GMP used during 3 and a half years to solve MIT's LCS35

Hans Åberg haberg-1 at
Wed May 8 16:25:25 UTC 2019

> On 8 May 2019, at 17:46, Bernard Fabrot <bfabrot at> wrote:
> If anyone could test this on an older GMP version / older hardware it would be amazing.  Actually tests on more recent hardware (like i9 or AMD) would be great too.
> That'd be 40 million iterations. It's from someone on mersenneforums (in a recent thread discussing the solution to LCS35 I found) and contains two tests (my code doesn't have any benchmarking in it and ain't as clean: it has all the code related to creating backup files with the intermediate results so that I could resume the computation).  On Core i7 CPUs it takes very roughly from 50 seconds to 30 seconds for the 100 m iterations.
> No biggie if nobody has the time to try this.

Just so that there is at least one reply :-), the program on a 3.2 GHz Intel Core i3 gives 114 and 97 seconds.

More information about the gmp-discuss mailing list