New Project: MD5CRK

[scoser]

Looks like another round of the MD5CRK project will be starting up Monday, if anyone wants to get involved. I just started up a team for us and it looks like the stats will be restarted for the newest round.

http://www.md5crk.com/

 

[amdxborg]

Thanx I'll check it out!

 

[LANMAN]

Linkified for your enjoyment

--LANMAN

 

[scoser]

Thanks LANMAN...When I posted, it said it would autoparse URLs and I thought it would linkify it for me as a result.

Also, the current version of the program goes from SSE2 -> MMX with no SSE level, which means it sucks on AMD chips as evidenced by a 2.5 Ghz P4 getting 11.2 MegaMD5/sec and my 3200+ getting only 9.1 MegaMD5/sec. With the new client they are releasing Monday, I'll do some rechecking for any improvements.

 

[scoser]

Also, for those of you giving it a test, results are not autosubmitted by the current program. You have to login into the MD5CRK website and upload your results file from there. Hopefully this will change when the project gets out of beta.

 

[kamper]

I've gotta say the project looks alright but that magic button thing is extremely uncool. Using someone's cpu (while they probably don't know it's happening) is not good form. It took me 10 minutes to figure out that it was stealing my seti crunching time, and I only realized that much because I was reading specifically about it. I hope nobody actually uses it.

 

[Stormgiant]

Ya, that IE thing not cool...

Just fired up on one of the P4 to see how it worked...

It ended the run to 100% and sent the results fine to the server, but doesn't appear on the stats yet.

Also, the speed you talk is a little be odd, cause i'm getting 15.3 average with a 2.8@3.45...

 

[scoser]

Stats seem to update hourly on the hour. And what my speed seems odd? I was running it on a lab computer, so it might be off a little bit because of account admin programs that were running on the computer.

 

[scoser]

Originally posted by: Stormgiant
Ya, that IE thing not cool...

Just fired up on one of the P4 to see how it worked...

It ended the run to 100% and sent the results fine to the server, but doesn't appear on the stats yet.

Also, the speed you talk is a little be odd, cause i'm getting 15.3 average with a 2.8@3.45...

As for not appearing on stats, you have to create an account on the webpage and upload your md5_user.xml to the database for it to associate your results with you.

 

[Stormgiant]

This is taken from the log :

Sun Feb 29 02:46:16 2004
Work unit sent

If it was not sent to be accounted, why the message saing so and the result was deleted from the disk ?
Sorry, do not like this project...

Test end....

 

[Spacehead]

Originally posted by: kamper
I've gotta say the project looks alright but that magic button thing is extremely uncool. Using someone's cpu (while they probably don't know it's happening) is not good form. It took me 10 minutes to figure out that it was stealing my seti crunching time, and I only realized that much because I was reading specifically about it. I hope nobody actually uses it.

Yeah, after reading your post i went back to the website & opened the task manager... sure enough "iexplore.exe' was gettin 99% CPU usage. There should at least be a warning message telling you what's going on while visiting.

 

[scoser]

Originally posted by: Stormgiant
This is taken from the log :

Sun Feb 29 02:46:16 2004
Work unit sent

If it was not sent to be accounted, why the message saing so and the result was deleted from the disk ?
Sorry, do not like this project...

Test end....

Well, it was sent and is being counted. It just needs the member number stored in your md5_user.xml to be uploaded under your account so that your member number can be associated with your account.

 

[ViRGE]

Bump for a worth project. Since this thread is a month old, here's an update as to the current status of the project:

* Project has gone live; clients are now all part of the version 1 series
* With version 1, all the necessary base features of the client have been added, including network WU submission, proxy support, headless mode, and test mode
* Stats are completely live too, including user stats and team stats. Individual stats are live, with "top" stats being updated every 24hrs
* TA has 0 WU's submitted for the project; get to work people!

All things considered, I'm relatively impressed with the client so far, although as a crypto nut, I may be a bit biased. Besides support to install itself for auto-loading, it already has all the major features a client needs, including headless mode(something even SETI is still lacking). The stats are good too; they're a little bare compared to the all out statistical assault SETI/Dnet offer, but they're more than adequate. As long as the operators are capable of running the project well, this is definitely going to be my favorite project real soon.

 

[n0cmonkey]

It looks interesting, but that magic button might keep me from playing with it too much. Definitely a worthwhile cause though. Thankfully, OpenBSD switched to blowfish for password hashes a long time ago

 

[MereMortal]

Although I disapprove of hijacking a user's resources, you can avoid the magic button by turning off cookies for the site.

 

[CyGoR]

My Amd A64 3000+ @ 2.2Ghz using the SSE2 core only gets 12.0MegaMD5/sec..
I wanted to test my Apple, but that client is a little lame.. Just as worse as the Seti cmd client Can't see anything about
progress..

Why isn't it possible to set the priority of this project? Besides the proxy nothing settings are available..

I'll be cruching some for the TeAm

 

[ViRGE]

If you want to set the priority, you can use the Task Manager to do it under Windows, or you can use renice on *nix/OSX. I haven't tried the Windows client, but the Mac client defaults to nice 19(on a scale of -20 to 20, 20 being the lowest), so it's already starting out pretty low. Anyhow, the Apples are still #1 in MD5/mhz in this contest, thanks to Altivec, although the P4 does surprisingly well. For your A64, I believe they recommend you use the SSE2ALU2 core, so if you aren't on that, it might be part of the problem.

 

[CyGoR]

Yes, it is running on that core..
My Apple (PB 867Mhz, 640Mb ram) unfortunatly only does 7.33 MegaMD5/sec
While in RC5/72 or whatever it outcruches my A64 with 18%

 

[yelo333]

Well, I would enjoy supporting such a project, as I believe with MD5 being currently known as unbreakable to many people, this is a worthy project, BUT...I believe that javascript thing is too uncool, and anyone who does that doesn't get my computer resources...

100th post!

 

[danzigrules]

I agree the hijacking deal is bad, but why not look around a bit and find things that disable the hijacking. This info is from thier site :disgust:

disable it and refresh the screen and WOW it isn't running anymore

 

[n0cmonkey]

Originally posted by: danzigrules
I agree the hijacking deal is bad, but why not look around a bit and find things that disable the hijacking. This info is from thier site :disgust:

disable it and refresh the screen and WOW it isn't running anymore

Off by default would be even better. And I don't have a problem with their site doing it, the problem will be if other sites pick it up, turn it on by default, and maybe hide it.

But this is looking like the project I'll go to after ecc2 is over.

 

[yelo333]

And, don't forget that he was talking about how cool it would be if he could get a few mainstream sites to run it. He is excited that this might be the way he could get his work done.

 

[nbarb99]

I'm a bit confused here... is this another "crunch for a prize" project like ECC2 or RC5?

 

[ViRGE]

Originally posted by: nbarb99
I'm a bit confused here... is this another "crunch for a prize" project like ECC2 or RC5?

There's no prize, this is just something to do for fun or to get rid of MD5, depending on how you look at it.

 

[Xemus]

I would have done it, but I disagree with the magic button theory. If they change that, I may join.
Worst part about the button is that you can't even collect the points from people who crunch at your site (not that I saw anyway).