[Shaktai]

Well it finally happened. Of course it would while I am on an almost two week business trip.

http://setiboinc.ssl.berkeley.edu/sah/
This is the URL for the new "IT's ALIVE" SETI running on Boinc. This is a great chance for Team MacNN SETI crunchers to get on board at the very beginning of a project and set the pace for all the other Mac teams. Team MacNN is already set up. Be sure to use the latest BOINC version, 3.18. Earlier versions will not get work.

All MacNN Seti crunchers should begin moving over. SETI Classic will remain active for a time, but if we want to be the Mac pace setters, now is the time to do it.

Of course if SETI isn't your cup of tea, there are plenty of other worthwhile projects that can use all the crunch power you have to spare. Just CRUNCH, CRUNCH CRUNCH...and do it for Team MacNN.

[Scotttheking]

Continue goal of top 100 in f@h project, or go for boinc, ...

Ok, how do I set it up, do I need to make a new user or does my seti1 account work, what OS / hardware works best, blah blah blah.

[Shaktai]

Originally posted by Scotttheking:
Continue goal of top 100 in f@h project, or go for boinc, ...

Ok, how do I set it up, do I need to make a new user or does my seti1 account work, what OS / hardware works best, blah blah blah.

The short version is that all SETI classic accounts have been transferred over. From the homepage select "Particpate in Seti@Home"

Select the option:
If you participated in the original 'SETI@home Classic' project:
We've transferred your account, including profile and team information.
You must now activate the account. (click the link) Note: the address of your original account must be current (or correctable via your old SETI log-in/password) so that you can get your SETI project key by e-mail. Otherwise you will need to re-register

BOINC is multi-processor friendly and likes Linux and Macs. It really likes G5's and Athlon 64s. It is okay now with Windows. The only GUI/Screensaver is Windows, but that is expected to be "redone" so that all clients are CLI with a seperate GUI/Screensaver front end. Most of the delays have been due to either Windows issues or backend stuff. The Linux and Mac clients have been stable for 2 or 3 months now (but still improved).

Because the source code is available, it is possible to create optimized "core clients" (the BOINC framework) for various hardware/software combinations and that is effective. The SETI application is not optimized for any platform, and as before likes lots of memory bandwidth. Expect possible additional SETI optimizations, depending on "popularity" of platforms.

[[APi]TheMan]

Done and done. One question... How important is that setting about allowing BOINC to use no more than 100 Gbytes of space for work blah blah...? I set it to 1 Gbyte and then downloaded the client... attached... and now I get:

Code:
2004-06-22 23:36:30 [SETI@home] Message from server: No work available (there was work but 
you don't have enough disk space allocated)

I tried to detach and reattach after I had changed the settings for my account on the webpage... What is the MacNN team number? The team serach function has been disabled.

edit: detaching and reattaching worked... or maybe it is just random. But I'm crunchin' Seti now.

[Shaktai]

Originally posted by [APi]TheMan:
I tried to detach and reattach after I had changed the settings for my account on the webpage... What is the MacNN team number? The team serach function has been disabled.

edit: detaching and reattaching worked... or maybe it is just random. But I'm crunchin' Seti now. [/B]

I suspect the servers are getting hit pretty heavy, while people switch over. Did you "activate" your old SETI account, or register a new one? Nothing like "start-up" bugs.

[[APi]TheMan]

Originally posted by Shaktai:
I suspect the servers are getting hit pretty heavy, while people switch over. Did you "activate" your old SETI account, or register a new one? Nothing like "start-up" bugs.

Yeah, I "activated" my old account... It wouldn't let me activate my older account (4.5 years old), maybe because I never crunched any units with that one. Hah. I'm definitely crunching away on the units now, though. Well, my computer was crunching when I left for work... 9 hours ago, I haven't been home to check since then. Hopefully I'll have some progress, though it ran all night while I slept lastnight and I didn't see any progress when I woke up.

So I guess we can't join a team yet... Bummer.

edit: I don't feel so bad about the team thing... according to this [setiboinc.ssl.berkeley.edu] page, every registered team has no more than 1 member... cool.

