Marconi

  1. /
  2. Systems
  3. /
  4. Hardware
  5. /
  6. Marconi

MARCONI

MARCONI is the Tier-0 system, co-designed by CINECA and based on the Lenovo NeXtScale platform, that replaced the former IBM BG/Q system (FERMI) in June 2016.

MARCONI, based on the next-generation of the Intel® Xeon Phi™ product family alongside with Intel® Xeon® processor E5-2600 v4 product family, offers the scientific community a technologically advanced and energy-efficient high performance computing system.


The new system has been gradually upgraded, between June 2016 and 2020, according to a plan based on a series of updates:

  • A1: a preliminary system was into production from June 2016, based on Intel® Xeon® processor E5-2600 v4 product family (Broadwell) with the computational power of ~ 1Pflop/s. (closed in September 2018)
  • A2: by the end of 2016 a section was added, equipped with the next-generation of the Intel Xeon Phi product family (Knights Landing), based on a many-core architecture, enabling an overall configuration of about 250 thousand cores with the additional computational power of ~ 11Pflop/s. (closed in January 2020)
  • A3:  in August 2017, a third partition was added, based on Intel Xeon 8160 (SkyLake). 1.512 nodes at first, followed by about 800 more, a few months later. In Nov 2018 a further enhancement of 912 SKL nodes, replacing the A1 partition, have finally completed the system, reaching a total computational power of about 20Pflop/s

This supercomputer takes advantage of the new Intel® Omni-Path Architecture, which provides the high-performance interconnectivity required to efficiently scale the system’s thousands of servers. A high-performance Lenovo GSS storage subsystem, that integrates the IBM Spectrum Scale™ (GPFS) file system, is connected to the Intel Omni-Path Fabric and provides data storage capacity.

MARCONI

Model: Lenovo NeXtScale
Architecture: Intel OmniPath Cluster
Internal Network: Intel OmniPath
Disk Space: 17PB (raw) of local storage
Total Peak Performance: ~ 20 PFlop/s

MARCONI – A3 (SKL)

Nodes: 3188
Processors: 2 x 24-cores Intel Xeon 8160 (SkyLake) at 2.10 GHz
Cores: 48 cores/node, 72.576 + 38.016 cores in total
RAM: 196 GB/node
Peak Performance: ~8 PFlop/s

Software