A parallel spectral model for atmospheric transport processes

Abstract

The paper describes a parallel implementation of a grand challenge problem: global atmospheric modeling. The novel contributions of our work include (1) a detailed investigation of opportunities for parallelism in atmospheric global modeling based on spectral solution methods, (2) the experimental evaluation of overheads arising from load imbalances and data movement for alternative parallelization methods, and (3) the development of a parallel code that can be monitored and steered interactively based on output data visualizations and animations of program functionality or performance. Code parallelization takes advantage of the relative independence of computations at different levels in the earth's atmosphere, resulting in parallelism of up to 40 processors, each independently performing computations for different atmospheric levels and requiring few communications between different levels across model time steps. Next, additional parallelism is attained within each level by taking advantage of the natural parallelism offered by the spectral computations being performed (e.g. taking advantage of independently computable terms in equations). Performance measurements are performed on a 64‐node KSR2 supercomputer. However, the parallel code has been ported to several shared memory parallel machines, including SGI multiprocessors, and has also been ported to distributed memory platforms like the IBM SP‐2.

A middleware toolkit for client-initiated service specialization

“…One use for this capability is remote data reduction. For example, consider event channels used for monitoring of scientific calculations, such as the global climate model described in [14]. Further, consider a client that may be interested in some property of the monitored data, such as an average value over the range of data.…”

“…This can be accomplished by deriving a channel using a function which performs the appropriate data reduction. For example, the E-Code function defined in Figure 5 performs such an average over atmospheric data generated by the atmospheric simulation described in [14], thereby reducing the amount of data to be transmitted by nearly four orders of magnitude.…”

ACDS: Adapting computational data streams for high performance

“…The sample application used in our research is a global atmospheric climate model [8]. The data streams of principal interest to this paper link the running model and/or data stored from previous model runs to visualizations employed by end users.…”

Evaluation of Rate-Based Adaptivity in Asynchronous Data Stream Joins

“…The window size can be fixed and static or can vary across streams; it can be a system-wide parameter set at startup, or can be specified by the user on a query-by-query basis by an extension to the query language. Our experience in applying a prototype continuous query system to several application domains (i.e., safety critical robot control [18], a global atmospheric transport model [10], and now severe storm forecasting) has led us to conclude that in order to work in a realistic setting, the join window size must be configurable by the user, and allowed to vary in size from stream to stream. User configurable is key because only the user has sufficient understanding of the latency and asynchronism of application-specific streams.…”