OT: Space Empires and Seti@Home

[Baron Munchausen]

No, you just need to learn the secrets of SETI@Home fanaticism.

First of all, what Version are you using? The Win95/98 screen saver? Set it to blank the screen after 0 minutes. No, not your screen saver timeout in the Windows settings. The blank screen setting in the actual SETI program. The pretty graphics use as much CPU as the actual calculations. By turning them off you'll save a lot of power. You'll probably more than double the speed at which you finish a work unit just by doing that. A 1 Ghz P III is actually fairly powerful. You could check the box to have it running all the time in the background, I think. I have an 800 Mhz Duron and it's no problem for me to have it running all the time.

Now after that you have to consider how you have configured your computer and what else is running on it. Sounds are cute, but use MUCH CPU time playing unless you've got a fancy sound subsystem with hardware buffering. It would be wise to turn sounds off. Lots of 'harmless' programs that load up and sit in memory waiting for you to do something are actually using a noticable portion of your CPU power. What have you got in your task bar? Think real hard about whether you need those things. Is it really that hard to just go to the menu or desktop icon to activate that program? Remove as many as possible.

There are some more obscure tweaks that will give you a tiny percentage more, like using lower color depth on your screen. But the best improvement will come from not using the screen saver Version at all. Get the Win NT Version, which DOES run in Win95/98, and use that instead. (Then color depth doesn't matter!) It's about 10 percent faster than the Screensaver even with the graphics turned off. I think you will find that everyone on the SE IV group has been using the NT Version for a long time now.

I suppose I should rejoin the group but I hate the competitiveness of it. It's supposed to be 'fun' and/or 'charity' in that you're donating spare computer power to a good non-profit cause, but it's become a huge competitve thing for so many people now. There are entire farms of computers in some people's basements dedicated to SETI and nothing else. Some 'gift of spare CPU time'... But at least they've got the largest accumulated computing project in history now. Just wait till SETI@HOME II starts up for the southern hemisphere...

[ September 29, 2002, 01:50: Message edited by: Baron Munchausen ]

[Thermodyne]

Wow, I was wondering where all of the new people were coming from, so now I know. I don�t visit here very often. For the guys looking for faster unit times, run the 3.03 command line Version and a caching program like SetiDriver or one of the others. Then OC the heck out of every old system that you have lying around.

The team was started by one of the SE III masters, Paul Dionne/Minervan Necores Alliance. And while it is a small team, we have a pair of �%�rs, and should have 2 people over 10,000 units by Thanksgiving. So come on over and join up, we need all the help we can get

http://setiathome.ssl.berkeley.edu/s...eam_73545.html

[Fyron]

Wait... that low average CPU time per work unit is a good thing? I wasn't sure what it meant. Are all of the work units the same size?

[Suicide Junkie]

The units are all the same size, but the amount of work done on them has increased in the Last year or two.

A while ago, IIRC, the SETI guys found that people were processing the units TOO fast, and they were going to run out, so they increased the detail of the analysis.

[Thermodyne]

Originally posted by Imperator Fyron:
Wait... that low average CPU time per work unit is a good thing? I wasn't sure what it meant. Are all of the work units the same size?

No two units are the same. The first Version had work units that were faster to work than the Version we have now, something to do with adding Gaussians I think. So the peeps that got in at the start had some quick units to work on, but with slower systems. The units we work with now vary by as much as 25% depending on the angle of the work unit. I won�t go into all of that here; it has been covered in detail on the seti forum. The key to high output is to have a bunch of systems doing the work, but a few fast systems can put out a lot of units. My slowest system is a P II 300 at about 14:30 a unit. And on the other end of the scale I have a P4 @ 2.8Ghz that does a unit every two hours.

[Thermodyne]

Originally posted by Suicide Junkie:

A while ago, IIRC, the SETI guys found that people were processing the units TOO fast, and they were going to run out, so they increased the detail of the analysis.

Not true, each unit is worked many times by many different systems, just to make sure that the results are reliable. This also eliminates the need to track lost units. Many of the units that are downloaded never get returned. The biggest problem we have had is when we used up the bandwidth at the UC Berkeley campus. It got real hard to send and receive work units for a few months. But all things must end, and after the first of the year the data from the northern hemisphere will be finished. There are a couple of new projects in the works, but it will remain to be seen if they will be as popular as seti@home is. They should hit 4,000,000 Users this week.

[ September 30, 2002, 17:23: Message edited by: Thermodyne ]

[sven]

Originally posted by Thermodyne:
My slowest system is a P II 300 at about 14:30 a unit.

Hehe, my slowest is a 486/66 which takes about 480 hours (20 days!) a unit.

[sachmo]

Good news, alien hunters!

The world's most popular ET-hunting program for home Users is about to get upgrades of both its software and the telescope that feeds data into it.

