Extra validation and comments on top of: Rewrite mul/sqr for 32bit/64bit #91

1. 
   sipa commented at 5:53 pm on November 3, 2014: contributor

   Builds further on #73.
2. Rewrite mul/sqr for 32bit/64bit

   - interleave calculation of the lower and upper partial product ranges, and reduction
   - less registers needed, more opportunities for parallel ops
   - "bench" performance: +3.5% for 64bit, +7.5% for 32bit (as tested on a 64-bit machine)

   5fcd6dd39c
3. Add range checks in secp256k1_fe_mul_inner a746cc55a6
4. Annotate inner_{mul,sqr} with relations between inputs and outputs. 0220d2cfb9
5. 
   sipa commented at 5:54 pm on November 3, 2014: contributor

   Resubmitting because Travis apparently missed the previous one.
6. 
   sipa commented at 6:51 am on November 4, 2014: contributor

   Closing as it’s based on outdated code.
7. sipa closed this on Nov 4, 2014

   ---

Contributors
sipa