[[APi]TheMan]

Ok so the BOINC command line progress notifications aren't very helpful or descriptive. I return home about 10 hours later and I have this:

2004-06-23 06:10:41 [SETI@home] Starting computation for result 11se03aa.2459.1745.1009650.129_2 using setiathome version 3.08
2004-06-23 06:10:41 [SETI@home] Started upload of 11se03aa.2459.497.229830.19_0_0
2004-06-23 06:10:43 [SETI@home] Finished upload of 11se03aa.2459.497.229830.19_0_0
2004-06-23 06:10:43 [SETI@home] Approximate throughput 284409.460455 bytes/sec
2004-06-23 13:49:01 [SETI@home] Computation for result 11se03aa.2459.1745.1009650.129 finished

I got a few of those... and the BOINC stats page say that I've returned 5 results...? Also, I have two of the same computer listed on the "computers" page for my account. Is that because I accidentally attached twice?

edit: I "merged" the two computers together on the computers page... Now there is only one.

[Shaktai]

Originally posted by [APi]TheMan:

So I guess we can't join a team yet... Bummer.

edit: I don't feel so bad about the team thing... according to this [setiboinc.ssl.berkeley.edu] page, every registered team has no more than 1 member... cool.

Some kind of a problem with team stats right now. If you were part of TeamMacNN under SETI Classic, you will automatically be part of Team MacNN under SETI-BOINC, unless you yourself change it. At least that is the theory...

I'll be back home tomorrow and then maybe can sort through some of the problems. As far as crunching, you may want to do a manual update, other wise it won't register any returned credits immediately.

Stop the client.
restart it using -update_prefs "URL" flag like so.

./setiboinc_3.18 -update_prefs http://setiboinc.ssl.berkeley.edu/sah/

This will force an update of completed work. At that point one of two things will happen. Either the credits earned will show up in your "pending list" or if the work unit has been validated by another computer it will show up as awarded credit. Cheating, which was so rampant under SETI Classic, will be very difficult under BOINC. The side effect however is that it may take a few days to see final credit. The more people get on the project though, the less time, validation should take.

[[APi]TheMan]

Originally posted by Shaktai:
Some kind of a problem with team stats right now. If you were part of TeamMacNN under SETI Classic, you will automatically be part of Team MacNN under SETI-BOINC, unless you yourself change it. At least that is the theory...

That's ok, my points should carry over to the team when we're finally able to join, right?

Originally posted by Shaktai:
I'll be back home tomorrow and then maybe can sort through some of the problems. As far as crunching, you may want to do a manual update, other wise it won't register any returned credits immediately.

Stop the client.
restart it using -update_prefs "URL" flag like so.

./setiboinc_3.18 -update_prefs http://setiboinc.ssl.berkeley.edu/sah/

This will force an update of completed work. At that point one of two things will happen. Either the credits earned will show up in your "pending list" or if the work unit has been validated by another computer it will show up as awarded credit. Cheating, which was so rampant under SETI Classic, will be very difficult under BOINC. The side effect however is that it may take a few days to see final credit. The more people get on the project though, the less time, validation should take.

Yeah, I read that in the forums over there; they're very helpful actually. So are you and I the only two active for team MacNN SETI BOINC? Good to be in it with ya. Previous to this I had been folding for about 6 months, and before that I was running the RC5 over at distributed.net for a few years. I decided that I wanted something constant, I am in it for the long haul so I want to be part of a project that is too, so I think I'm switching to SETI.

[djohnson]

OK downloaded the file, converted my acount and now when I try to run the boinc program, I get the following errors:

dyld: ./boinc_3.18_powerpc-apple-darwin Undefined symbols:
./boinc_3.18_powerpc-apple-darwin undefined reference to _btowc expected to be defined in /usr/lib/libSystem.B.dylib
./boinc_3.18_powerpc-apple-darwin undefined reference to _iswctype expected to be defined in /usr/lib/libSystem.B.dylib
./boinc_3.18_powerpc-apple-darwin undefined reference to _mbsrtowcs expected to be defined in /usr/lib/libSystem.B.dylib
./boinc_3.18_powerpc-apple-darwin undefined reference to _towlower expected to be defined in /usr/lib/libSystem.B.dylib
./boinc_3.18_powerpc-apple-darwin undefined reference to _towupper expected to be defined in /usr/lib/libSystem.B.dylib
./boinc_3.18_powerpc-apple-darwin undefined reference to _wcscoll expected to be defined in /usr/lib/libSystem.B.dylib
./boinc_3.18_powerpc-apple-darwin undefined reference to _wcsftime expected to be defined in /usr/lib/libSystem.B.dylib
./boinc_3.18_powerpc-apple-darwin undefined reference to _wcslen expected to be defined in /usr/lib/libTrace/BPT trap

OS X 10.2.8 Anyone have any clues?

