Difference between revisions of "Introduction"
| (6 intermediate revisions by one other user not shown) | |||
| Line 1: | Line 1: | ||
| + | [[ZeptoOS_Documentation|Top]] | [[Configuration]] | ||
| + | ---- | ||
| + | |||
ZeptoOS is a research project investigating operating systems and run-time systems and libraries for petascale architectures with 10,000 to 1 million CPUs. The ZeptoOS project is a collaboration between Argonne National Laboratory and the University of Oregon. | ZeptoOS is a research project investigating operating systems and run-time systems and libraries for petascale architectures with 10,000 to 1 million CPUs. The ZeptoOS project is a collaboration between Argonne National Laboratory and the University of Oregon. | ||
| − | For IBM | + | For IBM Blue Gene/P platform, we are investigating highly efficient Linux-based compute node and I/O node kernels and advanced-functionality libraries and tools that support efficient collective operations, parallel I/O, memory management, and performance monitoring tools. Most of the time, to accomplish our computer science research goals we must replace IBM's limited-functionality components with versions that we have either written from scratch or adapted from other open source projects. The new components are fast, small, reliable, infinitely flexible, and with lots of new features. |
| − | [ | + | ---- |
| + | [[ZeptoOS_Documentation|Top]] | [[Configuration]] | ||
Latest revision as of 15:01, 5 May 2009
ZeptoOS is a research project investigating operating systems and run-time systems and libraries for petascale architectures with 10,000 to 1 million CPUs. The ZeptoOS project is a collaboration between Argonne National Laboratory and the University of Oregon.
For IBM Blue Gene/P platform, we are investigating highly efficient Linux-based compute node and I/O node kernels and advanced-functionality libraries and tools that support efficient collective operations, parallel I/O, memory management, and performance monitoring tools. Most of the time, to accomplish our computer science research goals we must replace IBM's limited-functionality components with versions that we have either written from scratch or adapted from other open source projects. The new components are fast, small, reliable, infinitely flexible, and with lots of new features.
