Pentium III 550… 256 meg ram… Going to upgrade to an Athlon or PIII 700-800 soon… 11.7 hrs avg time… Yessss! The only person with a lower average than I is Gatsby. What you go, Gats?
–Tim
Pentium III 550… 256 meg ram… Going to upgrade to an Athlon or PIII 700-800 soon… 11.7 hrs avg time… Yessss! The only person with a lower average than I is Gatsby. What you go, Gats?
–Tim
Okay, call me stupid. I mean to type 17 hr avg.
–Tim
Presently, I look for a job. I hope to have one very soon, though. Previously, I’ve been in computer sales for the past 3 years. As far as the work unit per hour thing, see previous posts about not running the screen saver (graphics, anyway) or about running the CLI version instead.
I never understood why the SETI team made a resource hungry screen saver. I guess they figured they were competing with the Budweiser frogs or something…
Thanks to the other 7 but…
WE HAVE EIGHT MEMBERS! WE NEED MORE! WHAT IS YOUR COMPUTER DOING WHEN YOU AREN’T USING IT?
Read as: “We need more people”
C’mon, we need more!
I just joined!
average: 10 h 33 min. Are the rest of you doing the calculations by hand?
Seriously, though, are all the data packets similar as far as processing time is concerned, so that the computer is the only factor affecting the processing time? I have only done 18 data sets, but they have all been the same, give or take a few minutes. Is it possible that the data are heterogeneous enough that there are some data sets (or some sections of the sky) that are “easier” than others?
Is SETI a reliable speed testing program, at least as reliable as the “real-word” tests that use Photoshop or other processor-intensive applications?
FYI, I am using an iMac DV with a 400 MHz G3 processor.
malden
All of the packets are of identical size, which means that the amount of time it takes to process each will be approximately the same (within a few minutes, if not seconds).
IOW, yes, the speed of the computer is the only thing which affects the processing time.
It’s hardly a standardized benchmark, but it’s a lot more interesting and a lot easier to set up than a lot of the ‘real’ benchmarks out there.
The two fastest machines I’ve personally seen have been a 450 mhz G4, 256 MB ram, which took about 7 hours per packet, and an 800 mhz P3, 128 MB ram, running RedHat 6.2, which took about 3.5 hours per packet.
Macs on average seem to go faster than PCs, but the fastest PCs kill the fastest Macs. I think it’s because the application is very memory-intensive, and Macs tend to have a lot more physical memory in them than PCs (they need it more).
[cynical humor] I’m not sure who Tom Waits is, but my mom has been saying that phrase to my sister for at least 20 years now. Way before 1999. [/cynical humor]
::ahem:: Erm, I think you misunderstood me. I asked what you got, computerwise, not what you do… But thanks for the information… I’m in comp sales, too.
–Tim
Not quite. You will see quite a bit more variance than that. In fact, the work units that take longer to process have a better chance of being useful to the project.
The reason being that when the software finds strong signals it runs additional tests to make sure that they are not terrestrial.
Also, every now and then you will have a work unit that finishes up extremely quickly (like 10 minutes or less). Usually when that happens it’s because there is so much terrestrial interference the program basically says “screw it”.
The work units are all roughly the same size (though not identical) but if you are averaging 10.5 hours you will probably see your times vary by as much as 45 minutes or so.
aenea, happy? (I never heard the cross one before Waits but it’s pretty funny.)
Tim, actually, you asked what I “go”. Sorry, I thought you meant do, not got. Anyway, the computer is a Celeron 366 overclocked to 566 with 128 megs. It usually finishes work units in about 8.5 to 9 hours. My average is higher because I used to run it on another machine as well.
We have 10 members. Anyone else?
This needed to be bumped anyway…re the speed issue, I have noticed that my computer processes the data packets quite a bit slower when I am at school hooked up to their network. The times when on the network are usually 24+ hours, while when not on, it’s more like 12-15 hours. I’m not very familiar with how networks work, so can anyone explain why being on the network sucks up so much of my processing power? Is this typical, or should I talk to the admins about it?
[playful challenge]
Hey, come on, you guys - rack up some REAL units for Cecil!
[/playful challenge from an afca lurker and work unit donator to SetiTeam alt.fan.cecil-adams]
I’ve got 70 data units, but it takes forever to run each packet (and if I run it in non-screensaver mode, it slows down my clock!). I wonder if I’m using an old version or something.
Anyway, I’ll look into joining.
OK, I joined the Straight Dopers team (almost accidentally joined the AFCA one). It asked me for my password and I thought, “Yeah, right.” Luckily, they e-mail it to you right away.
Figured that even though I didn’t post much I would toss my stats in with you guys. Gotta love the straight dope. I currently have only one computer running it and that is at work. Keep forgeting to get my computers at home involved. Will have to do that soon to help out the cause. Maybe someday we can beat those guys at the top with the unix supercomputers. hehehe
Dave
Ahhhh… A thread that I have to pop into. I’ll join the team.
I’m running Seti@home at work, on a 500 mhz 256 megs of ram computer… I’ve been at it a while now, and racked up 5504 hours of processing time, and 496 data units completed.
For some reason, I’m really looking forward to hitting 500, my life is that boring. heh. Time to look into getting the text only version.
Off to join the group.
Lupin.
(back to lurking…)
I would join, but I’m already on a (small) team. I would feel bad if I left; I’ve already contributed 1800+ units and over 5 years CPU time.
PeeQueue
OK, I’ve added my 182 units. I have recently entered the rarified world where less than a thousand people share my rank.
At home, I started with a 200MHz MMX w/64MB running Win95 sometime last December. That took about 60 hours to process a work unit.
Now I have a P3/700 256MB Win98 at home, which takes about 16-18 hours.
At work is a P2/400 128MB NT4, which takes about 10 hours.
I’m working on getting the old 200MMX running Debian/GNU Linux and networked to the new box. When that happens, I’ll see how it runs on a real OS.
Cheers,
Rick
I’ve only posted here a couple of times, but figure what the heck, I’ll throw in my data units. I’ve got about 800.
Ok I signed up, waiting for my password to join the SD group.
I threw my 80-something units into the group. I’ll probably skew the CPU time ratings to the extreme low end…most of my work has been done on '486 machines running W95, although there are several of them crunching away. Even after disabling the screen display they take about three days to chunk through a unit.