Slashdot Mirror
SETI@home & RC5
abh writes "The SETI@home project is now sending new (fresh) work units, after having spent a few weeks in a rut sending the same 2 days worth of observations repeatedly. Read the announcement " As well, we were sent word from a reader that we've lost the #1 position at RC5. Head over and sign up!
He doesnt "get off" doing this. He has a large group of machines without easy internet access, so he uses the good 'ol floppy disk method of transferring completed blocks... I wouldnt worry about it, Anandtech will be consistently beating Slashdot soon enough, even without dumps, legit as they may be.
Joined rc5, joined slashdot last week!
.. go team!
My machine is running everyday except
Weekends. Before that I tried SETI@home
but I never got it right (my jobs did
not get registered at their site) so
I switched to rc5des.
Not that I know exactly what this is
about (nor do I care) but I do have
extra CPU cycles here (PII 450 Mhz
I think), running linux RH 6.0.
My way of saying thanks for the
everyday news I'm getting
JondZ
The FAQ does tell you sort of how to do it.. but when I search for my email address (JUST signed up..) it doesn't bring anything up. I'm assuming thats because I haven't actually cracked any blocks yet :) I think when you crack a block it will enter you into the database, THEN you can search for your email addy. (I think)
Actually, if you take the 2.4 million units received, and assume that productivity has been constant over the last 30 days since the launch (not a great assumption, but good enough for a very rough approximation), that's about 80K work units per day, so they're still oversubscribed by a factor of 5+, but it's not as bad as it has been.
If anybody keeps track of the completed figure on a day-by-day basis, we could come up with a better estimate.
I had a question for you programmers out there. I have an Indigo2 Extreme at home. The theoretical limit of the Geometry engines on the Video hardware is 256 MFLOPS (as opposed to the meager 18 or so on the R4400 I have as a CPU). How hard would it be to write the RC5 client to crack keys using this instead of the main CPU. Also, WOULD this be better than using the main CPU (general purpose CPU vs. dedicated task CPU). Just a thought.
Dan
If I read things right as they were happening (I don't have time to follow everything!), distributed.net was *originally* headed in this direction. However, they have restructured their goals to concentrate on their current projects, and the founder (beberg) has set off on his own to pursue the larger goal. It was called 'cosm' if I remember right. There's a web site somewhere, and I'm interested to know how things are going for them. Somebody slap a URL up here if they know it.
I'd like to join in, but there are no clients to download.
Me too -- I followed the clients link and found that none of them can be downloaded: 404 for all of the Linux clients. Logging into ftp.distributed.net is no different: all directories are empty and trying to "get" the files (in case they were hidden) did not work. They were not found.
does anyone know what the table full of numbers on the bottomh ome_1.html ;)
of http://seti athome.ssl.berkeley.edu/about_seti/about_seti_at_
is for? at least they don't seem to correspond to ascii, or do they?
we might need a new distributed computing client to solve this
Team Evangelista has been #1 for a looong time /. a whole while to catch up...
now. It'll take
Hmm? He has been dumping over 100000 blocks for a few days straight now, could it be he isn't saving them up??
Yep, he has temporary access to a network of 200
PII-machines. This will not last forever though.
Posted by dangerangel:
My question for the nerds out there isn't about number crunching... is SETI going to really tell us if we find something? If my computer finds a signal among a trillion radio waves I want it to light up like a pinball machine and spit quarters out of the floppy drive.
Here's a memo I sent to SETI. I can't get the Linux client to run. Anyone have an idea?
----
Here's your account info:
Email address: [xxx]
OK to show email address? no
Name: grinder
OK to show name? yes
Country: France
Postal code: 75003
Computer location: Work
Everything correct? (y/n) y
gethostbyname: Success
Server host unknown
This is the version information:
Platform: i386-pc-linux-gnulibc1-static
Version: 1.2
I have tried all three Linux ports with no luck.
i386-pc-linux-gnu-gnulibc2.1 (dumps core)
i386-pc-linux-gnulibc1 (gethostbyname fails)
i386-pc-linux-gnulibc1-static (see above output)
Curiously enough, nslookup knows about setiathome.ssl.berkeley.edu:
Non-authoritative answer:
Name: milkyway.SSL.Berkeley.EDU
Address: 128.32.18.165
Aliases: setiathome.ssl.berkeley.edu
I am lost. Please give me a clue.
