Hi, On Thu, Feb 05, 2009 at 01:42:59PM +0100, Richard B. Kreckel wrote:
Strange. I just ran current master HEAD ten times and it finished fine.
Also: it is stopping at the old code, right?
Short answer: no (it's a bug on its own). Long(er) answer: The benchmark is intended to a) check for bugs in mod_gcd (modular GCD algorithm for univariate polynomials using vector of coefficients to store polynomials), b) find out how (in)efficient GiNaC::ex is as a representation of univariate polynomials. To do this it calls different implementations of the same GCD calculation method. So, it's not supposed to call chinrem_gcd() in first place, but for some reason it does. It's a bug in the test itself. Fortunately this bug uncovered much more severe bug (in multivariate GCD code). I'll fix the test after I find out what is wrong with chinrem_gcd(). Best regards, Alexei