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

Re: [csmith-dev] Zero Checksum

On 02/09/11 15:52, John Regehr wrote:
Perhaps the larger problem here is that Csmith produces many programs
that are small and (worse) effectively identical.

My early driver scripts kept a hash of every tested program to avoid
testing duplicates, but eventually I realized it was simpler and more
effective to just throw away every program smaller than some size. We've
variously used numbers like 5 KB and 20 KB.

That's definitely an interesting idea. Thanks for the tip.
Did you use file size to determine if a test is big enough of source lines?

Also, is there any study on the size of programs generated by CSmith in general?

At some level it would be nice if Csmith just failed to produce these
programs, but on the other hand I strongly believe in keeping complexity
outside of Csmith when this is feasible (I'm not claiming we've done a
good job with this).

I agree that if there are things that can be easily performed by a postprocessor and kept outside CSmith, then better! :)


Member of the CSR plc group of companies. CSR plc registered in England and Wales, registered number 4187346, registered office Churchill House, Cambridge Business Park, Cowley Road, Cambridge, CB4 0WZ, United Kingdom
More information can be found at www.csr.com. Follow CSR on Twitter at http://twitter.com/CSR_PLC and read our blog at www.csr.com/blog