We're almost there. I've gotten back from Intel and AMD, they are digging out the needed info for us.
I went ahead and made the change to mpz_invert (and its test code). As soon as the assembly hacking is done, I'll start he tuneup runs. They'll take a couple of days. Then tom, hannah, biko, pile will be unavailable (and not run any virtualisation guests). * Make some other sec functions from Niels' list public? We really want to allow straightforward implementations of any common group used for cryptography. (4 h) * Split up outer loops of mpn/x86_64/coreisbr/mul_basecase.asm. * Split up outer loops of mpn/x86_64/coreihwl/mul_basecase.asm. (well underway, ~20 h remaining) * Make sure the test suite exercises all new sec code. (The coverage status can be gleaned at https://gmplib.org/devel/lcov/.) (? h) * Run tuneup with large limits for critical machines, update many gmp-mparam.h files. (Waiting for asm checkin.) (4 h + wait) * Update config.guess and mpn/x86*/fat/fat.c with upcoming AMD and Intel CPUs. (Waiting on input from AMD and Intel.) (2 h) Torbjörn _______________________________________________ gmp-devel mailing list [email protected] https://gmplib.org/mailman/listinfo/gmp-devel
