You can do it like me, and join other boinc projects.

There are some boinc projects, which are very happy about every CPU time and has adjusted there deadlines that older systems can easly compute it within the deadline.
There is an overview about the recomendations and crunching times of the different projects in the boinc wiki : http://boinc-wiki.ath.cx/index.php?title=Choosing_a_BOINC_Powered_Project

My old 800 MHz system doing SIMAP in about 3.5 hours (10 days deadline), malaria controll (beta project) in about 2.5 hours (3 days deadline) and TANPAKU (beta projec) in about 0.5 hours (6 hours deadline).
Predictor has also low crunching times and a good deadline.
At Rosetta you can configure the crunchintime from 1 to 20 hours (3 hours default), deadline is about 7 days.

Dotsch, it is very important to most people that their contribution is useful. Because of the way BOINC works, a lot of work was being wasted.

I can't stress enough that WCG won't miss results that have been completed weeks before. However, the percentage of people affected by this is very, very small, and it was more important to WCG to retain goodwill and increase efficiency by not wasting their members' work than to retain slower computers that can't keep up.

I think I was wrong about people with lots of BOINC projects. Normally, BOINC is clever enough to make the deadline and still maintain your project weightings. You will see BOINC work on WCG exclusively for a few days, then ignore it for a while.

If you have a slow computer and still want to contribute, then keep an eye on the Member News. Future projects may have lower system requirements. It's important to remember that the system requirements have a dual purpose: sure, they stop work taking too long and being wasted, but the most important reason is to protect members from a work unit that their computer can't handle, and prevent normal computer use from being affected.

Hope that cleared up a few points.

Ooops, I see I forgot to explain the central point. This isn't about the quorum. You need that to guarantee good science. But to get the quorum, BOINC sends out several work units. WCG used to send out 4 for a quorum of 3. This worked well, but a lot of work was wasted. Now, WCG send out only 3 for a quorum of 3, then send out additional work units only if something goes wrong.

Obviously, BOINC can't wait forever for a work unit. After a while, it has to assume the work unit isn't coming back, and send out another. This deadline change sets the deadline to the same time that BOINC gives up on the result, so if you meet the deadline, nothing is wasted.

Now, I can imagine what you're thinking: "why can't BOINC wait for a month or two before sending out extra copies?" The answer is twofold: BOINC only awards points when a work unit is validated. Reducing the replication to 3 already caused a lot more "pending validation" results, and the outcry was - well, a few people were unhappy. Having to wait months for points to be awarded would be far, far worse. WCG have to keep everybody happy, and this includes people who want lots of points. I think WCG have got the balance right, but it is (unfortunately) impossible to please everybody.

The second reason is purely practical. Results take up quite a bit of space on the WCG servers. While a result is pending, WCG have to keep the original work unit, and all the results that have been sent in so far. To keep the quantity of data manageable, WCG have to get batches finished off and returned to the project scientists within a reasonable time.

So, a few months ago when you returned a result after 2 or 3 weeks of crunching, BOINC had already got its quorum and your result really was wasted. I know this isn't what you want to hear, but (unlike some other projects) WCG have been entirely honest about this, and done their very best to fix the situation. Now, you can reassign your slower machines to a more useful purpose.

Your effort is still greatly appreciated, and I hope you will stay. It's not your fault that WCG are new to BOINC, and are still fine tuning to find the perfect balance. In fact, this is another advantage of WCG: with the grid techs concentrating on getting the grid technology absolutely optimal, the scientists can get on with doing science without any distractions.

Finally, Unix. I hope WCG do manage to add Unix as a supported platform. The only problem is there are so many Unices out there, WCG can never support them all. I think the only way we will get Unix support is if some large corporation offers their Unix boxes to the grid. Individual workstations are simply too rare to make the effort worthwhile. There's a lot of work involved in adding support for a platform, and to be quite honest, I think it was quite difficult to justify adding Mac support.

I do it for science, not for credits. And I have some problems to understand the reasons of the credit hunters....


Thank you for your clarifications about the situation of the project, and the backgrounds.
I hope, you did not misunderstood me. But your tone in your first posting sounded a little bit argoant, and without any background why what anyting done, it could/was misunderstood by some people (including me). It sounds like slow machines are not welcome.
I hope, that the project will announce future changes with more details and the background, that why was it done. From my expirance in other boinc projects, I have seen that a good communication is very needed.

But if I look at the supported OSes and the uptimes of my systems, which could run WCG, I will never complete within any deadline. So I have to stop crunching.
I think it is worse, because the project care more about the credit hunters and not about the people with older systems.


From my one practice in other BOINC projects, which also support the different UNIXes I have seen, that there is a lot of computing power available.
There are a lot of individuals, and companys which allow there employees to let run a boinc project which use the different UNIXes.
Also you should not forget, that a there are a lot of UNIX multi CPU systems out there, which have a hughe throughput, so the number of system is not so high, but they have a good RAC. - The last user I have helped at SETI had problems with a 16 and a 8 CPU Itanium...
For example, SIMAP has released Binaries for different Platforms, and the 28 UNIX systems have nearly the same throughput as 371 Windows 98 systems.
Simliar situation at SETI.

Porting to MacOS is not easy, in my oppinion Apple has to do a lot of work, to get the OS more closely to the other Unixes, and make it easier to port to there OS. From my sight it is difficulter to port from Linux to MacOS, that from Linux to HPUX, AIX, Solaris, *BSD.
If you have ported Linux, it must not be a hughe work to port it to HPUX, AIX, Solaris, *BSD. This OSes are quite simliar

So, I hope that WCG will support the big four (Solaris, HPUX, AIX, FreeBSD) anytimes.