Parallel I/O Benchmarks, Applications, Traces

Below is a list of parallel I/O benchmarks, applications, and traces I am aware of (in no particular order). Let me (Rajeev Thakur) know if you have any additions or corrections to the list.

Updated: November 1, 2021


  1. H5bench: A Parallel I/O Benchmark Suite for HDF-5

  2. DLIO Benchmark for Scientific Deep Learning Applications

  3. IOR and mdtest

  4. WRFIO

  5. MACSio Proxy App

  6. PnetCDF Benchmarks

  7. Performance Test in Parallel I/O (PIO) Library (see tests/performance directory)

  8. HACC I/O

  9. FLASH I/O
    An updated version for newer versions of parallel NetCDF and HDF-5 is available

  10. Effective I/O Bandwidth Benchmark (b_eff_io)

  11. NAS BTIO Benchmark (see BT in NPB3.3-MPI)

  12. MPI Tile I/O

  13. FS_TEST (see this as a README)

  14. Trace-Driven Scientific I/O Benchmarks

  15. HPIO and S3asim

  16. Noncontig

  17. ioperf test in Silo

  18. Pattern I/O Benchmark (John Shalf, LBNL)

  19. MADbench2

  20. QIO

  21. PIO-Bench
    An article describing it is at


  1. Earth System Modeling Framework (ESMF)

  2. Nek5000

  3. NekCEM

  4. PISM: Parallel Ice Sheet Model

  5. MOAB


  7. Cactus

  8. mpiBLAST

  9. FLASH
    Also see FLASH I/O Benchmark above.

  10. Community Earth System Model (CESM)

  11. Regional Ocean Modeling System (ROMS)

  12. Weather Research & Forecasting Model (WRF)

  13. CM1

  14. Qbox

  15. GADGET

  16. MILC

  17. MPQC

  18. NaSt3DGP

  19. SIESTA

  20. CP2K

  21. LAMMPS


  1. HPC Application I/O Traces

  2. Darshan

  3. Sandia

  4. MAMBO

  5. IO-SIG Software for I/O Signature