2

How many floating point operations(add, mul) per cycle an ARMv6(ARM1176JZFS - Raspberry Pi model B) processor can execute? This is for get the teorical peak performance of HPC(High Performance Computing) cluster:

Node performance in GFlops = (CPU speed in GHz) x (number of CPU cores) x (CPU instruction per cycle) x (number of CPUs per node)

carma03
  • 21
  • 2
  • http://www.arm.com/products/processors/classic/arm11/arm1176.php – Marc B Jun 23 '15 at 21:13
  • 1
    didnt someone just ask this? is this a duplicate? – old_timer Jun 23 '15 at 21:14
  • @dwelch, yes, I asked that a few days ago [peak-flops-per-cycle-for-arm11-and-cortex-a7-cores-in-raspberry-pi-1-and-2](http://stackoverflow.com/questions/30976071/peak-flops-per-cycle-for-arm11-and-cortex-a7-cores-in-raspberry-pi-1-and-2). I guess other people are interested in this as well. – Z boson Jun 24 '15 at 08:14
  • I can't imagine any HPC cluster using ARMv6. It may be a cluster using many (thousands) of ARMVv6/raspberry Pi 1 but it won't be HPC. – Z boson Jun 24 '15 at 08:17
  • Of course Z boson, "don't imagine it". It won't be a production cluster, it's only for educational or prototype purpose. Coming soon the ARM processors will come to dominate the computers and supercomputers ;) . Salut – carma03 Jun 24 '15 at 16:27
  • 1
    @Zboson: ARMv6 no, but ARMv8+GPU - [yep!](http://www.montblanc-project.eu/press-corner/news/innovative-embedded-architecture-hpc-made-europe) :) – Hristo Iliev Jun 25 '15 at 17:32

0 Answers0