Thread 'BOINC Workshop 2018'

Message boards : News : BOINC Workshop 2018
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · 4 · Next

AuthorMessage
Richard Haselgrove
Volunteer tester
Help desk expert

Send message
Joined: 5 Oct 06
Posts: 5129
United Kingdom
Message 87255 - Posted: 24 Jul 2018, 11:27:25 UTC

Matt Blumberg - Grid Republic and Charity Engine
ID: 87255 · Report as offensive
Richard Haselgrove
Volunteer tester
Help desk expert

Send message
Joined: 5 Oct 06
Posts: 5129
United Kingdom
Message 87256 - Posted: 24 Jul 2018, 12:04:54 UTC

Steven Clark - Nanohub

And we break for lunch - back in 1 hour.
ID: 87256 · Report as offensive
ProfileJord
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 29 Aug 05
Posts: 15569
Netherlands
Message 87257 - Posted: 24 Jul 2018, 12:19:53 UTC

Can you ask those climate guys to stop making it so darn hot? We're not in Australia or the desert where these temps are normal. :-)
ID: 87257 · Report as offensive
Richard Haselgrove
Volunteer tester
Help desk expert

Send message
Joined: 5 Oct 06
Posts: 5129
United Kingdom
Message 87258 - Posted: 24 Jul 2018, 13:11:09 UTC - in response to Message 87257.  

Just been out for a brief wander round the shadier bits of Oxford - phew, it's hot out there. The building the climate guys live in (and which we're borrowing for the week) has a tall atrium, good airflow, and north-facing windows: it's very tolerable by comparison. Maybe they knew something in advance before they moved in?

Listen to David's talk: he had some good factoids. If it's still this hot on Thursday, the UK National Grid will have to reduce electricity supply by 20%, because the wires will go into thermal runaway. And we're running at 97% load already...

While we wait to reconvene - that GPU question. World Community Grid manage the distribution of work to volunteers, but they don't originate the research themselves. That is submitted by partner organistations around the world, and the partners provide the original software they want to be run. WCG can tweak it, but they don't go to the extent of re-writing it for another platform - or for GPUs. If the research partners provide GPU software, WCG can and have distributed it through BOINC - but none of their current projects use GPUs. Not all scientific problems can be designed to run as many simple parallel threads at the same time (which is what a GPU provides): researchers won't write for GPUs unless they have a problem which can be solved that way.
ID: 87258 · Report as offensive
Richard Haselgrove
Volunteer tester
Help desk expert

Send message
Joined: 5 Oct 06
Posts: 5129
United Kingdom
Message 87259 - Posted: 24 Jul 2018, 13:30:54 UTC
Last modified: 31 Jul 2018, 13:43:26 UTC

CM - Gridcoin
ID: 87259 · Report as offensive
Adam Radocz

Send message
Joined: 29 Jul 17
Posts: 5
Hungary
Message 87261 - Posted: 24 Jul 2018, 13:55:49 UTC - in response to Message 87258.  

In response to WCG. Why don't they share the project's source code to others be able to contribute, maybe further optimize or port to other platforms?
ID: 87261 · Report as offensive
Richard Haselgrove
Volunteer tester
Help desk expert

Send message
Joined: 5 Oct 06
Posts: 5129
United Kingdom
Message 87262 - Posted: 24 Jul 2018, 14:04:12 UTC - in response to Message 87261.  
Last modified: 24 Jul 2018, 14:41:33 UTC

In response to WCG. Why don't they share the project's source code to others be able to contribute, maybe further optimize or port to other platforms?
As in my last answer, it isn't their code: they would at the very least need permission from the providing researchers - but that's not insurmountable.

A possibly more significant problem is covered in the talk: WCG is run by IBM, and IBM is careful to protect its reputation. So, for example, even the researcher's code is security-audited before acceptance. But I'll try to get one of the WCG team to reply further.

OK, I asked Juan in the coffee break. WCG tweak the submitted code to work with BOINC (checkpoints etc.), but they are very careful to avoid touching the scientific algorithms. Most of the research code that is submitted to them has been in use for several years, and the researchers are comfortable with its accuracy and reliability. Any 'improvements' would need not only the security vetting, but also a scientific re-audit to be acceptable in a research context.
ID: 87262 · Report as offensive
Richard Haselgrove
Volunteer tester
Help desk expert

Send message
Joined: 5 Oct 06
Posts: 5129
United Kingdom
Message 87263 - Posted: 24 Jul 2018, 14:16:17 UTC

Team Blocklink AG
(Jan Schotta, Lorenz Langgartner, Maximillian Weigel)
ID: 87263 · Report as offensive
Richard Haselgrove
Volunteer tester
Help desk expert

