Efficient parallel I/O in seismic imaging

Although high performance computers tend to be measured by their processor and communication speeds, the bottleneck for many large-scale applications is the I/O performance rather than the computational or communication performance. One such application is the processing of three-dimensional seismic data. Seismic data sets, consisting of recorded pressure waves, can be very large, sometimes more than a terabyte in size. Even if the computations can be performed in core, the time required to read the initial seismic data and velocity model and write images is substantial. In this paper, the authors discuss an approach in handling the massive I/O requirements of seismic processing and show the performance of their imaging code (Salvo) on the Intel Paragon(TM) computer.