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

[csmith-dev] Floating-point support in csmith

Hi All,

First of all, I apologize for my limited knowledge about csmith. I'm interested to know level of support for floating-point in current csmith. Here are some of the facts I observed (again, I'm sorry as I'm new to csmith and might not know a lot of facts):

 - Floating-point support was introduced in version 2.2.0 according to the release note [1]. Floating point support must be turned on passing --float command-line option. This note acknowledges work of Dr. Alastair Donaldson for the initial implementation.

According to Dr. Donaldson's 2015 paper [2]: " Like Csmith, CLsmith does not generate test floating point programs. We view this as an exciting open  challenge:  floating  point  imprecision  is  tolerated  in  the  accelerator programming domain, but the fuzzing methods we study demand precise results. "

 - Current TO-DO file [4] in the GitHub repository lists "generate floating point math" as a TO-DO task. However, it is also mentioned that this file is not updated.

 - I conducted a run of csmith 2.2.0 with --float argument on Ubuntu 10.04 (gcc 4.4.3). Number of crash-programs is significantly more now. Running csmith without --float found 3 crash-bugs, while running with --float found more than 300 crash-bugs within the (almost) same runtime. 

I am interested to know current support-level for floating points, any limitations and any future works you plan regarding float-point support.

I'll also appreciate any source which explains current support level of floating-point in csmith.

Thanks for your time!

[1] https://github.com/csmith-project/csmith/commit/dcef523ba47761eaabb0119e80a7bb42ffd8cd81

[2] http://www.doc.ic.ac.uk/~afd/homepages/papers/pdfs/2015/PLDI_Fuzzing.pdf

[3] CLsmith: http://multicore.doc.ic.ac.uk/tools/CLsmith/PLDI15/

[4] https://github.com/csmith-project/csmith/blob/master/TODO

 - Shafiul