Ltop

ltop@nikhef.nl

16 discussions

Recent news
by Rob Lambert 10 Feb '14

10 Feb '14

Hi Wilco, etc. Some useful progress from my side: a) We now have a configuration of the trigger (Moore) which runs only the tracking sequence, which should be a nice compliment to any tests we want to do in Brunel (the much simpler reconstruction application). b) We also have gotten together a set of "high-occupancy test files" which should be very useful for this proof-of-principle, and eventual profiling, examining memory footprints/patterns, since this is when it should be the most stressed out with lots of hist and tracks per event. c) We're thinking of providing a script where you can slice out an algorithm from a complicated Gaudi set of options for stand-alone running, which should further simplify the making of a sandbox (but since you're probably changing the data structures themselves it is unclear how useful this will be for you). Perhaps we should have a 15-minute coffee meeting in the next few days to discuss our first targets? Thursday or Monday would be good for me. Cheers, Rob ------------------------------------------ Robert Lambert FOM-VU-NIKHEF-Bfys LHCb Email: rob.lambert(a)cern.ch ------------------------------------------ Nikhef N251 Tel: +31 20 592 2131 Fax: +31 20 592 5155 ------------------------------------------ CERN, 13-1-018 Tel: +41 22 767 4024 Fax: +41 22 766 8109 ------------------------------------------

4 9

Tutorial pages etc.
by Rob Lambert 23 Jan '14

23 Jan '14

Hi Wilco, Summary of today's meeting: a) Installation of slc6 in changeroot environment on your laptop b) Installation of cvmfs following instructions from our internal IT guys i.e. sit at the helpdesk for a while (the online instructions are rubbish) c) Go through the simple tutorials briefly. The LHCb software tutorials are linked through the LHCbComputing page: http://lhcb-comp.web.cern.ch/lhcb-comp/ https://twiki.cern.ch/twiki/bin/view/LHCb/LHCbSoftwareTutorials You will want to read the following tutorials before we meet next week. https://twiki.cern.ch/twiki/bin/view/LHCb/LHCbSoftwareTrainingBasics https://twiki.cern.ch/twiki/bin/view/LHCb/LHCbSoftwareTrainingPrinting You take these tutorials by simply reading through the slides, and then the twiki. Reading through is good enough for now. You will also probably want to read the following guidelines. https://twiki.cern.ch/twiki/bin/view/LHCb/LHCbCodingGuidelines https://twiki.cern.ch/twiki/bin/view/LHCb/SVNUsageGuidelines There are no further tutorials specific to what you will be doing. All the other tutorials are aimed at data analysts not software developers. We keep our documentation in the following places: - Doxygen. One website per project. . e.g. Brunel http://lhcb-release-area.web.cern.ch/LHCb-release-area/DOC/brunel/releases/… . e.g Moore http://lhcb-release-area.web.cern.ch/LHCb-release-area/DOC/moore/releases/l… - Twiki. . fully googleable so long as you add CERN TWiki LHCb into the google search. i.e.: http://lmgtfy.com/?q=CERN+TWiki+LHCb+RawEvent . Starts here: https://twiki.cern.ch/twiki/bin/view/LHCb/WebHome - Savannah task tracker: . https://savannah.cern.ch/projects/lhcbcore/ - JIRA task tracker: . https://its.cern.ch/jira/browse/LBCORE - Browsing the repository: . svnweb: https://svnweb.cern.ch/trac/lhcb/browser . authenticated access: svn+ssh://svn.cern.ch/reps/lhcb . anon access: http://svn.cern.ch/guest/lhcb That's all for now, you probably want to bookmark all those links somewhere. Cheers, Rob ------------------------------------------ Robert Lambert FOM-VU-NIKHEF-Bfys LHCb Email: rob.lambert(a)cern.ch ------------------------------------------ Nikhef N251 Tel: +31 20 592 2131 Fax: +31 20 592 5155 ------------------------------------------ CERN, 13-1-018 Tel: +41 22 767 4024 Fax: +41 22 766 8109 ------------------------------------------

