PrimeGrid
Please visit donation page to help the project cover running costs for this month

Toggle Menu

Join PrimeGrid

Returning Participants

Community

Leader Boards

Results

Other

drummers-lowrise
21) Message boards : PrimeGrid : Feature developement/bug fixing list (Message 1530)
Posted 5535 days ago by Profile Pentium100
another small bug in the frontpage ant user of the day:)

If user is using BBCode (like bold) then the code (not the result of it) appears in the frontpage. Text looks OK in the profile.

For example:
text in profile is : "I am the best"
text in frontpage appears: "I am the [b ]best[/b ]"
22) Message boards : Number crunching : A few questions about RSA (Message 1245)
Posted 5565 days ago by Profile Pentium100
I see PrimeGrid's intractible algorithm as a bad thing to the Boinc community.

How come?

Because people don't pay attention to the finer details of a project, don't realise there is no realistic chance of finding the factors, and therefore crunching an intractible problem has little scientific benefit - hence more science would come out of running one of the other projects. You were being ironic, but you are literally correct when you say you might as well add 1 up a billion times.

If the point of the project is to test the scheduler out - that's slightly different. But when will you have tested it enough - and what further testing do you need to do to convince yourself the scheduler works?


The point about whether the project cost anything to develop is mostly irrelevant when it comes to making a decision about whether to devote CPU cycles to it for the following reasons:
1) It costs me a non-trivial amount of money in electricity to do the number crunching.
2) I could be using those CPU cycles to do something more useful - like looking for Protein Structures.

Your choice, as I said earlier, I do not for you to crunch for me or for anyone else. Heck, I even do not force you to crunch at all! Save your money, if so.

I have already (mostly) deactivated PrimeGrid crunching and, in the absence of further information on the testing requirements for the scheduler, would suggest others take similar consideration.

I am not anti your scheduler development - or even a project to factor numbers. I am anti the idea that your approach to factoring is in anyway a useful thing to do. And I think the intractibility of the problem needs publicity exactly so other people can choose whether it's worth allocating cycles to it.

Regards
Chris


This is the only project where I can stay near the top:) And I know we will never find the factors or others will find sooner than us.
23) Message boards : Number crunching : Old and New URL (Message 1186)
Posted 5567 days ago by Profile Pentium100
Hi

are there any consiquences to ataching to the new url but not detaching from the old


Set that your client will not get new work from the old URL. Otherwise you will get work from both URLs and when the old URL is closed you will not be able to report results and get credits:)
24) Message boards : Number crunching : A few questions about RSA (Message 1148)
Posted 5568 days ago by Profile Pentium100
Go read: http://www.x5.net/faqs/crypto/q48.html then http://www.math.uiuc.edu/~landquis/nfsieve.pdf and http://www.crypto-world.com/code/NFS and http://www.math.ttu.edu/~cmonico/software/ggnfs

Seriously - I would suggest you pause the project until you have implemented (or otherwise obtained code for) something like the "number field sieve" algorithm. Have a search around on google - you can probably find other links than the above.

And is perl really the best choice of language to do this sort of number crunching with?

Regards
Chris



Perl is used only for sheduler - program which sends workunits and receives results.

The crunching program is written in c++.

25) Message boards : Number crunching : A few questions about RSA (Message 1145)
Posted 5568 days ago by Profile Pentium100
Oh just one other thing - what I was thinking when I was originally talking about prime number algorithms is... One way to approach this is to generate a list of all prime numbers and see if any of them are a factor - but I don't know how efficient generating sequential primes is relative to what you're doing.

Checking all the odd numbers is unnecessary and too much work. If you skip multiples of 3 (which is trivial to code), you only need to check a third of all numbers. If you skip multiples of 5 as well (again trivial), you do about 26.7% of total numbers. Add in multiple's of 7 and it's about 22.9%. So for a few minutes extra work, you've reduced your work load by over half - it's still intractable though.

