1. Ken_g6 (Why 'g6'? That's coming.)
2. Ken Brazier
3. Colorado
4. Computer science student struggling to make it into computer science work.
I probably first read about the idea of distributed computing in a Compute! magazine from the late 80's or early 90's (sorry, I don't want to spend time digging it up). It mentioned the possibility that in the future people would be able to share their processing power to help other people do processor-intensive work faster. This was a fascinating dream, but since I was only about 12, and had only a Commodore 128 with no modem, I wasn't the best candidate.
Fast forward to 1997. I was 16 and had just gotten a nice, fast Cyrix 6x86 P150+ @133MHz
, along with dial-up internet access. Now I've liked prime numbers for as long as I can remember. I've probably written a prime finding program in every language I know, except maybe assembly. So I was naturally drawn to
The Prime Page. It had, among other things, a list of the top 5000 known primes, and
a new program called Proth that could allow a nice fast CPU like mine to find primes in this range, in its idle time! So I found a prime and e-mailed it to Chris Caldwell, who still runs the Prime Page. He e-mailed back and asked me to test an automated sumission form. So I signed up on the form, got
the prover code g6 as the sixth person using Gallot's Proth, and submitted some more primes.
The payoff here was just the wonder of being the discoverer of several of the 5000 largest known primes in the world!
Since then, my CPUs have almost never been idle, although what they work on continues to vary. But for a couple of years, the Proth manually-distributed project was the only one I knew about. I convince my boss on a summer job to put Proth on his machine and some others, my first "assimilations", although I didn't know the term yet. I also registered a range of numbers to check for "the reisel project", my first organized distributed computing, and found a prime that eliminated my range (a good thing).
In 1999 I got to college and, like many freshman students, I got high-speed internet for the first time. Some searched for music, some searched for pr0n, but I searched for distributed computing. On pages that included Proth, such as Yahoo's list of Distributed Computing projects and DMOZ.org, there were other projects. Proth was nice, and fun, but I thought doing something that made a difference in the world would be nice. And I also got a Dell PII-400 for college, so I could actually make a contribution. The search went something like this:
Hmmm... let's see... mersenne.org. The GIMPS project. I already knew about this one. It's another prime-finder, but it takes much longer than Proth. Moving on...
Distributed.net. Distributed, sounds nice. Huh? All they do is a brute-force attack on encrypted text, to see if it can be cracked? The average time to do that could be calculated by hand! Next!
SETI@Home. Ooh, this looks nice! A scientific project to search for alien life in the galaxy! Wow! Let's check it out some more. [/me Googles SETI@Home, or maybe it was MetaCrawler] Huh, this guy (sorry, I don't remember who) says SETI has too many members already, and is intentionally not optimizing their client, like the inventor of the typewriter placed his keys randomly to slow typists. Well, I love optimization, and you don't need me. Next!
DCypher.net. "de-cypher dot net"? Another encryption cracking project? Well, the description says something about other projects, so let's check it out. Hmmm... Says they're only doing another encryption project right now, but they might do something else later. Something to keep in mind for the future, but not now. Next!
Distributed Gamma Flux Simulation. That sounds scientific, and maybe useful. Hmmm, the project looks nice, simulating nuclear waste containment vessels to see how safe they are. Where's the program? Oh, the program's running right now! It's a Java applet! Well, I've heard Java is slow, but they say it can be up to half the speed of C code, so let's try it.
Username? Hmmm. I want a username, so my contribution can be recognized, like in Proth, but I don't really want to put my real name on the internet. Maybe I should reference my other distributed computing efforts (cryptically, my name's there). There are now something like 100 "Gallot" provers, so "g6" is something special. So let's use that, with some of my real name. "Ken" + "g6". They don't allow spaces, but they do take underscores. "Ken_g6". Yeah!
By November 1999, in the stats I have saved, I had completed 1% of the entire goal myself, and I was ranked #7 out of 470 users. A nice rank, although 6 might be better.
