This program must be run with GMP (https://gmplib.org/)
This program using many number theoretic functions in GMP library (see https://gmplib.org/manual/Number-Theoretic-Functions.html), and we use the GMP function mpz_probab_prime_p to test the probable primality of the numbers, this function is combination of the Baillie-PSW probable primality test (https://en.wikipedia.org/wiki/Baillie%E2%80%93PSW_primality_test, https://mathworld.wolfram.com/Baillie-PSWPrimalityTest.html), the Miller-Rabin probable primality test (https://en.wikipedia.org/wiki/Miller%E2%80%93Rabin_primality_test, https://primes.utm.edu/glossary/xpage/MillersTest.html, https://www.rieselprime.de/ziki/Miller-Rabin_pseudoprimality_test, https://mathworld.wolfram.com/Rabin-MillerStrongPseudoprimeTest.html) of first 26 prime bases (see https://oeis.org/A014233) (we use reps = 50, thus the first 50−24 = 26 prime bases), and trial division (https://en.wikipedia.org/wiki/Trial_division, https://primes.utm.edu/glossary/xpage/TrialDivision.html, https://www.rieselprime.de/ziki/Trial_factoring, https://mathworld.wolfram.com/TrialDivision.html) to about 109, thus all numbers in the data are Baillie-PSW probable primes (i.e. both strong probable primes to base 2 (see https://oeis.org/A001262) and strong Lucas pseudoprimes with parameters (P, Q) defined by Selfridge's Method A (see https://oeis.org/A217255)), i.e. either primes or Baillie-PSW pseudoprimes, and no known composites which pass the Baillie–PSW probable prime test, and no composites < 264 pass the Baillie–PSW probable prime test (see http://ntheory.org/pseudoprimes.html and https://archive.ph/IuzWs), thus if a number in the data is in fact composite, it will be a pseudoprime to the Baillie–PSW probable prime test, which currently no single example is known!