David
----
Here ya go:
http://cosm.mithral.com/
there's a scary scenario...
Large print giveth, and the small print taketh away
I saw a discussion a few weeks back of an open distributed client, that researchers needing some computing power could just send out plugin for it, and anybody who wanted to contribute could just plug it in and go.
Sounds like a very cool idea - SETI@Home has already used over 11,000 years of CPU time that would have otherwise gone to waste. Imagine what could happen if any serious researcher who needed more power could plug into that?
The next Cmdr Taco duplicate will be ready soon, but subscribers can beat the rush and see it early!
There's the new Retire-to feature recently implemented. I think you need to email one of the distributed.net people to get blocks from an old email address retired to a current email. All the gory details are somewhere in the distributed net pages. Have a look in the plans page they have.
"For I am a Bear of Very Little Brain, and Long Words Bother Me"
just bookmark your personal stats page. it'll take you there every time. I use it for my home page now.
"For I am a Bear of Very Little Brain, and Long Words Bother Me"
http://setiathome.ssl.berkeley.edu/tech_news.html is the actual announcement, if you follow the link to SETI@Home that for some reason Hemos posted as the actual link to the announcement you can then click the "Technical News Reports" link that is right there on the first part of the page just a few lines below the line you read. It's not that hard to find. :) They posted on that page on June 4th when they first discovered they were sending the same 115 units (an OS bug, they claim) and fixed it so that they were all of the units they had at that time, 31415 total. Yesterday they posted an announcement stating that the work unit pipeline is fully operational, so we should see even more units being sent.
The link should be http://setiathome.ssl.berkeley.edu/tech_news.html which takes you to the announcements. The link Hemos posted is to the main page of the SETI@Home site which just carries a 1 line explanation.
If you read the actual announcements, you'll see it was only one day when the same 2 days worth of work units (actually not even two complete days, just 115 work units from 2 different days) were being sent, then the project moved to about 30k units, which were all that were avaible at the time. Now several months worth of data is avaible. They go into detail as to problems they have had and their solutions on this page, it's rather interesting...
A. One of those people that store up a huge number of blocks for a month or so, then send them all in, and bask in the glory of seeing themselves as #1 in the daily rankings.
This seems the most likely by far. He just had better luck than I did the day a couple months back when I checked in a bit over 100K blocks, but:
a) the last 8K missed the stat run so I only showed 92K (Doh!), and
b) some other guy in Japan had to pick the same day to check in his 160K, so I was #2 (Arrghh!)
Is there anything wrong with that? It's harmless and it makes this more competitive and hence more fun and leads to higher key rates. Sure, it's at least a little bit silly, maybe even immature, but anyone who says that the resultant "spiking" of the stat distribution does some kind of real harm is just showing that he takes this even more seriously than I, when I'm already acknowledging that I'm taking it way too seriously and need a life. The only real harm I can think of is the possibility that hoarding the blocks could cause duplicate work, but that is so remote as to be insignificant.
As for SlashDot losing the #1 spot, AnandTech checked in half a million blocks yesterday, but they've only done 38 million total, to our 140 million in the same 600 days, so unless they've got a new key-cracking supercomputer or otherwise massively increased their horsepower, it must also have been a matter of saving them up. I think it's kinda cool that they managed to do it, even though it knocked us down a rung, which, of course, means war. Like I said, competition is fun.
Hey, wait, it looks like they were also #2 the previous day. Maybe they have got the horsepower to keep this up. Do something! Let's see, how many more clients can I install here at work?
David Gould
David Gould
main(i){putchar(340056100>>(i-1)*5&31|!!(i<6)<< 6)&&main(++i);}
The stats database updates daily from ~0:00-4:00 GMT. Until some blocks that you've checked make it to the stats server for an update, you won't see your email. To be safe, do a flush at about 23:00 GMT to make sure your blocks get included in that days stats run.
