In the second volume of The Art of Computer Programming, Donald Knuth included a puzzle. A literary quotation is encrypted using the RSA algorithm, and we will decrypt it by factoring the public encryption key! This effort will use the lower-memory lasievee application, so more of you will be able to participate. The workunits for this challenge start with Gtaocp, and they will start going out shortly. Let's crack this as quickly as possible!

My slight concern is that the score seems little low for these at 36.

This one is a lasievee task awarding 44 credits each. The somewhat easier tasks using lasieved are awarding 36 credits each. If you wish, you can disable the lasieved tasks in your NFS@Home preferences.

Have all the Gtaocp tasks been sent out? Reason I'm asking is because as I write this, on the Status of numbers page It still shows TAOCP (Puzzle time!) status as Sieving. I requested tasks not so long ago and got eight 1621_73 tasks. Currently 81.28% Pushed.

The "Status of numbers page" is updated by hand, so there is a delay between finishing the sieving and updating the Status page.

Thanks I weren't aware it was updated by hand

I'm just waiting for the last few results to come in, and I'll start postprocessing probably later this evening.

Thanks for the update Greg

I started the linear algebra run this morning. It should take a bit over two weeks to finish.

Thanks Greg. May I how many CPU'S & at what speed it will take to do the linear algebra run?

The LA is running on eight computers, each containing a 2.4GHz Core 2 Quad processor and DDR2 memory, connected by Infiniband. It's currently 19.8% done with 280 hours remaining.

That is some serious processor power 78.8Ghz. I gather all the output data is sent to 1 hard drive or location? Once you have all the post processed data how is it stored How much space in total have you used with completed work from the project? Thanks for taking the time to answer my questions.

The files for this factorization total ~49 GB. That consists of ~40 GB of relations from NFS@Home sievers stored in a single plain text file, a ~9 GB binary matrix file, and miscellaneous smaller files.

For the largest so far, 2,1061-, the files total 125 GB including an 86 GB text file containing the relations and a 37 GB binary matrix file.

Once we have completed a factorization, I typically delete the associated files. However, I have saved the recent largest factorizations:

58G ./6p379
111G ./2p997
125G ./2m1061
57G ./5p433
118G ./2m1031
113G ./5m409
58G ./3m607
636G .

6p379, 5p433, and 3m607 are smaller because I've compressed the text relations file. I just haven't bothered to do that for the others. Fortunately, disk space is now cheap!

Hello Greg!

Why the linear algebra must run on your computer, not us?

The linear algebra algorithm requires nearly constant communication of a large amount of data between all of the computers used in the calculation, so it must be done on a local cluster connected with a high-speed network. The Infiniband network that the cluster has provides 10 Gbit/s bandwidth in each direction for each computer with a communication latency less than 2 microseconds.

Thanks!

What is the prograss of the lineal algebra now, Greg?

linear algebra completed 19054871 of 23456239 dimensions (81.2%, ETA 65h30m)

We should have the factors by this weekend.

Are we there yet?

I can take a sleep tonight and there is the result at tommorow morning(sorry, I am in vietnam and it's 10 o'clock now, GMT+7)