(405b) Many Integrated Core and the Coming Stampede: Intel Responds to the GPU



Stampede About Stampede:

The Texas Advanced Computing Center (TACC) at The University of Texas at Austin will deploy and support a world-class supercomputer, called Stampede, with comprehensive computing and visualization capabilities as part of a National Science Foundation grant.

When deployed in 2013, the new system, built by TACC in partnership with Dell and Intel, will be the most powerful system in the NSF's eXtreme Digital (XD) program, and will support the nation's scientists in addressing the most challenging scientific and engineering problems over four years.

Stampede will have a peak performance of 10 petaflops, with 272 terabytes of total memory, and 14 petabytes of disk storage. Compute nodes:

  • Several thousand Dell Stallion servers, with each server having dual 8-core processorsfrom the forthcoming Intel® Xeon® Processor E5 Family (formerly codenamed Sandy Bridge-EP) and 32 gigabytes of memory
  • A total of 2 petaflops of performance using the Xeon processors.

Innovative capabilities:

  • The cluster will also include a special pre-release shipment of Intel® Many Integrated Core (MIC) architecture based co-processors connected to the compute nodes.
  • The Intel MIC processors will use the popular x86 instruction set and are designed to process highly parallel workloads.
  • The MIC processors will provide an additional 8 petaflops of performance.
  • Future generations of Intel MIC co-processors will be added when they become available and are expected to increase Stampede's aggregate peak performance to at least 15 petaflops.

Visualization and Data Analysis:

  • Stampede will offer 128 next-generation NVIDIA graphics processing units (GPUs) for remote visualization.
  • 16 Dell servers with 1 terabyte of shared memory and 2 GPUs each for large data analysis.
  • A high-performance Lustre file system for data-intensive computing.

System integration:

  • All components—compute nodes, visualization nodes, large shared memory nodes, and file system—will be integrated with an FDR 56 Gb/s InfiniBand network for extreme scalability.
  • Grid Engine software will be used to submit batch jobs, and tools for submitting large ensembles of jobs will be provided.

For further information on Stampede, please see our press release or watch a video with TACC director, Jay Boisseau, describing the new system.