GMP 6.2 is upwardly source and binary compatible with 6.1, 6.0, 5.1, 5.0, 4.x, and 3.x, except for applications that use the semi-documented mpn_bdivmod function.

Changes in GMP 6.2.1

BUGS FIXED

• A possible overflow of type int is avoided for mpz_cmp on huge operands.
• Overflows are more carefully detected and reported for mpz_pow_ui.
• A bug in longlong.h for aarch64 sub_ddmmss, not affecting GMP, was healed.
• Mini-GMP: mpz_out_str and mpq_out_str now correctly handle out of range bases.

FEATURES

• C90 compliance.
• Initial support for Darwin on arm64, and improved portability.
• Support for more processors.

SPEEDUPS

• None, except indirectly through recognition of new CPUs.

Changes between GMP version 6.1.* and 6.2.0

BUGS FIXED

• gmp_snprintf now correctly reports an error (returns -1) when snprintf does.
• Conversion to double is now more robust even if the configuration process does not recognize the float format.
• mpz_powm could return a not completely normalised value, when the exponent was 1 and the base < 0.
• mpf_set_str could create invalid mpf_t variables for input strings with many leading zeros.

FEATURES

• New C++ functions factorial, primorial and fibonacci for mpz_class.
• Functions to detect primality now substitute the first 24 Miller-Rabin iterations with the BPSW test.
• Mini-GMP: new functions mpz_2fac_ui and mpz_mfac_uiui.
• Mini-GMP: mpz_sizeinbase, mpz_get_str, and mpz_set_str now support bases up to 62.
• Mini-GMP: added support for the mpq_t layer.
• MIPS r6 cores are now supported.

SPEEDUPS

• Major speedup for AMD Ryzen and Epyc thanks to an extensive set of assembly code.
• Major speedup for IBM POWER9 thanks to assembly code making use of new madd instruction.
• Speedup for 64-bit ARM CPUs thanks to new/improved assembly code.
• The n-over-k function mpz_bin_ui has been reimplemented for great speedups for large operands.
• Speedup for the worst case of mpz_perfect_power_p.
• Speedup for gcd for small and medium size operands.
• Speedup for really huge multiplies thanks to much larger FFT tables.

MISC

• Internal representation of the mpz_t variables now supports lazy allocation; memory is allocated only when a value is stored.
• Small improvements and better coverage for the test suite.
• The tune/speed program can measure some more functions.
• The low-level function mpn_mul no longer diverts to mpn_sqr, users should call mpn_sqr directly when applicable.
• New installed files gmp.pc and gmpxx.pc, for use with pkg-config. Contributed by Hugh McMaster.

This release would not have been possible without the continuous work of improvement, revision and refinement by the whole GMP team: Torbjörn Granlund, Niels Möller, Marco Bodrato, and Marc Glisse. Please see the GMP manual for a complete list of GMP contributors.

There is a public repository for GMP, please see the GMP repository usage instructions for more information.