[Shaktai]

Originally posted by djohnson:
OK downloaded the file, converted my acount and now when I try to run the boinc program, I get the following errors:
........

OS X 10.2.8 Anyone have any clues?

Whoops, forgot to mention that. The Mac client has a problem with OS 10.2.x. They are working on it. 10.3.x is currently required.

[djohnson]

Yup, I just figured that out. Here is their Mac forum with many problems like this:

http://setiboinc.ssl.berkeley.edu/sa...orum.php?id=14

Well, I guess I will have to wait until next week when I get my new PowerBook before trying to get SetiLogger X updated for BOINC...

[mikkyo]

Can't someone just rebuild boinc on 10.2.8 and post the binary for folks?

[djohnson]

Attempting now on 10.2.8....

[djohnson]

Unsuccessful using nightly builds and the cvs 3.07 entry. It is missing some files... I am going to try and remove the references to those files and see how it works.

UPDATE: Now it seems to be working... just a little slow. I get to order my PowerBook on Monday!

[mikkyo]

The source code archives look to be a nightmare.
Missing files abound. Scary.

Where is the CVS server???
I can't find where to pull actual current source from.

[Scotttheking]

The project being out of work is really nice...

[bborofka]

My 1GHz PowerBook G4 at Home:

Code:
Measured floating point speed: 458.25 million ops/sec
Measured integer speed: 1353.85 million ops/sec

My 1.6 GHz P4 Gateway at Work:

Code:
Measured floating point speed: 989.18 million ops/sec
Measured integer speed: 2063.65 million ops/sec

Why the huge disparity? The Gateway has 600 more MHz obviously, but nearly doubles my PB in both categories. Is it because of poorly optimized software on the Mac or just the slower architecture of the Mac?

[Shaktai]

Originally posted by eVo:
Why the huge disparity? The Gateway has 600 more MHz obviously, but nearly doubles my PB in both categories. Is it because of poorly optimized software on the Mac or just the slower architecture of the Mac?

It appears to be heavily affected by "bus" speed. G5's do very well.

Keep in mind all, that in many ways, though they have pushed it to be "public", in reality it is just a public beta.

They have added an additional splitter and will add more, to meet the demand. However, to do so they will have to pull resources from SETI classic most likely. The current situation is not ideal, but it will get better. I lost several days of production myself, because they made the switch-over while I was on a 5 day business trip, that got extended to 10 days. However in a few short hours, I have already aquired 169.73 "pending" credits since getting my clients all updated.

[bborofka]

Double post

[bborofka]

Originally posted by Shaktai:
It appears to be heavily affected by "bus" speed. G5's do very well.

Keep in mind all, that in many ways, though they have pushed it to be "public", in reality it is just a public beta.

They have added an additional splitter and will add more, to meet the demand. However, to do so they will have to pull resources from SETI classic most likely. The current situation is not ideal, but it will get better. I lost several days of production myself, because they made the switch-over while I was on a 5 day business trip, that got extended to 10 days. However in a few short hours, I have already aquired 169.73 "pending" credits since getting my clients all updated.

So it's not a matter of the SETI software not being optimized for the Mac as much as PCs? Sorry, I'm just new to SETI, trying to understand it more.

According to this thread, AMD Athlon 64 3400+ crushes a Dual 2 GHz G5. Is the because of a superior architecture of the AMD?

[Shaktai]

Originally posted by bborofka:
According to this thread, AMD Athlon 64 3400+ crushes a Dual 2 GHz G5. Is the because of a superior architecture of the AMD?

Actually, if you read closer, you will see that in one of the messages, it was acknowledged that the Athlon 64 originally discussed was running a specially compiled version of the BOINC client, that was free of the debug code. It is easy to bandy about "benchmark" numbers, but that doesn't necessiarily relate to real world performance. The G5's do very well. The downloadable clients are not "optimized" for any of the architectures, but in Beta Testing it was discovered that specially re-compiled versions of BOINC for the Athlon 64's could turn in some very impressive benchmarks numbers.

Also, what is not known is the conditions under which the bencmarks are being ran. The client Benchmark is nice for people who like to play with numbers, but the reality is that the actual "application" be it SETI, Predictor, Climate Prediction, or whatever will vary depending on many factors. Also keep in mind that while the Athlon 64 is cruncing one unit, the dual G5 will be crunching two. In Beta, the G5's consistently did very well with the standard compiles. It is too early to tell in the live production, what actual production will be for any of the machines.

