An Open Repository for Real-Time Benchmarks

FAQ

What is the difference between this benchmark and others

EEMBC is one of the earliest efforts towards developing performance benchmarks for use in embedded systems. It organizes the test suites by specific focus of embedded systems hardware and software development, for application domains including automotive, consumer electronics, and telecommunications, or to address specific design concerns such as energy and floating point performance. MiBench, collected in the early 2000s, follows the model of EEMBC by dividing a set of 35 applications (available as C code) into 6 categories, including automotive the test suites by specific focus of embedded systems hardware and software development, for application domains including automotive, consumer electronics, and telecommunications, or to address specific design concerns such as energy and floating point performance. MiBench, collected in the early 2000s, follows the model of EEMBC by dividing a set of 35 applications (available as C code) into 6 categories, including automotive and industrial control, consumer devices, office automation, networking, security, and telecommunications. For a specific purpose of evaluating WCET analysis methods and tools, the M¨alardalen WCET research group maintains a benchmark, containing 35 programs (provided in C source files) collected from several different research groups and tool vendors around the world. However, these benchmarks have their limitations. EEMBC is not freely available, access requires a membership with the associated cost. MiBench (besides its out-ofdate) and M¨alardalen benchmark are collections of programs rather than entire systems with a defined task structure.

How to download the benchmark

Please click this.

How to use the benchmark

How to contribute

Please click this.
Contact info: chuansheng.dong@mail.mcgill.ca