[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [csmith-dev] Random configuration file for csmith



Thanks. So,in csmith, we need to add the weight for the configure parameter to generate the complex _expression_.

Haihao

On Thu, May 12, 2011 at 6:13 PM, <derek@knosof.co.uk> wrote:
haihao,

> Unfortunately, the empirical data is not direct for csmith usage and
> sometimes random configuration test may expose more bugs.

In many ways compilers are tuned to work well with source that
has the characteristics given in usefigtab, i.e., the common cases.

The way to uncover faults is to jump off the beaten track.
There are two ways of doing this, using more complicated types
and using expressions containing many more operators than usual
(more details of the common case expressions in
http://www.knosof.co.uk/cbook/accu06.html).


> Thanks,
> Haihao
> On Thu, May 12, 2011 at 5:59 PM, <derek@knosof.co.uk> wrote:
> > haihao,
> >
> > > I am wondering whether there is any empirical data for random
> > configuration.
> >
> > Lots of empirical probability data is available in:
> > www.knosof.co.uk/cbook/usefigtab.pdf
> >
> >
> >
>