world to do beam studies. Like much of CERN's older software it is written in Fortran. On the initiative of Eric McIntosh it is also used as part of the SPEC4 floating point benchmark. So CERN, when buying new computers, never have to estimate how well a new processor will run SixTrack, but can look it up directly. Another added benefit is that manufacturers will optimize their hardware and compilers to get the best benchmark score, so modern CPUs and compilers are directly optimized for SixTrack performance." Get BOINC WIKIed |
Send message Joined: 2 Sep 04 Posts: 19 Credit: 26,799 RAC: 0 |
|
Send message Joined: 2 Sep 04 Posts: 309 Credit: 715,258 RAC: 0 |
More interesting snippets..... "At the time we started this analysis, an average WU would take about one hour to compute on a 2 GHz P4 machine. This number is also highly influenced by the options the physicist specify. During our work with the application we had to change Fortran compilers many times, sometimes just versions and sometimes also the compiler manufacturer, and each time the execution time for a WU became smaller. The version we finally used for the public had almost halved the time an average WU takes." "LHC@home delivered more than 600,000 finished jobs to SixTrack during our public run. For a comparison CPSS has delivered 200,000 finished jobs during its lifetime, which is more than a year. This means that LHC@home has delivered more than three times the amount of results in a quarter of the time as CPSS has. Furthermore the complexity of the jobs run on LHC@home has been much higher." "At the end of the project, we experienced a lot of performance issues with the database server, so the LHC@home hardware will also be upgraded during this period. The newest types of server machines are not put directly into production use at CERN, but are tested for a period of time to measure the stability. Luckily, the LHC@home project has been chosen to be a part of this test, so we can exchange our deprecated servers with some of the newest and fastest, and we hope this will increase the amount of users the system is able to handle." "The positive result of the project which was initially an experiment led by a small sub group of the IT division at CERN has impressed many people at CERN. So the decision has been taken to make BOINC a more permanent facility at CERN. BOINC will probably be implemented as one of the services the IT division offers. This means that an experiment can approach the IT division and have their simulations ported and run through the IT divisions BOINC servers." "For a comparison LCG-2 consists of approximately 10,000 CPUs, whereas LHC@home could deliver the equivalent of 360 dedicated CPUs and it was only so small because our server capacity had limited the amount of users to 6000. We believe that it is possible to extend the size of the user base to at least the 500,000 users that SETI@home has." These were some of the postive points in my skimming of the thesis....there were some not so positive points (but who wants to hear about those....) ;) Live long and crunch. Paul (S@H1 8888) BOINC/SAH BETA |
Send message Joined: 2 Sep 04 Posts: 545 Credit: 148,912 RAC: 0 |
> These were some of the postive points in my skimming of the thesis....there > were some not so positive points (but who wants to hear about those....) ;) Me ... I think that knowing the good and bad is important. The difficulty we find ourselves in so often is that it is so easy to let a minor bit of froth to taint the whole soup ... That is why I tried to write that eloquent little number on SETI@Home as we start the monthly BOINC is not as good as classic because ... |
Send message Joined: 17 Sep 04 Posts: 190 Credit: 649,637 RAC: 0 |
... > ...start the monthly BOINC is not as good as classic because ... > Hm really? I see some difficulties to do LHC work with the "seti@classic client" ;) Boinc is basically not as bad as "some" people want to make believe us. There is more than only good or bad.. It would be interesting to turn the cow and list why "boinc" is "better" overall seen. |
Send message Joined: 2 Sep 04 Posts: 545 Credit: 148,912 RAC: 0 |
I was just saying, it seems every couple weeks, roughly once a month, we get a thread yelling that SETI@Home Classic is perfectly fine and that they hate BOINC because ... So, we try to find out why they think BOINC is so bad, and it USUALLY boils down to the fact that it is different and does not appear to be as "easy" to use. The fact that the Classic application was a bear to run unless you had one or two "helper" programs is ignored. The fact that you had almost no information about the system and how it was performing is ignored. The perception that BOINC is less reliable is usually due to the fact that we now have greater visibility as to what is going on is used to "prove" Classic is better. Mostly it is because they don't see any reason for change. And most people don't like change that much. |
©2024 CERN