$30,000! (no, not spam)

[tommyboy123x]

haha, basically... the past ppl have used like 100 P2's, but i figure that a fraction of newer computers will be enough

also, i forgot to add this to ur PM - you can set the affinity of a process so that it is ONLY in "processor 0" or "processor 1" by going to the task manager, right clicking a process (such as firefox.exe or iexplore.exe, going to affinity, and then clicking the processor(s) you want the task to be able to run in

 

[Nitestick]

yeah i know. but is it possible to run multiple instances of it on different work?

 

[tommyboy123x]

PM - has a lot of spec. details

Anyone who just skipped to the last page or is already working on the crunching, please read the first post i made... it has been updated mainly with a precise payment plan (so you understand what to expect)

 

[Nubius]

A "payment plan" has been worked out and we have decided that the discoverer of the money will be awarded $2000 (leaving 28,000 to split) no matter what. The cut off for getting an equal share of the money (not the bonus $2000 for getting hte correct solution) is May 1st. If you simply help but do not find the answer after May 1st, you will get no money. No extra money will be given for those who do anything beyond what is asked, that is what the $2000 "royalty check" is for.

Aright, so if I'm getting this right, lets suppose there are 15 machines.

Machine number 15 finds the winning number. So machine 15 gets $2000 and then whatever his share is of the remaining $28,000? And what about this whole after may 1st deal. May 1st is pretty close. So if you're just crunching numbers after May 1st, you get absolutely nothing or you're just saying you didn't get that $2000, you'd only get what's split between 15 people and $28,000?

Why dont you just go for the biggun $200,000

 

[Mr_Threepwood]

I wanna know how the program you wrote works. You didnt just make a strait up brute number cruncher right. I mean you can obviously skip every 2 sumbers since nothing even will be divisable, plus you can also do checks for each number, like for example if you know that the large number isnt divisible by 3 then you can elimate ever 3rd number, same with 5, etc. But there will become a point where doing that check becomes slower then brute checking.

 

[Nubius]

Dont know, it's using a java applet with methods I've never heard of before.

Anywho, I've got 3 instances of it running right now and both cores are maxed out, could probably do 4 instances with 2 on each core since right now I've got 2 instances on core1 and 1 instance on core0

 

[V-tek]

All you need is a program that can perform recursion of the number until it reaches the factor

 

[tommyboy123x]

yea, right now were using the ECM (Elliptical curve meathod) and it is indeed a java applet

this is really just kind of "busy work" you could say for your computer. It is ruling out the idea of smaller numberd factors

in terms of going for the $200k one... you have to be joking me... lol... yeah, im going to find 2 300+ digit factor from a 600+ digit number... ON ONLY A FEW MACHINES!!

and yes, May 1st is close, but this "plan" has been worked out by a few people and myself and was agreed apon. This way you do not do part of the work and get the same prize (you have to be there for most of it)

and finally, yes, as you said nubius you can have multiple processes running at the same time, just record what "curve" you are at and stay within 10k curves of what i said (so for example, i "told" myself to start at curve 0, so i will go from 0-10k, but have three processes running it, so 0, 3333, and 6666). also, if you have a dual core cpu, you should use more than one process because IE and FireFox are not really supportive of multiple cores, therefore you will only reach 50% at the most with one window


EDIT: oh yea, and i just actually had an idea last night (which put me to sleep ) about how to make the program (just try finding out how to use the GNFS (General Number Field Sieve) and impliment it in a program!!). It would start out at 1e105 or 106 (because all the other factors have had the same number of digits in each factor, making 212) and simply skipping every multiple of 2, 3, and 5 (the rest would most likely bog down the program as well as possibly mess it up).

Probly in C++ or C...

 

[Nubius]

I just opened avant and IE and have one instance running in each.

and yes, May 1st is close, but this "plan" has been worked out by a few people and myself and was agreed apon. This way you do not do part of the work and get the same prize (you have to be there for most of it)

Ok so once may 1st hits and it hasn't found a factor I'm supposed to stop then is that what you're telling me?

 

[tommyboy123x]

no no no... if someone wants to "join" or "help" or w/e, and they say this AFTER May 1st, then they are welcome to join, however they will not get any of the prize unless they find the factor. If they do find the factor, they will be given 2000 dollars, and the remaining 28,000 will be divided equally among the rest