Moo!
dB!
distributed.net Human Interface
Yep.
Change all of your clients to use the new email address, and submit a block or two. On the main stats page, click on "Edit Your Information". Enter your email address and your d.net password (If you don't have a password, instrictions for retrieving it are found elsewhere in this thread). Go to the bottom of the page and retire your address.
Where is this announcement? All I see is one line mentioning that new works is being done continuously. Some announcement.
According to the technews, they generate 15.000 blocks a day from the tapedrives.
According to the stats, it takes each of the 600.000 users about 1 1/2 CPU-days to process a block.
So they generate 15.000, and consume about 400.000 units per day. Go figure.
My computers stay with PrimeNet until they are overbooked by less than a factor of 3. (but THEN I'll be back!)
http://www.mersenne.org. At least there are enough prime numbers to go around.
http://www.distributed.net/clients.html
lol! me too
Yes, there is a existing project to use distributed computing for real work and allow researchers and others to plug in.
Cosm
- Adam L. Beberg - The Cosm Project - http://www.mithral.com/
That too is covered in a FAQ. To be exact, it's covered in the StatsFAQ
Is it really that hard to just go to the main distributed.net page, and type "change email" in the search box on the bottom-left? First result that popped up. Took me about 10 seconds to find it.
Really... READ those FAQ's please people. USE the search engines. It makes life so much easier. Cheers.
----------
'We have no choice in what we are. Yet what are we,
but the sum of our choices.' --Rob Grant
----------
'We have no choice in what we are. Yet what are we,
but the sum of our choices.' --Rob Grant
Lesson #1 in computing: When all else fails, RTFM, or in this case, the FAQ, which, if you bothered to look, is linked from every single stats page right at the top, and contains full instructions for how to join a team. As for the username, you answered your own question. If the only thing it asked for was an e-mail address, then it's a fairly safe guess that your e-mail address is your username, and you have to go fill in a password somewhere (that is answered by the FAQ)
Distributed.net Stats FAQ
-Cheetah
I sent you the password... (I sent it to fuddoson@hotmail.com)
Now check the slashdot stats and choose to sign in.
BTW,
According to the stats your avg. rate is 200kk/s and not 800kk/s.
(or is it because you ran it only part of the time)
---
---
I'm going to live forever, or die in the attempt.
---
-B
Ash and Hickory, straight-grained and true, make excellent bludgeons, dandy for the cudgeling of vegetarians.
and just how do I sign up for this any way?
I am running distributed.net on my machince at work. and I avg about 800 keys per sec and always running it.
I have gone to the stats site, gone though it, click on "I want to join this team" and got a usernamer and PW box, I have tried several things in it, nothing seems to work
and when I set up my cleint I don't recall getting a username and password, all it wanted was a email address.
so if u want me on ur team, tell me how to join up.
-------------------------------------------------
even though due to poor documentation here and at the FAQ. I managaed it.
first make sure you have at least run the distributed.net client at least once, and useing the email u setup in ur client. goto stats page
goto view stats. and in the participant search box, insert the same email u used in ur client, go though that, and at the bottom click on mail me my password...
then check ur mail. get ur password, go back to the site, and now in ur username use ur email and for the password use the password that was in the email sent to u
BTW I find this to be an extremely ODD login process. does anyone else think so?
-------------------------------------------------
while I actually figured this process out on me own.. (see "how to sign up...") /sec
I would like thank you all for the effort u wasted in tellins me and everyone else that is possible wondering just how to sign up to this thing
and yes my stats say 200 Keys
but that will change to 800, because right now I am avg. 900 keys / sec and over night, I avg more than 1000.
and since I am always connected to the net, I always have this thing runnings. so its cool.
-------------------------------------------------
Wow, I love RC5 trash talk. I think it may be time for me to join a team. Slashdot sounds better and better, especially after these threats from anandtech... :)
-Chris
CComp> Too many people look at a thing and say,
CComp> 'This should be used for X instead of Y,
CComp> and right now! Change it! Change it!!