1 0

CERN account, tutorial, etc?
by Rob Lambert 21 Jan '14

21 Jan '14

Since this mailing list is short enough I thought may as well chat about these details here. CERN account: ----------------- Wilko will need a CERN account with LHCb (z5) to ensure everything can run smoothly for him. I'm not sure if Jeff is able to set that up for you or if we should go through Marcel (He's back on Thursday). We should try and sort this account this week so that we can go over some basics early next week. Brief introduction to LHCb software: ------------------------------------------ As I'm *that guy* who gives the LHCb software tutorials we should make a slot to go over the basics of LHCb software and software environment. It will take around about half a day, and I think early next week is good for that. Of course this will be the real initial starting point vaguely covering the structure, conventions and framework of our software. There's going to be much more to see and play with *under the hood* as it were. Let me know when you want to do that, I'd prefer if you already had a cern account before that, so a morning or afternoon early next week. Cheers, Rob ------------------------------------------ Robert Lambert FOM-VU-NIKHEF-Bfys LHCb Email: rob.lambert(a)cern.ch ------------------------------------------ Nikhef N251 Tel: +31 20 592 2131 Fax: +31 20 592 5155 ------------------------------------------ CERN, 13-1-018 Tel: +41 22 767 4024 Fax: +41 22 766 8109 ------------------------------------------

2 1

Fwd: REMINDER : Registration is open: Second Thematic CERN school of computing: June 2014, Split, Croatia
by Gerhard Raven 17 Jan '14

17 Jan '14

FYI — data-oriented design is an explicit topic of this school… — Gerhard Begin forwarded message: > From: Andrzej Nowak <andrzej.nowak(a)cern.ch> > Subject: FW: REMINDER : Registration is open: Second Thematic CERN school of computing: June 2014, Split, Croatia > Date: 17 Jan 2014 12:58:00 GMT+1 > To: "rd-multicore (General announcements and discussion forum for the Multi-...)" <rd-multicore(a)cern.ch>, "lhcb-manycores (Information exchange on many-core computing in LHCb)" <lhcb-manycores(a)cern.ch> > Cc: Alberto Pace <Alberto.Pace(a)cern.ch> > > Dear all, > > Please find attached the announcement from Alberto Pace, concerning the Second Thematic CERN School of Computing (independent of the main summer CSC 2014). All participants are welcome, in particular the software developers and programmers in LHC experiments. > > Regards, > Andrzej Nowak > > From: Yasemin Hauser > Sent: 17 January 2014 11:21 > > > REMINDER : > only 2 weeks left to register !!! > > > From: Alberto Pace > Sent: 06 December 2013 15:59 > To: Computing School > Subject: Registration is open: Second Thematic school of computing: June 2014, Split, Croatia > > > Dear colleague, > > The registration is open for the 2014 thematic CERN School of Computing focused on future high-throughput scientific computing on the topic of developing the next generation physics analysis applications that benefits from the new emerging hardware architectures. > > This tCSC2014 will last one week and take place the 16-20 June 2014 in Split, Croatia and is independent from the main 2014 CERN School of Computing (to be held in August). See http://cern.ch/csc for all CSC upcoming schools. > > The thematic school can be attended by any participant, including former and future CSC participants of the main summer school. Registration will close on January 31st, 2014. We are fully aware of the difficult financial period of many public institutions and we managed to arrange a very reasonable fee (550 EUR). Be aware that also this year the school will accept a limited number of participants. > > The theme of the 2nd thematic CSC is “future high-throughput scientific computing”, covering: > 1. Programming for Concurrency: Modern and performing C++, Expressing parallelism pragmatically, Resource protection and thread safety > 2. Data oriented design: Designing for data, Data-intensive applications, Vectorization > 3. Memory programming: Memory effects in hardware, Choosing data structures, Non-uniform memory > 4. Acceleration: Principles of programming for accelerators, Direct, Offload, Symmetric mode, MIC > 5. Efficient computing: Architecture refresher (big and small cores), Accurate and efficient floating point, Compilers: their strengths and weaknesses, Advanced performance monitoring and tuning > > Here are the links to the detailed schedule and to the practical information. To apply to the school, please fill the registration form (and ensure you have done a login to indico using your CERN account or a lightweight account as explained in the registration page). > > You are most welcome to pass-on this information to colleagues potentially interested, within or outside CERN. I also apologize to those who will receive this announcement through multiple channels. > > Best Regards. > Alberto Pace > Director, CERN School of Computing > > For enquiries and correspondence related to the school, contact information is available at: http://csc.web.cern.ch/contacts