There are numbers, and then there are damned numbers. The thread you referenced leans more towards the latter. Too many variables to have real meaning. Running "standard compiled" Boinc clients, the G5 does not suffer compared to any others where it counts.

[bborofka]

Originally posted by Shaktai:
Actually, if you read closer, you will see that in one of the messages, it was acknowledged that the Athlon 64 originally discussed was running a specially compiled version of the BOINC client, that was free of the debug code. It is easy to bandy about "benchmark" numbers, but that doesn't necessiarily relate to real world performance. The G5's do very well. The downloadable clients are not "optimized" for any of the architectures, but in Beta Testing it was discovered that specially re-compiled versions of BOINC for the Athlon 64's could turn in some very impressive benchmarks numbers.

Also, what is not known is the conditions under which the bencmarks are being ran. The client Benchmark is nice for people who like to play with numbers, but the reality is that the actual "application" be it SETI, Predictor, Climate Prediction, or whatever will vary depending on many factors. Also keep in mind that while the Athlon 64 is cruncing one unit, the dual G5 will be crunching two. In Beta, the G5's consistently did very well with the standard compiles. It is too early to tell in the live production, what actual production will be for any of the machines.

There are numbers, and then there are damned numbers. The thread you referenced leans more towards the latter. Too many variables to have real meaning. Running "standard compiled" Boinc clients, the G5 does not suffer compared to any others where it counts.

Thanks for the explanation. So it is just a matter of the clients being optimized for a platform. I just wasn't sure.

The other question I have is, you mention the project URL of http://setiboinc.ssl.berkeley.edu/sah/ . The email I got from SETI (as well as their front news page) told me to use http://setiathome.berkeley.edu/. What is the difference and why should I use http://setiboinc.ssl.berkeley.edu/sah/?

[Scotttheking]

It did a WU, uploaded it, but I show no credit or pending credit. How long should it take to show in the stats? And why am I awake?

[Shaktai]

Originally posted by bborofka:
The other question I have is, you mention the project URL of http://setiboinc.ssl.berkeley.edu/sah/ . The email I got from SETI (as well as their front news page) told me to use http://setiathome.berkeley.edu/. What is the difference and why should I use http://setiboinc.ssl.berkeley.edu/sah/?

The orginal URL was http://setiboinc.ssl.berkeley.edu/sah/ They have updated their system to now point to http://setiathome.berkeley.edu/ Use the new URL when attaching a client or requesting updates. Both work, but eventually only the new one, http://setiathome.berkeley.edu/, will work. This URL actually points to SETI classic, but eventually it will be the only URL, when Seti Classice "someday" shuts down.

[Shaktai]

Originally posted by Scotttheking:
It did a WU, uploaded it, but I show no credit or pending credit. How long should it take to show in the stats? And why am I awake?

As part of the "anti cheating" process and to improve the Science, all work units must be validated by two other computers before credit is actually awarded. The credit actually given will the the "middle requested score" of the three. If you look under "pending credits" in your profile, you should see your completed and returned work unit. If it is not there, then run the flag
-update_prefs http://setiathome.berkeley.edu/

That should force an update of your stats and preferences.

NOTE to ALL: There have been problems getting work. This is due to the demand placed on the system for work. Resources are being diverted to help with this problem by the SETI team, but for the interim, please request no more then 1 days work in your preferences, unless you have a very real need to cache more then that.

Also, SETI now has a two week expiration on work units. If you don't complete a work unit within two weeks of when it was issued to your computer, then it will be reissued to someone else, and you will not get credit.

Now why the heck am I awake at this hour??

[djohnson]

Oops, forgot to hit submit last night... Anyways, as reported, the code will not compile. I don't see how they can claim that these are nightly builds when they are missing about 10 files! Well, I will have SetiLogger X updated once I get my PowerBook next week... Maybe it will be finally released in another week or so?

[Shaktai]

Originally posted by djohnson:
Oops, forgot to hit submit last night... Anyways, as reported, the code will not compile. I don't see how they can claim that these are nightly builds when they are missing about 10 files! Well, I will have SetiLogger X updated once I get my PowerBook next week... Maybe it will be finally released in another week or so?

Well you will have one of the first new utilities for SETI-Boinc then, and it will likely be quite popular with Mac users. I look forward to seeing it when ready.

[Scotttheking]

2004-06-25 17:35:22 [SETI@home] Message from server: Project is temporarily shut down for maintenance

[Shaktai]

June 25, 2004 (FROM SETI@HOME)
We added a second splitter and then hit the next bottleneck which was the transitioner. We are building a new transitioner now with a more efficient workunit database query. In addition we are bringing a third splitter online for good measure. The project will be down for a bit while we put this in place.