For three years, SETI@home has used the spare processing power of computers across the world, in the guise of a screensaver, to examine radio telescope data for signs of intelligent extraterrestrial life. The data comes from the Arecibo Observatory in Puerto Rico, which can scan only a 30-degree patch of the sky in the Northern Hemisphere.

Starting early next year, the Arecibo recorder will be shut off and a new but similar recorder will be turned on at the Parkes Observatory in Australia, which can observe 70 degrees of the sky -- and a more advantageous part of the sky, too.



The odds go up for finding ET from the Southern Hemisphere. Why? Because most of our Milky Way Galaxy's stars, in the dense central bulge, are visible only from south of the equator. While northern skywatchers see our galaxy's main disk -- itself rich in stars -- the denser bulge barely rises above the horizon.

And the chances of getting a signal from outside of the Milky Way "is pretty small," says David Anderson, the SETI@home project director at the University of California at Berkeley.

New software

At around the same time as the Parkes Observatory recorder comes Online, SETI@home will also release new software designed to broaden its scientific applications and streamline the program.

The first change expected to come Online is AstroPulse, a program that will scan the three years of Arecibo data stored on tapes for broad band signals, or "a sudden burst of energy that's spread across a wide frequency range," Anderson explained in a recent interview with SPACE.com. The current SETI@home algorithms search for narrow-band signals.

One of the most exciting explanations for such an observation, Anderson says, is the evaporation of quantum black holes, a phenomenon predicted by Stephen Hawking but that has yet to be observed.

According to Hawking's theory, "black holes give off radiation and therefore lose mass," Anderson explains. "So small black holes will basically kind of dry up and go away. In the moment of their disappearance, the theory predicts that they will give up a short burst of broad band radio radiation. Our data from Arecibo is an ideal place to look for that sort of thing."

Going BOINCers

With the release of AstroPulse will come the inauguration of BOINC, or Berkeley Open Infrastructure for Networking Computing.

BOINC, in essence, is "a new layer of software" that separates the different components of the SETI@home program. It will allow changes to be installed without interrupting the screensaver or asking the user to download upgrades.

About 4 million people have downloaded SETI@home, and roughly 600,000 use it on a regular basis. In the past, when new Versions of SETI@home were released and were finally installed by the Users, "it's taken between six months and a year," says Anderson. "BOINC will streamline that."

BOINC will also make it possible to seamlessly integrate different kinds of computing projects into SETI@home, such as programs that analyze biological functions, global warming, or anything else researchers can dream up.

"We're opening things up to the world," Anderson said. "It turns out there are many areas of science�that can also benefit from huge amounts of computer time."

A project called Folding@home, which simulates protein folding and its related diseases, already exists.

Democratizing science

Users will be able to integrate projects such as Folding@home into their desktop Versions of SETI@home by clicking a box, Anderson says. This will also allow Users to divide their processing power between projects. "They won't be forced to choose between one project and another � essentially bringing a form of democracy to science research," Anderson says.

SETI@home was founded in 1995 by David Getty, a former computer science graduate student of Anderson's at Berkeley. The original idea was always to tap into the unused power of computers everywhere for scientific data analysis. But, "to get SETI@home off the ground at all, we really had to limit our ambitions and just try to do things in as simple a way as possible," Anderson says.

With BOINC and the increasing interest of scientists everywhere in this computing power, SETI@home is expanding.

At its core, still, is the search for extraterrestrial intelligence. And switching to the Parkes Observatory will make work easier for the 5-member SETI@home staff. One of the main tasks of the staff (besides writing software such as BOINC) is to examine any possible extraterrestrial signals detected by the SETI@home program.

Most, if not all of the signals, are a result of television, radio and other signals leaking into the telescope from the one civilization we know to exist.

To solve this problem, SETI@home will take advantage of Parkes' multi-beam receiver. It is one of the first observatories in the world to install one. Instead of examining only one point in the sky at a time, as Arecibo does, Parkes will record data from several.

"If you get a signal that seems to be coming from several points in the sky at one time," Anderson explains, "then it's probably not coming from the sky at all. It's coming from Earth and bouncing off the atmosphere and back into the telescope."

Anderson expects SETI@home Users to begin analyzing Parkes data soon after AstroPulse is released in February or March of next year. But, he adds, "every scheduling prediction that I've made has been way off."

[zen.]

Suh-weet! Thanks for the article, Sachmo...I haven't been up to date at all with SETI news.

Which reminds me, I need to download a Win2k SETI@Home Version so I can reinstall after getting my new hard drive.

zen

[Thermodyne]

Seti hit 4 MILLION Users today! That�s 4,000,000 for you guys across the pond. I hit 8000 units at almost the same time, should only take about 400k more to get to be #1