Search Results

You are looking at 1 - 2 of 2 items for

  • Author or Editor: Patrick H. Worley x
  • All content x
Clear All Modify Search
Ian T. Foster, Brian Toonen, and Patrick H. Worley

Abstract

Massively parallel processing (MPP) computer systems use high-speed interconnection networks to link hundreds or thousands of RISC microprocessors. With each microprocessor having a peak performance of 100 or more megaflops per second, there is at least the possibility of achieving very high performance. However, the question of exactly how to achieve this performance remains unanswered. MPP systems and vector multi-processors require very different coding styles. Different MPP systems have widely varying architectures and performance characteristics. For most problems, a range of different parallel implementations is possible, again with varying performance characteristics. In this paper, we provide a detailed evaluation of MPP performance for a spectral transform kernel as used in weather and climate modeling applications. Using a specially designed spectral transform code, the authors study performance on three different MPP systems: Intel Paragon, IBM SP2, and Cray T3D. Great care is taken to tune the implementation for efficient execution on each platform. The results yield insights into MPP performance characteristics, parallel spectral transform algorithms, and coding style for MPP systems. The authors conclude that it is possible to construct parallel models that achieve multigigaflops-per-second performance on a range of MPPS, if the models are constructed to allow compile- or run-time selection of some parallel implementation options.

Full access
Peter R. Gent, Gokhan Danabasoglu, Leo J. Donner, Marika M. Holland, Elizabeth C. Hunke, Steve R. Jayne, David M. Lawrence, Richard B. Neale, Philip J. Rasch, Mariana Vertenstein, Patrick H. Worley, Zong-Liang Yang, and Minghua Zhang

Abstract

The fourth version of the Community Climate System Model (CCSM4) was recently completed and released to the climate community. This paper describes developments to all CCSM components, and documents fully coupled preindustrial control runs compared to the previous version, CCSM3. Using the standard atmosphere and land resolution of 1° results in the sea surface temperature biases in the major upwelling regions being comparable to the 1.4°-resolution CCSM3. Two changes to the deep convection scheme in the atmosphere component result in CCSM4 producing El Niño–Southern Oscillation variability with a much more realistic frequency distribution than in CCSM3, although the amplitude is too large compared to observations. These changes also improve the Madden–Julian oscillation and the frequency distribution of tropical precipitation. A new overflow parameterization in the ocean component leads to an improved simulation of the Gulf Stream path and the North Atlantic Ocean meridional overturning circulation. Changes to the CCSM4 land component lead to a much improved annual cycle of water storage, especially in the tropics. The CCSM4 sea ice component uses much more realistic albedos than CCSM3, and for several reasons the Arctic sea ice concentration is improved in CCSM4. An ensemble of twentieth-century simulations produces a good match to the observed September Arctic sea ice extent from 1979 to 2005. The CCSM4 ensemble mean increase in globally averaged surface temperature between 1850 and 2005 is larger than the observed increase by about 0.4°C. This is consistent with the fact that CCSM4 does not include a representation of the indirect effects of aerosols, although other factors may come into play. The CCSM4 still has significant biases, such as the mean precipitation distribution in the tropical Pacific Ocean, too much low cloud in the Arctic, and the latitudinal distributions of shortwave and longwave cloud forcings.

Full access