Re: [gambit-list] Some floating-point benchmarks