This is a collection of programs containing floating-point arithmetic with the immediate goal of aiding the development and evaluation of floating-point precision and correctness analysis tools. The long-term goal is to support and enable the scaling of such tools to provide insights for high-performance computing (HPC) applications. This collection aims to address the "middle ground" between real-valued expressions (such as those contained in the FPBench collection) and full HPC applications.
The benchmarks themselves are located in the bench folder and they can be
built using the standard make command. See the benchmark folders for more info
about each individual benchmark.
Because this suite is intended to provide test cases for precision and
correctness analysis tools, one goal of the suite is to provide correctness
criteria for the benchmarks. For the benchmarks where these routines have been
written, they can be run using the make test command.
Some of the benchmarks are not directly hosted in this repository in order to avoid licensing issues. Those benchmarks are acquired in a reproducible way (often by downloading or cloning a particular version) as specified by the provided Makefiles.
Currently, the only targetted and supported platform is x86-64 Linux.
This software effort is still in early stages of development, and as such contributions and feedback are very welcome! Please use the Issues tab to submit bug reports or see a list of what still needs to be done.
The code in this repository is distributed under the MIT license.