Send message
Joined: 5 Oct 06
Posts: 5129
United Kingdom
Message 87265 - Posted: 24 Jul 2018, 14:33:20 UTC

The Innovation Grid
(Macauley Lynch, Richard Phillips)
ID: 87265 · Report as offensive
Juha
Volunteer developer
Volunteer tester
Help desk expert

Send message
Joined: 20 Nov 12
Posts: 801
Finland
Message 87267 - Posted: 24 Jul 2018, 15:00:33 UTC

Slides?
ID: 87267 · Report as offensive
Richard Haselgrove
Volunteer tester
Help desk expert

Send message
Joined: 5 Oct 06
Posts: 5129
United Kingdom
Message 87269 - Posted: 24 Jul 2018, 15:21:46 UTC - in response to Message 87267.  

... will be made available by the workshop administrators after it's all over. I've suggested to David Wallom that he should try to associate the slides with these recordings in the official report.
ID: 87269 · Report as offensive
Richard Haselgrove
Volunteer tester
Help desk expert

Send message
Joined: 5 Oct 06
Posts: 5129
United Kingdom
Message 87270 - Posted: 24 Jul 2018, 15:22:38 UTC

Patrick Schoefer - SETI.Germany
(about team competitions like their Pentathlon)
ID: 87270 · Report as offensive
Richard Haselgrove
Volunteer tester
Help desk expert

Send message
Joined: 5 Oct 06
Posts: 5129
United Kingdom
Message 87273 - Posted: 24 Jul 2018, 15:51:02 UTC

Ilia Chernov - Inst Applied Math
ID: 87273 · Report as offensive
Richard Haselgrove
Volunteer tester
Help desk expert

Send message
Joined: 5 Oct 06
Posts: 5129
United Kingdom
Message 87276 - Posted: 24 Jul 2018, 16:52:29 UTC
Last modified: 24 Jul 2018, 17:05:03 UTC

David Anderson - BOINC

We're moving to more general discussion for the last 15 minutes, I'll record it, but not bother to upload it if we ramble too much[*].

Tomorrow is more centrally devoted to BOINC itself, rather than the projects. I'll have to ask the participants again if it's OK to record them.

* Not really anything useful, except which pub we're adjourning to - the Bear in Blue Boar Street, for anyone local).
ID: 87276 · Report as offensive
Richard Haselgrove
Volunteer tester
Help desk expert

Send message
Joined: 5 Oct 06
Posts: 5129
United Kingdom
Message 87286 - Posted: 25 Jul 2018, 8:17:49 UTC
Last modified: 25 Jul 2018, 8:20:29 UTC

And here we go again. Recordings will appear on the same link as yesterday:

BOINC Workshop 2018 - talks audio

Just a brief intro to day 2 to start with.

And this is how we expect the rest of the day to run:

Day 2 schedule
ID: 87286 · Report as offensive
Richard Haselgrove
Volunteer tester
Help desk expert

Send message
Joined: 5 Oct 06
Posts: 5129
United Kingdom
Message 87287 - Posted: 25 Jul 2018, 9:20:30 UTC

Marius Millea - BOINC server using Docker
(technical - sorry, I can't show the live demo)
ID: 87287 · Report as offensive
Richard Haselgrove
Volunteer tester
Help desk expert

Send message
Joined: 5 Oct 06
Posts: 5129
United Kingdom
Message 87288 - Posted: 25 Jul 2018, 9:33:25 UTC

Kevin Reed - server release process
ID: 87288 · Report as offensive
Richard Haselgrove
Volunteer tester
Help desk expert

Send message
Joined: 5 Oct 06
Posts: 5129
United Kingdom
Message 87289 - Posted: 25 Jul 2018, 9:47:33 UTC

Client releases
(Kevin Reed and Richard Haselgrove)
ID: 87289 · Report as offensive
Richard Haselgrove
Volunteer tester
Help desk expert

Send message
Joined: 5 Oct 06
Posts: 5129
United Kingdom
Message 87291 - Posted: 25 Jul 2018, 10:29:51 UTC

Christian Beer - Software testing
(Software testing and Continuous Integration mechanism review)
ID: 87291 · Report as offensive
Richard Haselgrove
Volunteer tester
Help desk expert

Send message
Joined: 5 Oct 06
Posts: 5129
United Kingdom
Message 87292 - Posted: 25 Jul 2018, 10:45:12 UTC

Keith Uplinger - unit testing code

Break - we ALL need more coffee.
ID: 87292 · Report as offensive
Previous · 1 · 2 · 3 · 4 · Next

Message boards : News : BOINC Workshop 2018

Copyright © 2024 University of California.
Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.