I had already become a stats addict. But as it turned out, DCypher.net was going to lead me to my future. On DCypher.net they had a link to a "bulletin board" type system. This would be the place where any new project would appear. I don't remember if I posted anything or just lurked, but at some point I found a post by somebody called "The Magicman" who said he was from a very nice team called "Team Anandtech", and he had a link to more forums. I went there (came here), and these forums had much nicer people than the DCypher forums, with more interesting posts. But I still didn't post here yet.
Meanwhile, the operator of the Gamma Flux project posted that he was having problems with his web service, something about too much bandwidth, and that the project would be forced to end soon if he didn't find something else. Well, I knew that DCypher still hadn't found another project, and he wasn't mentioning them, so I e-mailed him and told him about DCypher. I never got proof that this caused what happened next, but two days later, DCypher announced that they would soon start the Gamma Flux project!
When the original Gamma Flux shut down, I signed up over at DCypher. Their stats had not only usernames, but
teams! I looked over the teams available. There was one for Get3DNow!, the bulletin board that hosted DCypher much like FreeDC hosts SeventeenorBust.com's board today, as well as one for Anandtech. The people at Anandtech were so nice and helpful that I decided to join them.
Meanwhile, the DCypher Gamma Flux client took awhile to get going. And various people from Anandtech were posting messages like "We need help in SETI", "We need help in DCypher", "We need help in RC5". So I signed up for SETI and did a little work there. I needed to be signed up for DCypher for Gamma Flux, so I did some work on their cryptography project. I didn't really think the cryptography project was useful, but they said its stats would be combined with the Gamma Flux stats later. At this point, I began to care less about usefulness, and more about
STATS!!!
Hi, my name is Ken, and I'm a statsaholic.
When DCypher.net folded (that's "closed their doors", not "did protein folding") several months later, I found I had a choice between three fairly useless projects. I could go back to Proth, but nobody really needs big prime numbers (well, cryptographers do, but Proth numbers are useless for that). I could do Distributed.net RC5-64, but like I said the information in that could be obtained from a simple calculation. Or I could do SETI, a science project that didn't need me and wished some people would quit so their servers wouldn't be overloaded.
So I made my choices (RC5 and SETI) based on the "we need help" threads, and the stats, since Proth didn't have team stats. Other scientific projects came along, but many required continuous internet access (by that time I was back on dial-up). Of those that didn't, I heard that United Devices had a weird stats system involving how much RAM you had and how much disk space you gave to the client. I'm not sure that was true, and I doubt it's true now, but I avoided the most useful science project out there because I didn't like its stats.
Instead, I focused on "assimilating" more computers wherever I could. At one point, my school's computer science department installed 16 Athlon 800's with no hard drives; only removable hard drive bays, so that students could put whatever they wanted on the drives. But I wanted to run DC on all of them at the same time, so I found a way to put it SETI on a floppy Linux (LRP), and ran it on the computer other people weren't using. Of course this meant anyone who used the computers for useful educational purposes turned off SETI, so I dropped by the lab every spare moment I had to boot my floppy on the machines other people had turned off.
It got me to 10,000 points, but I think I could have found better ways to spend my spare time.
At this point, I should mention the "permission" issue. I got permission (more or less) to install DC clients on machines I didn't own. But I know people who didn't, mainly from college and not from this forum. One guy who shall remain nameless bragged to me that he put RC5 on all the machines in the physics building offices. I also found an RC5 client being run on the Linux programming lab machines by another student, who denied it when I asked him, although his web page bragged about his stats.
As far as I know, these people weren't caught, but one prominent member of Team Anandtech was.
David McOwen's story made headlines even outside the DC community. You can read more about his ordeal
here.
Right now I'm less addicted to stats than I have been. I finally found a useful science project with nice stats and a decent client that works on dial-up,
Find-a-Drug, but I still respond to the occasional "we need help in SETI" thread when I see it. My CPU will probably never be idle, and I hope that it can sometimes be useful too.
Facebook
Twitter