Latest boinc version thread

[Paladin]

Heh, just messed with the Ubuntu 64-bit one a couple days ago, worked fine on Gentoo and should work for any other distro.

I found that it breaks Synaptic if you force a newer version that it has in it's database. Tells you to do a "sudo apt-get -f" (deleting the newer BOINC) before it will allow anything in the entire system to be updated.

[sierra_bound]

5.10.43 beta is now available for Windows, Mac OS X and Linux.

http://boinc.berkeley.edu/download_all.php

[sierra_bound]

5.10.44 beta is now available for Windows users.

32-bit version

64-bit version

[sierra_bound]

5.10.44 beta is now available for Mac OS X users.

Standard GUI

Unix command-line version

[sierra_bound]

5.10.44 beta is now available for Linux/Ubuntu users.

http://boinc.berkeley.edu/download_all.php

[sierra_bound]

5.10.45 beta is now available for Windows and Linux users.

http://boinc.berkeley.edu/download_all.php

[ownage]

I saw there are several optimized Boinc versions for SETI. Are there no Boinc versions optimized for WCG?

I still use 5.10.30, still OK, or should i install a newer one?

[sierra_bound]

The problem with using optimized clients is the BOINC quorum. If your results are way out of line with those of other quorum members, you run the risk of having your result declared invalid. Or in the case of two-member quorums, the lower of the two results will often be used for granted credit.

There are better options than optimized clients. Running a 64-bit OS and client will give you a boost in production. And you won't necessarily be punished in the quorums.

As far as which client version to use, each person has his or her favorite. You just have to try other versions to see if they are better than the recommended client.

[sierra_bound]

5.10.45 beta is now available for Mac OS X users.

Standard GUI

Unix command-line version

[rcofell]

About the optimized clients, the only reason to use an optimized [boinc] client is to match up with an optimized science application, so that your reported score(benchmark * time) is somewhat accurate when compared to someone doing the workunit with the standard science app. A workunit should be worth the same no matter how it's crunched, no? The optimized client evens things back up by boosting the benchmark, usually by compiling with SSE(1/2/3/4) optimizations through the compiler.

Granted there's the fact that the existing benchmark system isn't so great to begin with, but how else could you do an accurate comparison when it's a unique workunit being crunched by only you? The best would be a benchmark built into the specific workunit that's based on of the algorithms actually being used, but that would probably be too complex to implement, plus you'd have all the complaints about inter-project scoring... Truth be told, running a large quorum is actually the most fair of all under the existing system, since you're being paired up with a good mixture of the other computers running the project and you're all deciding on what the fair value for a workunit is. Despite the attitude sometimes expressed here, you're not getting cheated out of anything. Sure it may seem like other computers are dragging your score down, but the truth is if you have a faster computer you would have finished the workunit faster and will be able to get more done in the same time period as a slower computer. Total throughput is the factor here, score per workunit(should be constant across different machines to be fair as a scoring system) * workunits per day, much like processor speed is based on the balance of architectural efficiency and clock speed

This works out for seti@home, since the source code for the science app is open source and people are allowed to optimize it; this would mostly include taking advantage of newer processor instructions like SSE, or I suppose any other optimization techniques, like inline code perhaps. Now in the case of WCG, all the projects being run are closed source, so we can't tamper with the code and optimize it, even with just compiler options. This is understandable, since we're mostly dealing with medical computations, where accuracy is very important(which the quorum system should help with too).

It would be great if these projects would optimize for SSE, but the fact is they have to make it so the program runs on practically any x86 processor(i386 you could say), so it's pretty unoptimized for modern processors. Of course, boinc added the capability to read cpu instruction flags later on, so this does give the ability to select what's the most optimized science app a computer can handle, but I'm sure it won't be much use until a majority of users have a version supporting that(not sure what version it was added, maybe 5.10.x).

So this basically leaves us with this, if you run an optimized boinc client without an optimized science app, then you gain nothing. It could be considered cheating in a sense, since you're knowingly(hopefully) inflating your score, somewhat like the lazy person claiming he did the most work... It doesn't really make anything go faster computation wise, maybe the boinc client will be a little snappier, but that has almost no bearing on the workunit being crunched.

About upgrading clients, technically there should be nothing wrong staying back at 5.10.30(or probably even 5.10.23). It's just what you prefer, I like to stay up to date if I can, but sometimes I don't pay attention and run behind. New features are always nice, but there hasn't been anything major added, I think they're just maintaining the 5.10.x code base until they release 6.x.

I didn't intent on writing a book on this, but it was quite an issue a while back and thought I should voice my thoughts on it

P.S.
I tried running all the 6.1.x clients after 6.1.0 and they all seem to have that CA certificate issue with WCG only, since they're the only project to require a secure connection. Word is it'll be fixed in 5.10.9, so in the meantime stick to 6.1.0 or a 5.10.x release.

[sierra_bound]

People searching for space aliens need all the help they can get.

[Hymay]

Truth be told, running a large quorum is actually the most fair of all under the existing system, since you're being paired up with a good mixture of the other computers running the project and you're all deciding on what the fair value for a workunit is. Despite the attitude sometimes expressed here, you're not getting cheated out of anything. Sure it may seem like other computers are dragging your score down, but the truth is if you have a faster computer you would have finished the workunit faster and will be able to get more done in the same time period as a slower computer. Total throughput is the factor here, score per workunit(should be constant across different machines to be fair as a scoring system) * workunits per day, much like processor speed is based on the balance of architectural efficiency and clock speed

Now, if only they could get the computers that can complete the WU in similar times to actually claim a similar value... I'd be relatively happy with that.