1 0

Monday 3rd Feb
by Rob Lambert 17 Jan '14

17 Jan '14

Hi All, There is a trigger-optimization session in the next trigger workshop. > https://indico.cern.ch/conferenceOtherViews.py?view=standard&confId=293170 (I think it's probably OK to tell you the LHCb password for that, but I'd prefer not to email it around myself, so pls pass by for the password). 9AM->12:30. I think that might be a good idea to see the current state-of-play and the key players inside LHCb working on similar (orthogonal) problems. You can connect remotely and simply listen in. Probably not everything is useful and you might want to have it playing as "background noise" most of the time, but anyway, it's there if you want it ;) Here is a summary of the session: ---------------------- Monday, 3 February 2014 09:00 - 10:45 Tracking 2015 09:00 Tracking performances a la 2015 20' Speaker: Maurizio Martinelli (NIKHEF (NL)) 09:20 Improving VeloTT 20' Speaker: Espen Eie Bowen (Universitaet Zuerich (CH)) 09:40 What Can a Thoughtful Man Hope for 2015 Tracking, Given the Experience of the Past Year of Upgrade Tracking 20' Speaker: Michel De Cian (Ruprecht-Karls-Universitaet Heidelberg (DE)) 10:05 2015 tracking preparation 20' Speaker: Maurizio Martinelli (NIKHEF (NL)) 10:55 - 12:35 Tracking Upgrade 10:55 VeloPix 20' Speaker: Heinrich Schindler (CERN) 11:15 Upgrade Pattern Recognition 20' Speaker: Thomas Nikodem (Ruprecht-Karls-Universitaet Heidelberg (DE)) 11:35 Everything you wanted to know about auto-vectorization 20' Speaker: Stefan Lohn (CERN) 11:55 Make your code faster ! 20' Speaker: Tim Head (CERN) 12:15 Track Fitting 20' Speaker: Wouter Hulsbergen (NIKHEF (NL)) ----------------------------------- Cheers, Rob ------------------------------------------ Robert Lambert FOM-VU-NIKHEF-Bfys LHCb Email: rob.lambert(a)cern.ch ------------------------------------------ Nikhef N251 Tel: +31 20 592 2131 Fax: +31 20 592 5155 ------------------------------------------ CERN, 13-1-018 Tel: +41 22 767 4024 Fax: +41 22 766 8109 ------------------------------------------

1 0

Toys, talks, documentation, examples, ...
by Gerhard Raven 16 Jan '14

16 Jan '14

First Post! The talk I showed this morning is available at https://indico.cern.ch/contributionDisplay.py?contribId=2&sessionId=0&confI… (the CERN indico for the last LHCb computing week, which was the 1st week of November) For those that don’t have access to that, I’ve attached a copy — note that it has more than just C++11 and ‘data layout’; feel free to ignore those other bits (and on the 14th slide, thinking about it for a 2nd time, it is better have two pointers in the ‘red’ stuff, one to the blue and one to the green directly, and pair up the offsets and thus creating a ‘zipped’ iteration; anyway, just an example that there is room for obvious improvement compared to what I wrote). The talks I mentioned can be found at http://channel9.msdn.com/Events/GoingNative/2013 esp. the talks from Sean Parent (Adobe) and STL (MS) & Alexandrescu (Facebook) are inspiring. Then there is http://channel9.msdn.com/Events/GoingNative/GoingNative-2012 and https://www.facebook.com/notes/facebook-engineering/three-optimization-tips… An example of what looks to me as very clean C++11 code are the ‘ranges’ introduced here: https://github.com/ericniebler/range-v3 — this is work in progress, but Eric Niebler is the author boost::proto, which I wish I could understand in detail, as it is probably one of the most clever bits of C++ ‘out there’. Then there is an entire website devoted to ‘data oriented programming’ — not surprisingly, it is http://www.dataorienteddesign.com/dodmain/ (go one level up for several more links), and there is http://gameprogrammingpatterns.com and finally, the toy example (which has morphed around a bit) is attached as a tar.gz file. As mentioned it was inspired by ArrowStreet (see https://github.com/ExaScience/arrow-street and https://indico.cern.ch/contributionDisplay.py?contribId=475&sessionId=7&con… Then there is some documentation on the algorithm which I find the most suited to ‘attack’, namely the ‘forward patter recognition’ here: https://twiki.cern.ch/twiki/bin/view/LHCb/PatForward The code lives eg. at http://svnweb.cern.ch/world/wsvn/lhcb/Rec/trunk/Tf/PatAlgorithms/src/?#a3ed… and to be concrete, what matters is everything invoked from the ‘execute’ method of http://svnweb.cern.ch/world/wsvn/lhcb/Rec/trunk/Tf/PatAlgorithms/src/PatFor… I guess that covers most of what we discussed this morning ;-) Cheers, — Gerhard

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

Ltop