Compute Node Compute Node

Many scientiic applications that run on today's multiprocessors, such as weather forecasting and seismic analysis, are bottlenecked by their le-I/O needs. Even if the multiprocessor is conngured with suucient I/O hardware, the le-systemsoftware often fails to provide the available bandwidth to the application. Although libraries and enhanced le-system interfaces can make a signiicant improvement, we believe that fundamental changes are needed in the le-server software. We propose a new technique, disk-directed I/O, to allow the disk servers to determine the ow of data for maximum performance. Our simulations show that tremendous performance gains are possible both for simple reads and writes and for an out-of-core application. Indeed, our disk-directed I/O technique provided consistent high performance that was largely independent of data distribution, and obtained up to 93% of peak disk bandwidth. It was as much as 18 times faster than either a typical parallel le system, or a two-phase-I/O library.