Hiya Drudge,

One other thing I noticed, you're running a Pentium4 on Windows right? Are you using TJM's optimized apps for Windows? I don't know if your P4 is overclocked or not (it's reported at 3.4GHz, but I know you can OC up to 4GHz or more with the right setup.) Anyhoo, here's one stat from your account here on Enigma@Home....

resultid=5013178 hceyz72_0_5589153_r0 4,067.16 seconds

...it took 1hr 7min and 47.16 seconds to complete the smallest WU Engima@Home has to offer.

For comparison, here my latest hecyz100_0 that has been uploaded....

resultid=5359269 hceyz72_0_5865949_r0 1,109.04 seconds

...so it only took me and my AMD Phenom 9950 OC'ed at 3.05GHz - 18 min 29.04 secs to perform the same amount of work. Yes, I have a newer processor than you, and your Pentium 4 may or may not have the ability to go 64-bit. Also, you seem to be on the BOINC Australia team, so you're spread amongst other projects as well. Here's the point....

If you add another message to be worked on we will delay the project completion (and in reality, we're working on 2 sections of the same message, which computationally is the same as two distinct messages [because the message may be garbled; hceyz's are the beginning and awgly's are the middle of the message, so if there's any mis-translation or missing letters, we can still break the message] so the project length has already been extended due to the nature of this break process). There are over 9600 active accounts on Enigma@Home, and about the 1st 2600 are active crunchers. This means that in order to "make up" for the extra cpu time you'd need to decipher another message without delaying the completion date yet again you'd need to attract another 1300-2600 computers (depending on cpu and whether they're spread over other projects).

Since you don't have an extra 1300 cpu's just lying around (or do you? :-) ) the result would be that MY computer would be doing more of the work on another message than YOUR computer. The extra effort TJM and I have placed into optimization would have been for nothing then. No, I'm not being selfish here, just trying to put a little perspective on the request to add another message to what has already taken a very long time.

Do me, and us here at Engima@Home a favor. If you aren't using one of the optimized apps TJM has placed out there (there's a sticky at the top of the "Number Crunching" section), then please download it and use it. It shaves about 20% or so off your computation time, and will help us get to the finish line sooner, rather than later......


....THEN we'll need a new message to decipher. ;-)

Please note, this is not a personal attack on you, but you gotta understand that adding more work at this point does not help the situation.

Mike Doerner

---

Finally a discussion! Well, it started out like a discussion. I understand that most of the people in the dc hobby really look at this as a hobby and invest a lot of money in computers. Well, I don't give a f**k about points.
I've started running dc applications because I liked the idea of using spare cycles for something useful. For me, the key word here is spare. Of course, I'm watching my rank, joining competitions (on a small scale), all that stuff, but I'll never buy a computer solely for the purpose of running something full time, consuming power. My greatest motivation is doing something useful, something that can be finished. I'm not judging anybody, just stating my point.
I don't want to speed up the project by attacting more people from other projects, that's marketing and that's not my area. This algorithm can run on forever, there's no finish line if we don't set one, and the message may very well be corrupted. We should switch to other messages (I think most of them will break in months), then come back to the hard ones. We'll fill in more holes in history regardless of computing power.

Adding more ciphertexts isn't an easy task. The server was designed as a wrapper for M4 project, so in theory it only sends out workunits downloaded from original M4 server, and there are only two parts of the same ciphertext available.
However, at some point I needed to test new versions of project daemons and/or apps, so I added support for 'local' ciphertexts. I think it worked fine, one additional ciphertext taken from here: http://cryptocellar.org/bgac/ - the one starting with KEJNQ was available for a short period of time, assigned to early version of current application.

So theoretically the server can support multiple ciphertexts, 'local' and 'wrapped' at the same time. But there's a problem - while some of the backend daemons are shared between apps, some others are not - for example validators. Adding more ciphertexts increases the server load and database size (which also has negative impact on the server's performance), especially if the workunits are short like the M3 messages.
The server is just an average home PC (well, at least it was average when I bought it), and already has some problems (like an 'Internal server error' when contacting scheduler) when the load gets quite high, so stressing it even more is not a good idea.

---

M4 Project homepage
M4 Project wiki