http://sage.math.washington.edu/home/kstueve/T_R_NICELY/
These are the same data that Thomas R. Nicely has published at
but with more data points included. The granularity is 1e10 or
better, which in places is 10 to 1000 times denser than the next
densest available table of the prime counting function available at
http://www.primefan.ru/stuff/primes/table.html
Thomas R. Nicely considers primes, twin primes, prime triplets (both
types) and quadruplets in the newly published tables.
These tables were being used by Thomas R. Nicely to calculate Brun's
constant (the harmonic sum of twin primes) when he discovered the
infamous Pentium FDIV bug.
My main interest in tables of prime counts is that they can be used to
calculate prime counting functions faster than the combinatorial
method.
I am in deep gratitude to Thomas R. Nicely for mailing me these tables
on a CD-ROM.
Kevin Stueve