Turing : Introduction

Turing is the massively parallel architecture of IDRIS. This is an IBM Blue Gene/Q machine which is particularly well-balanced (network, I/O, computing power and memory bandwidth) and has a remarkable energy efficiency (2.17 Gflops/W).

Hardware characteristics of Turing :

  • 6 racks, each containing 1,024 compute nodes and each node having 16 cores, for a total of 98,304 cores
  • Cumulated peak performance of 1.258 Pflop/s (42nd in the TOP500 worldwide ranking of November 2014)
  • 96 TiB of total memory (1 GiB per core)
  • GPFS parallel file system (WORKDIR) shared by Ada, Adapp and Turing, with a bandwidth of 50 GiB/s in write and in read.

The codes which run on Turing must have a sufficiently high degree of parallelism (ideally about 512 or more execution cores) and not use more than 1 GiB of memory per core. In production, up to 4 racks of resources may be accessed (or 65,536 cores). For the phases of development or code porting, specific classes are available (from 1 to 64 compute nodes). You may consult the batch limits for Turing at class structures.

For more information concerning the available software or librairies, the architecture, and the usage of the Turing machine resources, see here.