Check recent changes haven't broken anything

[JoshKeegan]

Need to check a full generation & some searches are all still working after the changes this year.
To Test:

• Generate 4bit digit from y-cruncher output
• Generate suffix array
• Generate precomputed search results for lengths up to 9
• Run a few searches and compare against results from pisearch.joshkeegan.co.uk

SHA1 hashes for generated files:
pi_raw_5bil.4bitdigit - 7b36f3ea7da790c3458ebd5d3c1495e476a6e36c
pi_suf_5bil.bitAligned - 25fc8ca5d8ca9fb81e11c08dccc5355575ee3ad7

Precomputed search results:
1 - 50204d9ddcec042dde5ae1ede7270f779884286b
2 - daf4c1f560755a7ed8a85b2eb8c337bd2614ad0c
3 - 04534a9705d38c1f898eb664ae0ec9bb443b3ea3
4 - 9d80facff09fcfd12142ce0256720c7a7904a003
5 - 407585f3d25fc89bf74dbd6679f2e9dfad08d9ab
6 - f1bc44d362b3c52553be0ebcb5ff710b551b27d0
7 - 3f0b991bfda50689dd1c48f1e29a4b727f5e7300
8 - 142c4ccb55261ea902ef4ced00693c264ae9a908
9 - ee25b617376d7ef991138187d4d92ec033907e51

Should also make a note of how to do each step so that I can write a quick howto

[JoshKeegan]

Generated & checked hash for 4 bit digit file & correct.

Instructions:

• Generate digits for desired constant (example follows y-cruncher decimal output)
• Launch StringSearchConsole
• Use option 4 to set the working directory to the location of the y-cruncher output
• Use option 17 to convert the y-cruncher decimal output text file to a text file of just the raw decimal places
• Use option 7 to convert the text file of just the raw decimal places to a file that uses 4 bits per digit

[JoshKeegan]

Generated & checked hash for the bit aligned suffix array & correct.

Instructions:

• Use option 8 to load the 4bitdigit file into memory
• Use option 10 to generate the suffix array for the loaded 4bitdigit file
  • Add note about memory usage & how to calculate it
• Use option 20 to save the suffix array to disk

[JoshKeegan]

Maths for calculating total bits required for suffix array:

Where:

• s is total bits
• b is bits per value
• n is number of values to store
• m is number of possible values to be stored (assuming can be compressed to use all combination of bits, which in this case requires no work as they're a continuous series of integers. Also in this case, m== n since they're all unique indices.
• S_B is total number of bytes

To calculate total memory usage for a suffix array generation, would also need to add the total data size, which is trivial to calculate: ceil(n/2) (ceil as can only store whole bytes, div 2 because 8 bits per byte & 4 bits per digit stored).

Would perhaps make more sense for anyone trying to reproduce this for this to be built into the program so you don't need to run the numbers by hand.
If so, should a reversed version of the formula also be included so you can enter a given quantity of RAM and calculate the number of digits you could handle (in memory)?