I am working on a condensed simplified write-up of my paper to post here. Will be posted soon. In the meantime, download my paper for detailed information about how the random numbers are generated.
Random Number Determination with Floating Point Resolution PDF Download
The generators currently consist of three independent Linux computers with 3.17ghz duo core processors. Those computers are solely used for generating the numbers and are not used for any other purpose. The numbers are not generated by the website server. The setup can scale by adding additional computers as needed. Each box runs a custom C program that implements the algorithm as described in my paper. (I am still working on an abbreviated version to post here).
Each box is independently seeded with a 64 digit random number and does the reseed process a few dozen times then is given a fresh initial seed. The seed numbers were generated by using one generator with an arbitrary seed that ran for a while (I went to lunch) then pausing and grabbing a list from the current numbers.
Using the vernacular from my paper, the raw summation uses 22,271 Voices to generate 19,661 Points. That correlates to 437,870,131 terms, which requires calculating 875,740,262 Sine and Cosine function calls. After conditioning and filtering, the final count of random numbers is about 16k from each cycle. Blocks of numbers are added to the pool of numbers available on the server as needed.
More information will be available soon.
Random Number Determination with Floating Point Resolution PDF Download
The generators currently consist of three independent Linux computers with 3.17ghz duo core processors. Those computers are solely used for generating the numbers and are not used for any other purpose. The numbers are not generated by the website server. The setup can scale by adding additional computers as needed. Each box runs a custom C program that implements the algorithm as described in my paper. (I am still working on an abbreviated version to post here).
Each box is independently seeded with a 64 digit random number and does the reseed process a few dozen times then is given a fresh initial seed. The seed numbers were generated by using one generator with an arbitrary seed that ran for a while (I went to lunch) then pausing and grabbing a list from the current numbers.
Using the vernacular from my paper, the raw summation uses 22,271 Voices to generate 19,661 Points. That correlates to 437,870,131 terms, which requires calculating 875,740,262 Sine and Cosine function calls. After conditioning and filtering, the final count of random numbers is about 16k from each cycle. Blocks of numbers are added to the pool of numbers available on the server as needed.
More information will be available soon.