And I'm sure there are much better approaches than this. If I get some time, I might do some research - but I have a life...

I'm still bemused by why you think it's fun to attempt to solve a problem using an intractible approach...

Regards
Chris


To generate all primes would be less effective, because:
1.Prime list would require a lot of diskspace.
2.Checking if number is a prime is longer then dividing the big number.

Personally, I think that there will be a few updates to the algorithm:) This project started as MD5 hash cracking (in case you do not know) which were generated by Rytis himself. Now we do something more useful:)
26) Message boards : Number crunching : A few questions about RSA (Message 1115)
Posted 5569 days ago by Profile Pentium100
1) What is the expected amount of CPU time it will take to complete (in some well understood unit such as 3Ghz Pentium 4 CPU hours)
2) Is the time to find an answer partly down to chance: e.g. is it as likely that an answer will be found tomorrow as an answer being found in a years time - or is there something like a bell curve (or some other shape) probability curve distribution for when an answer is likely.
3) Is there a definite upper bound to when an answer will be found - and if so approximately what is it
4) Once the $20,000 answer is found, is the intention to try the next prize up?

Regards
Chris


1.Very long. Unless we have more hosts than SETI we will finish by chance the answer is near or will not finish at all.
2.I do not understand.
3.It's square root of the big number, which is ~17627870661228609438117057367762263
850932991564221073301231703278767800
25762012912130622946534606
4.I do not know.

About time: One WU checks 1000000 numbers that can be factors. The numbers checked are all odd, so, divide the square root by 2000000 and you will get how many WUs are required to complete, of course we wil find answer sooner:)
BTW, one WU is completed in 55min on my computer (AMD Athlon XP (Palomino) 2000+@2100+).
27) Message boards : Problems and Help : We Are (Kind of) Back Up (Message 991)
Posted 5572 days ago by Profile Pentium100
Hey! Welcome back, project.

My results are currently uploading. I let boinc handle this in its own time. If I get it right, we should avoid reporting in the results for the moment, right? So we should prevent boinc from contacting your scheduler. How can we do this, without keeping the other projects from accessing the network? Or is the scheduler still down? Maybe it would be better if it is.


Sheduler is offline.
28) Message boards : Problems and Help : We Are (Kind of) Back Up (Message 986)
Posted 5573 days ago by Profile Pentium100
Ummmmmm Errrrrr, what am I supposed to do with the 700+ Finished WU's that are all past their Deadlines to Report now. Delete them, Reset the Project or what ... ???

Their really messing me up at the moment, I have to keep my Network Access Disabled or they keep trying to Upload and they lock up the BOINC Manager because theres so many of them ... So I have to just access the Network just a little bit at a time to report WU's from other Projects & then Disable again before the Manager locks up ...



Upload the results. Rytis will know if you find the answer. If Rytis gets the database back (recovered) then you should update the project to report and claim credit. If Rytis does not get the database then you can reset the project.
29) Message boards : Problems and Help : We Are (Kind of) Back Up (Message 984)
Posted 5573 days ago by Profile Pentium100
Your uptime would be very suitable for the server :)

Anyway, I'm getting the money and buying the real hosting for m@h tomorrow (if they do not charge me for the whole month ;)).


But my upload speed is 12KB/s mainly consumed by BitTorrent (I download ~60GB/month).
30) Message boards : Problems and Help : We Are (Kind of) Back Up (Message 981)
Posted 5573 days ago by Profile Pentium100
Two my main projects were down - SETI and Message, now message will be back up, then I will finish LHC and continue to work for Message:)


Next 10 posts
[Return to PrimeGrid main page]
DNS Powered by DNSEXIT.COM
Copyright © 2005 - 2020 Rytis Slatkevičius (contact) and PrimeGrid community. Server load 4.59, 3.82, 3.59
Generated 1 Dec 2020 | 6:01:37 UTC