raphaeltfool / iosig Goto Github PK

IOSIG software is a toolkit developed by Scalable Computing Software Lab at Illinois Institute of Technology. In our research work, we designed and developed several performance optimization techniques for parallel I/O systems. Many of the techniques is application-aware that can utilize the I/O characteristics information of the applications to improve the efficiency of I/O system . We use IOSIG to collect and analyze applications I/O characteristics, which works as follows:

1. IOSIG trace collector records the trace of all the I/O operations of the application.
2. The IOSIG trace analyzer analyzes the trace files and recognizes the data access patterns.
3. The analysis results can be displayed using the browser-based dashboard.

See INSTALL.md for how to install and use the software. For more examples on how to use IOSIG to profile parallel I/O benchmarks and applications, check the GitHub project: https://github.com/yinyanlong/profiling_parallel_io.

Contact:

• Yanlong Yin ([email protected])
• Xian-He Sun ([email protected])