p Statistics: Preparations
by JVSchmidt


p Data
Data for my calculations was received from the official and public ftp-site of the Yasumasa-Kanada-Laboratory. It holds the decimal representation of the first 4.2 x 109 digits of Ludolph's number. Data was provided in packed form (two decimal digits=one byte) containing 100 million decimal digits in a single file. First step was to repack the digits and compare single digit frequencies with those given by Kanada himself.
This comparision was really necessary because we obtain a transmission error in one of these 42 files.

Hardware
All tests carried out by the software finally lead to simple integer counting. That's why one doesn't need an extraordinary dreammachine to do this calculations. Results can optionally be stored to harddisk, but upper limit for these files sizes are some 100 kByte.
These are the parameters of my ordinary consumer pc used to do the analysis:
Intel Pentium 4; 2,5 GHz; 512 MB RAM; 120 GB HD

Software
I have written a specialized software to analyze PI data: THE p-COLLECTOR.
Here are the reasons I decieded to develop an own program for my needs.

  • The software should handle multipe file sets (42 files) of huge files (100 mill. byte) and analyze any area from digit d1 to digit d2, no matter in which file d1 and d2 are located.
  • The software should perform special calculations which are no standard tests like chi-quadrat for digit frequencies or something like that. Even to proof very long sequences I have designed new test ideas not common. To do this we need a statistic tool that can be programmed. So you can't get out of the way of programming when nostandard evaluations are required.
  • The user interface should allow you to analyze and monitor data at once. You should be able to "jump" to any position in the 4 billion sequence and see what's going on there. More than this, you can localize any given proof sequence in the data.
  • A software that has at least some of the features mentioned above is very expensive and it takes a pretty long time to manage with.

    <= Back to Main