Effective File I/O Bandwidth Benchmark
The effective I/O bandwidth benchmark (b{_}eff{_}io) covers two goals: (1) to achieve a characteristic average number for the I/O bandwidth achievable with parallel MPI-I/O applications, and (2) to get detailed information about several access patterns and buffer lengths. The benchmark examines ''first write'', ''rewrite'' and ''read'' access, strided (individual and shared pointers) and segmented collective patterns on one file per application and non-collective access to one file per process. The number of parallel accessing processes is also varied and well-formed I/O is compared with non-well formed. On systems, meeting the rule that the total memory can be written to disk in 10 minutes, the benchmark should not need more than 15 minutes for a first pass of all patterns. The benchmark is designed analogously to the effective bandwidth benchmark for message passing (b{_}eff) that characterizes the message passing capabilities of a system in a few minutes. First results of the b{_}eff{_}io benchmark are given for IBM SP and Cray T3E systems and compared with existing benchmarks based on parallel Posix-I/O.
- Research Organization:
- Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
- Sponsoring Organization:
- USDOE Office of Defense Programs (DP) (US)
- DOE Contract Number:
- W-7405-Eng-48
- OSTI ID:
- 792810
- Report Number(s):
- UCRL-JC-137586; TRN: US200302%%506
- Resource Relation:
- Journal Volume: 1900; Conference: Euro-Par 2000, Muchen (DE), 08/29/2000--09/01/2000; Other Information: PBD: 15 Feb 2000
- Country of Publication:
- United States
- Language:
- English
Input/output characteristics of scalable parallel applications
|
January 1995 | |
On implementing MPI-IO portably and with high performance
|
January 1999 | |
A performance study of two-phase I/O
|
January 1998 |
Similar Records
Effective Communication and File-I/O Bandwidth Benchmarks
HPC Global File System Performance Analysis Using A Scientific-Application Derived Benchmark