The website is back up, but they are still working on the back end. Getting the new splitters online is a priority so that they can supply sufficient work units to participants. Now that I am back home I have gotten all my boxes working again, and have received 197 units (cobblestones) in granted credit and have almost another 1000 pending in the past 24 hours. Team MacNN has been granted 482 units total. SETI@The Anderson House is currently the team MacNN leader with 208 cobblestones.

Team MacAddict has gotten a jump on us at #52.
Our current team rank is #114 (and climbing) out of all teams, most of those ahead of us got a 24-48 hour head start.
Mac OS Rumors is #134

[[APi]TheMan]

I wish I could join a team... Damnit. I have a few boxes running SETI and I'm pullin' lots of units down but they're not counting for anyone... "Team lookup disabled"

[djohnson]

Originally posted by Shaktai:
Well you will have one of the first new utilities for SETI-Boinc then, and it will likely be quite popular with Mac users. I look forward to seeing it when ready.

Here is a new idea... I need someone to send me the files outputted my the BOINC client. I can use those to update the current program. I must have a program finished by Sunday night for WWDC, but I can get this done pretty quick with those files. Send them to djohnson(at)tfworld.com. Thanks!!!!!!!! Oh and first one that sends them will possible get credit on both the home page and the program itself. Well, everyone that sends the files will get front page credit at http://www.tfworld.com/seti

[djohnson]

Anyone please?

[Shaktai]

Originally posted by djohnson:
Anyone please?

You should have mail. Didn't know what you needed, so sent everything except the account specific info.

[mikkyo]

I found the official cvs server.
Maybe now we can get working source and build something for 10.2.8 folks.
cvs -d :pserver:anonymous@alien.ssl.berkeley.edu:/home/cvs/cvsroot login
do a login with no password and then a co boinc.
You'll need mysql 3.23 or higher installed as well as gnu dev stuff.
Specifics here http://boinc.berkeley.edu/software.php

[djohnson]

Excellent work mikkyo! See if you can find all of the files... There are slightly more BOINC files then there were with the seti@home program. I should have something working in a few days. Oh and thanks to Shaktai for assisting with the files!

[mikkyo]

So ignoring the projectbuilder file, you might get a build going by doing:
cd boinc
setenv MYSQL_CONFIG nope
./configure
cd client
make

If the build works, I'll see if I can find a 10.2.8 system to build on.

[djohnson]

I have one right here if it works. Send it to me and I will test it out.

[mikkyo]

I got a 10.2.8 version built and it seems to work.
Anyone want to host it?

[djohnson]

I might be able to host it... assuming it isn't too bad. Let me think about it and I will get back with you.

[Scotttheking]

Originally posted by mikkyo:
I got a 10.2.8 version built and it seems to work.
Anyone want to host it?

email it to team AT macnn DOT com and I'll drop it on my server.

[Scotttheking]

How do they calculate credit? Their stats pages, especially the teams list, as well as members on a team, is horrible.
20 per page is pointless.

All but one of the WU that I've done still show as pending

[Scotttheking]

As sent by Mikkyo:
http://teamstats.macnn.com/~distcomp...9_ppc10.2.8.gz

[Shaktai]

Originally posted by Scotttheking:
How do they calculate credit? Their stats pages, especially the teams list, as well as members on a team, is horrible.
20 per page is pointless.

All but one of the WU that I've done still show as pending

I agree, that their team stats pages is a mess. Moving over the teams from SETI Classic sounded good, but it means that the stats files are huge, and unless hamm reappears from somewhere, nobody can clean up the list to get rid of all those who never crunched. Folks have been clammoring for them to at least make the team lists sort by work done.

Work units once crunched must be validated by at least two other computers. Once validated, then the mean pending credit is granted. This serves two purposes, 1 is to improve the quality of the science results (faulty work units are automatically ruled out) and 2 is to prevent cheating as it was over in SETI Classic. It can delay the time to get back work depending on when the work unit gets validated. In theory, running at full speed, there should not be significant delays most of the time, but if someone downloads work and doesn't complete it for any reason, then it will take two weeks before it is resent out to someone else. If that is a work unit you are waiting on for validation, then it could take more then two weeks for you to receive credit.

After about 2-3 weeks of crunching though, the credit flow evens out, and you will see credit coming in more consistently.

