EFFICIENT PARALLEL I O IN SEISMIC IMAGING

Although high performance computers tend to be measured by their proce ssor and communication speeds, the bottleneck for many large-scale app lications is the I/O performance rather than the computational or comm unication performance. One such application is the processing of three -dimensional seismic data. Seismic data sets, consisting of recorded p ressure waves, can be very large, sometimes more than a terabyte in si ze. Even if the computations can be performed in core, the time requir ed to read the initial seismic data and velocity model and write image s is substantial. In this paper, the authors discuss an approach in ha ndling the massive I/O requirements of seismic processing and show the performance of their imaging code (Salvo) on the Intel Paragon(TM) co mputer.