Interview with SETI@home Director David Anderson

CowboyRobot writes "ACM's Queue magazine interviews David P. Anderson, a research scientist at the U.C. Berkeley Space Sciences Laboratory, who directs the SETI@home and BOINC (Berkeley Open Infrastructure for Network Computing) projects. SETI@home uses hundreds of thousands of home computers in the search for extraterrestrial intelligence. FTA: "volunteer computing arose because projects such as SETI@home needed $100 million worth of computing power but didn't have the money. But there's no free lunch--a project must give participants something in return for their computer time.""

Comments

Give them a way to keep score

[winkydink]

It seems that many of us are competitive enough to donate cpu time and only get back a scorecard.

Re:Give them a way to keep score

[DurendalMac]

The way a lot of the SETI competitors see it, they get a bigger e-pecker in return for their number crunching efforts.

Re:Give them a way to keep score

[HTTP Error 403 403.9]

It seems that many of us are competitive enough to donate cpu time and only get back a scorecard.

How about using the MMORPG method of rewarding participants. Have SETI members level up after certain of work units. "I'm a level 42 SETI warrior!" Or maybe have SETI members find "rare" artifacts. "I have the sword of Cocconi!"

I can't understand how my nephew will play WOW for an entire weekend to change a number from 47 to 48.

How about a free probing?

[Datagod]

When you find the aliens, perhaps give all the Seti@home volunteers a good probing?

Power usage?

[Rikkochet]

Now correct me if I'm wrong, but doesn't utilizing 100% of a CPU result in a significant increase in power consumption on the system versus the processor simply being idle? Sure, it's nothing compared to leaving your big CRT monitor on, but still.. I definitely notice my CPU and case temperatures are substantially higher when I have high CPU utilization going on - I can't help but wonder how much energy we're actually consuming here.

Re:Power usage?

[Duncan3]

1. Yes.

2. Lots.

The cost is just spread out over thousands of people, instead of having them all in one place.

Re:Power usage?

[Mr Guy]

I have a 45 watt CPU. I'm going to assume for simplicity's sake that all other power drain is roughly equal and they only burning CPU time. We'll say, for the ease of the numbers, it burns 4 watts idle, so the ramp up to full cpu is 41 watts. That gives me 1 Kilowatt-hour per day. I pay about 8 cents a kilowatt-hour. So the way I figure it, for my two computers, I'm donating about $2.40 a month to cancer research with folding@home.

Re:Power usage?

[squidfood]

What about the cost to the environment?

From this link a good average differential between a processor at load and idle is 40W. If you turn the computer off instead, that's maybe 80W. (Broad average over many computers).

Now Here we see that 2million years of computing time has been used, so (times 40W/hr) that comes to 700,000MWHr.

No the 2000 U.S. consumption of energy was ~21 billion MWHr. (Here, and trust the government to use quadrillions of BTUs as a unit). So to date, SETI has used 0.003% of U.S. annual energy consumption. And that's almost enough energy to power the City of Red Deer, Alberta for 17 months! Someone else can tell us how many libraries of congress you could have read with that much light.

Feel free to check my units and zeros, I've been wrong before, as long as someone can tell the Brits what a quadrillion is.

Re:New client

[VJ42]

I carried on using the old client, it still works. (I couldn't get BOINC to work either).

How Timely

[ipxodi]

How timely considering Seti@home has been offline for a week and all the users have this really keen "Boinc is currently idle" floating screensaver.
Maybe they've been hacked by Aliens who didn't want to be discovered.
"I for one welcome our new alien hacker overlords."

.

Well, that's sort of the point.

[Mr Guy]

Calling it a resource hog may not be the right term depending on what resources you are talking about. The whole point of the programs is to run your CPU to max when it would be otherwise idle. In that sense you are deliberately contributing to the wear and tear of your system, as well as any heating issues you may be concerned about. You are choosing to offset this against the value of the research, which is why I can't understand why people will donate cycles to SETI and not to something more directly useful like folding@home, but that's a value judgement.

It, however, should NOT be a resource hog in the sense of Microsoft Office, in that it slows down other programs. These programs are designed to utilize any resources you aren't using, and immediately give them back if you need to use them. This is done by setting the priority of the process just over system idle. Any cycles that would be spent idle are spent on processing instead, but when a program wants cycles, it gives them up.

Re:New client

[lgw]

SETI is actually looking on the frequency band that makes sense, given the nearby stars have already been searched more broadly. Remember, SETI is only listening for aliens who are trying to send us a message, not looking for radio traffic leaking from some alien planet.

Anyone doing radio astonomy is going to be listening on or near the 21cm "hydrogen band", as there's only "a very narrow frequency band" that works for radio astronomy at any distance. If you're going to send a signal to someone you know noting about, this is the one frequency range that you can be sure they'll be listening on, if they're listening at all. It's not just chosen arbitrarily.

Certainly, the chance of finding alien intelligence after we checked the easy targets is small - small enough that I'm happy SETI is orivately funded, not fighting for funds from the NSF. But for a volunteer effort, support what makes you happy to support.

Re:I wish BOINC could...

[SETIGuy]

I wish BOINC could also be designed to use graphics cards - ala the BrookGPU project - to help with the number crunching duties.

So do I. In fact I keep looking for people to help us develop this.... To no avail. :( Aparently the people who want this most don't have the ability to implement it, and the people who have the ability (assuming they exist) aren't interested.

If anyone wants to help, join the boinc_opt mailing list and send a message.

BTW, David is the titular director of SETI@home, but currently has no managerial duties beyond the BOINC project.