Predictor@Home is similar but requires only one other computer to validate. However it also uses homogenous (sp?) validation because of the need for extreme precision. As a result a unit crunched by an AMD machine can only be validated by another AMD, Pentium by a Pentium and Mac by a Mac. This has something to do with the "floating point" issues between different chips and OS's Also, windows must be validated by windows and linux by linux.

If you are running a Mac for instance, and get a specific work unit, that unit should then only be sent out to other Macs for validation. They did have "glitch" the other day though that caused a mix-up. At least they admit that they are in Alpha/beta stage and don't claim a product ready for prime time.

[Lateralus]

I'm having some trouble getting BOINC going.

The steps I've been through:

1) Download it.
2) Place the BOINC file in a folder.
3) Open Terminal - 'cd /Users/ed/Applications/Distributed\ Computing/SETI@Home\ -\ BOINC' - Return
4) '/Users/ed/Applications/Distributed\ Computing/SETI@Home\ -\ BOINC/boinc_3.18_powerpc-apple-darwin' - Return
5) For the URL, I put: http://setiathome.berkeley.edu
6) I put in my user ID, as was sent to me via e-mail.
7) And then I get this:

2004-07-02 11:57:54 [SETI@home] Project prefs: using your defaults
2004-07-02 11:57:54 [SETI@home] Host ID is 41504
2004-07-02 11:57:54 [---] General prefs: from unknown project http://setiweb.ssl.berkeley.edu/ (last modified 2004-07-01 22:41:48)
2004-07-02 11:57:54 [---] General prefs: using your defaults
2004-07-02 11:57:54 [---] Fewer active results than CPUs; requesting more work
2004-07-02 11:57:54 [SETI@home] Requesting 172408 seconds of work
2004-07-02 11:57:54 [SETI@home] Sending request to scheduler: http://setiboinc.ssl.berkeley.edu/sah_cgi/cgi
2004-07-02 11:57:54 [SETI@home] Scheduler RPC to http://setiboinc.ssl.berkeley.edu/sah_cgi/cgi failed
2004-07-02 11:57:54 [SETI@home] Scheduler RPC to http://setiboinc.ssl.berkeley.edu/sah_cgi/cgi failed
2004-07-02 11:57:54 [SETI@home] No schedulers responded
2004-07-02 11:57:54 [SETI@home] No schedulers responded
2004-07-02 11:57:54 [SETI@home] Deferring communication with project for 34 minutes and 53 seconds
2004-07-02 11:57:54 [SETI@home] Deferring communication with project for 34 minutes and 53 seconds

????????

[mikkyo]

Just let it run and see if something shows up after the delay expires.

[Lateralus]

I ran it for a few hours last night. The delays continued, changing in length.

[Shaktai]

Originally posted by PowerMacMan:
I ran it for a few hours last night. The delays continued, changing in length.

The SETI servers have been down a lot. They are having to redesign some of the backend servers to handle the demand. What you are seeing is an indication the servers are down. It has assigned a host ID for your computer, so you are good to go as soon as they get their servers back up.

From Seti-Boinc homepage:
July 2, 2004
We've been having problems with server performance. To address these problems we're making some changes to our database. The project will be shut down most of today while this is being done.

However, Predictor@home which is also BOINC has been mostly running well, and it is advertised as a public alpha. You can attach to that project at a low priority and crunch there until Seti gets back up if you like. Once seti has work available it will shift over. (In preferences on the websites set something like resource share 10 for Predictor and 100 for SETI, or whatever works for you.

Predictor has even posted their first results. Not bad for an alpha that runs better then a lot of betas.

[Lateralus]

Are the SETI servers back up? I am still getting the same message.

[Shaktai]

Originally posted by PowerMacMan:
Are the SETI servers back up? I am still getting the same message.

They took the entire site down earlier today to do some work on the backend. There was a notice on the website earlier, but right now, even the website is down. The note, when readable, simply said the project would be down for "several hours".

It was up last night. Work availability continues to be off and on. What some SETI folks have done is attach to the Predictor project as well, and set Predictor with a low resource share like 10, in the web preferences. This leaves SETI with a 10 to 1 resource preference advantage, but if SETI isn't giving out work, but predictor is, your CPU will stay busy with Predictor work units, but will still periodically try to get SETI units. Keep the cache set small to less then one day if possible, for both projects. I recommend .5 to 1 day at first, then you can get a feel for how your particular equipment works.

The reality is, that Public SETI-BOINC, is really just a continuation of the Beta project, though they don't admit to it. They still have a lot of backend work to get done. SETI Classic remains active and will for a while still.