Once again:
RNA World allows for disabling the very long tasks. So, everybody can decide on his/her own whether these highly complex tasks are acceptable or not - and the fact (!) that there are multiple people out there who already have successfully processed tasks that run for approx. a full year is a prime example on what distributed computing based on volunteers can do. It is remarkable! And it is THIS project that set landmarks.
By the way, we are very proud that we managed to implement checkpointing using the virtual machine approach. Actually, it was us who originally suggested this technology on a BOINC workshop more than 10 years ago. And it universally enabled the use of software in a volunteer environment which originally was designed for specialized HPC deployment, only. After CERN nicely implemented the idea, a lot of projects make use of it even today.
And finally: We are a team of laymen.
I had spoken to the so-called experts at different research institutes and universities in advance. They laughed at us when they heard what we were planning. But: We did it. And we did it well. And already at the initial setup of RNA World in 2009, this project had more compute power than RNA-focused research centers of the EMBL. Imagine that and, as one of our volunteers, be proud that you were part of it!
Michael.
P.S.: You mentioned the Yoyo@home project where some tasks require 8 GB of RAM. RNA World has CMCALIBRATE tasks that require more than 128 GB of RAM. Since the research experts (!) who developed the software we use in RNA World never thought of implementing a forecast for RAM requirement and since we, the laymen (!), never thought that someone could have the idea of developing software not taking possible limitations of RAM into account, initially we did not even know this. So we found out by try & error. After we identified the issue, we deployed these tasks to specialized machines. And we made sure that the original programmers got informed to re-work their software (which they did nicely).
Can't catch a WU
- Michael H.W. Weber
- Vereinsvorstand
- Beiträge: 22436
- Registriert: 07.01.2002 01:00
- Wohnort: Marpurk
- Kontaktdaten:
Re: Can't catch a WU
Fördern, kooperieren und konstruieren statt fordern, konkurrieren und konsumieren.
http://signature.statseb.fr I: Kaputte Seite A
http://signature.statseb.fr II: Kaputte Seite B
http://signature.statseb.fr I: Kaputte Seite A
http://signature.statseb.fr II: Kaputte Seite B
-
- Fingerzähler
- Beiträge: 2
- Registriert: 18.07.2019 20:40
Re: Can't catch a WU
Why don't you work units send much users?
Many users from our team (Czech National Team) want to count RNA, but even after a few years they didn't even get one work units.
Send each unit to a hundred computers and at least the job will be 100% complete to one year !!!
Many users from our team (Czech National Team) want to count RNA, but even after a few years they didn't even get one work units.
Send each unit to a hundred computers and at least the job will be 100% complete to one year !!!
- Michael H.W. Weber
- Vereinsvorstand
- Beiträge: 22436
- Registriert: 07.01.2002 01:00
- Wohnort: Marpurk
- Kontaktdaten:
Re: Can't catch a WU
That would actually be a massive waste of ressources.forest CNT hat geschrieben: ↑18.07.2019 21:23Send each unit to a hundred computers and at least the job will be 100% complete to one year !!!
Michael.
Fördern, kooperieren und konstruieren statt fordern, konkurrieren und konsumieren.
http://signature.statseb.fr I: Kaputte Seite A
http://signature.statseb.fr II: Kaputte Seite B
http://signature.statseb.fr I: Kaputte Seite A
http://signature.statseb.fr II: Kaputte Seite B
-
- Fingerzähler
- Beiträge: 2
- Registriert: 18.07.2019 20:40
Re: Can't catch a WU
On the part of the project, this would be a huge benefit. Also accelerate this phase.
Each user decides about the resources themselves, in which project they will put them. The project does not have to think about it.
Only two units were completed in two months. There are 55 left to finish. That's three for years. Is that what you really want? Or do you want to have results as soon as possible and start the next stage?
Each user decides about the resources themselves, in which project they will put them. The project does not have to think about it.
Only two units were completed in two months. There are 55 left to finish. That's three for years. Is that what you really want? Or do you want to have results as soon as possible and start the next stage?
Re: Can't catch a WU
The only resource that truly matters in life is TIME.Michael H.W. Weber hat geschrieben: ↑20.07.2019 18:48That would actually be a massive waste of ressources.