A new type of work unit

[PcPerf bot]

We are preparing for the public release of a new work unit category: extra-large advanced methods work units. Some background information is provided below; we'll update this thread with more details as they emerge.

 

 

 

 

Why a new work unit category?

 

 

We have some specific projects where we 1) have large simulation systems and 2) want to get results fast. As multi-core processors get more powerful, we can perform calculations on Folding@Home that

 

 

previously required supercomputing clusters.

 

 

 

 

 

 

What's different about these work units from the donor perspective?

 

 

These work units are special SMP work units that have larger upload and download sizes, shorter deadlines, and require more memory and CPU resources. That's why we've created a new category.

 

 

 

 

 

 

Is there any points incentive for running these work units?

 

 

The base value of these work units corresponds roughly to what an SMP work unit using the A2 core would yield on an equivalent calculation. However, because fast completion is a scientific priority for these work units, we are doing a **trial** of a new bonus scheme where faster WU completion yields a points bonus.

 

 

 

 

 

 

What systems can run these work units?

 

 

Right now, only Linux and OS/X systems can run these work units, and they require 8 or more cores. We prefer 8+ *physical* cores, although fast Core i7 machines that are dedicated folders have proven

 

 

sufficient during the testing process. The points incentives are designed to match appropriate resources to points value; if your machine is marginal for the extra-large work units, you're probably

 

 

better off running standard SMP.

 

 

 

 

 

 

Does this have any relation to the large-points value work units and recent high-scoring users?

 

 

Yes. The initial projects are 2681 and 2682, valued at ~25K points base. Although these point values seem high, the work units are correspondingly larger, so the base PPD (points per day) value is

 

 

roughly comparable to standard SMP.

 

 

 

 

 

 

A collaborator has donated a large amount of compute time to this project; those clients were initially running under username Anonymous/team 1. To give proper credit for the donation, we have

 

 

changed the username to PDC, team 1. During the period of this donation, there are at any time between 100 and 400 8-core clients running under this username (800-3200 cores total).

 

 

 

 

 

 

Please stay tuned for further details regarding the upcoming release.

 

 

 

 

Voir l'article complet

[foo_fight]

La traduction ici -> http://www.fah-addict.net/news/news.php?id=29

[Thor]

Il marches pas le lien

[foo_fight]

Le site devait pas être accessible à ce moment là car à cet instant ça fonctionne.

[Thor]

Oui mais non, en tout cas chez moi il pointes sur la page d'accueil.

Le lien vers la news c'est http://www.fah-addict.net/news/news-0-29.php

[foo_fight]

Ah oui zut je me suis planté...