CComp> Nownownow!!' Mebbe the intermediate steps
CComp> are NECESSARY to get to where they want it,
CComp> but no one cares aboutthat. Gotta be
CComp> instantaneous or nothing. Call it the
CComp> McDonald's Syndrome.
Their FAQ makes little mention of using this factorization software in the larger context of generalized scientific and mathematical research. The closest they come is stating a goal of "feasibility of cooperative networked multiprocessing," with no mention of client security issues, or an API set for future developers, or how timesharing would work. That's what my question was about -- indeed, a cursory search of distributed.net's pages revel no developments along these lines. Or have you seen some place where they address these questions? Call it my McDonald's curiosity.
"Whatever happened to fair use?"
-- Duff-Man
Good points. However, there are a *lot* of research problems (especially ones in, say, combinatorics, or modeling) that can be solved with brute-force parallelism. Back in school the computer science LAN was almost perpetually bogged down by a math teacher's parallelized research (combinatorics, IIRC), running as background processes.
Assuming that the researcher's computational problem can be cut into fine-grained enough chunks (the way SETI's work and prime number factorization can), a sufficiently large generalized distributed computational system would be a god-send.
"Whatever happened to fair use?"
-- Duff-Man
Could this stuff be put to a *really* practical use? Surely some math professors out there have some linear algebra number crunching to do, or some physicist has weather modelling simulations to make, or the like. Not that generating primes and scanning gads of radio waves for intelligent signs isn't nice and all, but it seems to me that gargantuan amounts of parallelized computation time could really help out researchers on tight budgets.
Is there an easy process for researchers to utilize this stuff? If not, it might be a Good Thing to set up.
(It'd have to be secure, I suppose, and in a best-case scenario could involve timesharing without administrative hassle...)
"Whatever happened to fair use?"
-- Duff-Man
Some guy from Anand Tech must have been storing up blocks:
mark_hodgkinson@bigfoot.com did 120,403 blocks yesterday, at an average of 374,079 KKeys/sec.
He jumped 25988 places in the indivdual rankings from yesterday to today.
Why do people get off doing this?
His average keyrate for the past 170 days is 6338 KKeys/sec.
That's not nearly consistant with his huge 'spurt' of blocks yesterday.
This is:
A. One of those people that store up a huge number of blocks for a month or so, then send them all in, and bask in the glory of seeing themselves as #1 in the daily rankings.
B. A stats glich, that only affects this one guy.
C. A hack.
Which do you think is correct?
While I'm at it...
Anandtech is averaging 199,219 KKeys/sec.
slashdot.org is averageing 731,000 KKeys/sec.
Heh. Lots of luck catching up at those rates. You're only 102,868,204 blocks behind.
I've got 3 machines running the seti client. All 3 were sent, and are currently working on, the _same_ work unit, recorded on Feb. 21, at 4:44:20. Coincidence? The new unit retrieval took place, for all 3, over a period of more than 5 hours. Well, that's more than coincidence, in my book.
Yeah but Slashdot will topple Team EvangeList from overall 1st place at this rate in about a year (if the contest lasts that long). AnandTech, even if they double their rate it would take them 3 years to catch up.
my 28,000+ blocks? Kinda sad that my one computer puts me at about 76th in the team rankings. For anyone that cares, it is a PII 400.
Do really dense people warp space more than others?
I didn't look at the fact the team listings were for one day only. Ah well.
Do really dense people warp space more than others?
Are you referring to SETI@HOME? Because on their website to search for a group, people are typing "slashdot" instead of "team slashdot". So everybody, make sure you all join "TEAM SLASHDOT" not "SLASHDOT"
Why are there 2 slashdot teams? Both teams should be combined for best effect!
There can be only one.
Out of overclocked celerons! No Shit! It's called Warpcore.
Here's the link:
http://www.montac.com/
V2K
You need to go to your stats page and at the very very very bottom it has a link with "Email me my password." That's how you get your password. Then use that email and password to log in and change your team affiliation
ufdraco