[rcofell]

Yeah, it's never spectacular using a synthetic benchmark to judge real world performance, especially when it varies between boinc versions... I take it the whetstone/drystone benchmark doesn't really stress the memory subsystem very much, whereas a science app may, or else the real mixture of integer and floating point operations. I'll have to admit it looks like my claimed scores are usually not too far off from the granted, at least within 20%, usually 10%. Using a larger quorum should even out the results, since they're basically taking the mean(with some outliers discarded) of your result and a decent of the project population(well, better than just you and another machine). Of course, I wonder if it would even be worth scrutinizing single workunit results on quorums of just two, since over time the law of large numbers should take care of discrepancies. I hate to advocate statistics, the word makes me shudder(I took a random processes course instead ), but that's basically that's what this becomes; the long term results should be what we're focused on, not the short term. Perhaps one day a new method will be found to calibrate the scoring, a good way would be to have each project require you to crunch a short benchmark workunit before being allowed real workunits to download and work on. That would probably yield much more consistent results and keep scoring pretty calibrated, unless the workunits vary that much; they do in execution time, but I'm not sure about the execution characteristics(instruction/algorithm usage ratios).

I wish more information on results was given to us, like the specs of the other machines in the quorum... It would be nice to have a database of what different machine specs are getting long term, but you'd have to be careful and separate by projects run. It should be realized that by running multiple wcg projects it's very similar to running different boinc projects, since each project has its own science app and thus own performance characteristics. I'm also curious how workunits are actually split up, if they keep all the windows ones separate from linux, along with x86 vs x86_64, or if they overlap them at all and it's just the client that varies.

Again I typed more than I expected, lol. I don't think it's just this screen being smaller than my normal one.

[sierra_bound]

6.1.10 beta is now available for Mac OS X users.

Standard GUI

Unix command-line version

[sierra_bound]

5.10.45 is now the recommended version for Windows, Mac OS X and Linux users.

http://boinc.berkeley.edu/download_all.php

FYI, I've been using this version on one machine for about a week. Seems fine.

[karl_eller]

Just a quick question: How do you upgrade from one version to another (eg 5.10.30 to 5.10.45) without losing current work units? Is it just a matter of installing the new version to the same folder as the old one, and let Boinc do the rest?

This is a question for both Windows (Vista) and Ubuntu (where Boinc is installed by a .sh file, not the dep package).

Eller

[sierra_bound]

Just install it on top of the old version. That's what I always do. The installer will remove files that are no longer needed.

[[XC] NetburstXE]

About the optimized clients, the only reason to use an optimized [boinc] client is to match up with an optimized science application, so that your reported score(benchmark * time) is somewhat accurate when compared to someone doing the workunit with the standard science app. A workunit should be worth the same no matter how it's crunched, no? The optimized client evens things back up by boosting the benchmark, usually by compiling with SSE(1/2/3/4) optimizations through the compiler.

NO. Then my sh1tty Pentium M would get as much credit for completing a task as an E8400, it's just that the E8400 could complete 3 times as many workunits in a given period of time, even on only one core. Another project I participate in, Primegrid, has an app (PSP Sieve) that always gives the same amount of credit, regardless of CPU time. My Pentium III gets the same 4.99 points as my friend's E6300 for completing a task. It takes my Pentium III 40 minutes to complete a task, and the E6300 12. This seems hardly fair, as my Pentium III is still working as hard as it can.

[rcofell]

So you're basically saying you're basically in favor of getting points based on how much hardware you have running it, regardless of speed? Recycling old hardware and giving it a continued use is a good thing, but the goal of these projects is to crunch the most data the fastest, so logically the scoring system could be based on how you contribute to it in terms of work done. Besides the fact that running a bunch of old machines that run slower than a few new ones wouldn't make sense once you take into account energy costs and the such, hence I have a bunch of PIIIs I don't run(well, plus I'm at college and I prefer to sleep at night too ).

Sure, it'd be nice to involve effort somehow, but that's always a tough metric to measure. Of course it could also be said that if you're putting forth that much effort into a project you probably aren't in it for the points and just care for the science(how many times have I heard this before...), I couldn't say for sure though. Granted points are pretty much a game though, just a friendly competition that happens to help the project out, it gives people a reason to spend money towards the hobby Seems awful free market to me, another thing people tend to have a problem with sometimes

[The Warlock]

My vista machine did NOT continue work after installing 5.10.45, hmmm, lost hours of work. But the xp machine continued on as if nothing happened.

[sierra_bound]

6.1.12 beta is now available for Windows 32-bit users

Download link

[Philly_Boy]

Says it is not valid for XP 32-bit...

[sierra_bound]

The installer says that? I have not tried it yet. It's technically an alpha version. When the middle number is odd (as in 1), that means alpha. An even number indicates beta or recommended version.

[Sunfire]

It's working with Vista x86, but it's always showing the graphics window
Back to 5.10.45...

[sierra_bound]

6.1.14 is now available for Windows 32-bit users.

Download link

This is a development version, so use at your own risk.