Enhancing Computational Capabilities of the TeraGrid

德克萨斯大学奥斯汀分校将部署和支持一个强大的HPC系统，为广泛的科学应用提供支持。该系统是一个302 TF戴尔群集，具有44 TB内存和1.2 PB磁盘。这种平衡的系统具有5倍的峰值性能，也将提供2倍的？真实的？通过比当前系统大得多的内存和互连带宽，在科学应用上提高每核性能。它还提供了4倍的总内存和10倍的磁盘容量，并具有可扩展的路径来增加更多的计算和存储容量。 该系统将为当前的TeraGrid分配过程提供7500万小时的计算机时间，并通过TACC分配过程提供至少1500万小时的额外时间。 因此，大约50%的系统将用于NSF同行评审的科学研究。

The University of Texas at Austin will deploy and support a powerful HPC system that will provide support for a broad range of scientific applications. The system is a 302 TF Dell cluster with 44 TB memory and 1.2 PB disk. This balanced system has 5x the peak performance, and will also deliver 2x the ?real? performance per-core on science applications through much greater memory and interconnect bandwidth than the current system. It also offers 4x the total memory and 10x the disk capacity, with a scalable path to adding more computational and storage capacity. The system will provide 75 million hours of computer time to the current TeraGrid allocation process and an additional minimum of 15 million hours through the TACC allocation process. Thus approximately 50% of the system will go to NSF peer reviewed scientific research.