Skip to content

Bianca hardware

Technical Summary

  • 148 (ordinary/thin) compute nodes with
    • single or dual CPUs with 256 GB RAM and
    • one 4TB mechanical drive or 1 TB SSD
    • each CPU has 16 cores
  • 10 fat compute nodes
    • 6 nodes with 512 GB memory each
    • 4 nodes with 1 TB memory each
  • 27 GPU nodes
    • 9 GPU compute nodes each equipped with 2xNVIDIA A100 (40 GB) GPUs
    • 17 GPU compute nodes each equipped with 1xNVIDIA T4 (16 GB) GPUs
  • 35 Zen4 type compute nodes, 768 GB each
    • 105 virtual nodes รก 3 256 GB each
  • Total number of CPU cores is 4800
  • Login nodes have 4vCPU each and 50 GB memory
  • Dual 10 Gigabit Ethernet for all nodes

Parameters

Parameter Bianca
Purpose Sensitive data
Reserved for NAISS-SENS projects
Nodes (Intel) 272 + 9 nodes of 2 NVIDIA A100 GPUs + 17 nodes of NVIDIA T4 GPUs
Nodes (AMD) 45 (3 virtual nodes per physical node)
Cores per node 16
Memory per node 256GB
Fat nodes 512 & 1 TB
Local disk (scratch) 4TB
Network Dual 10Gbit/s
Operating System CentOS 7
Login nodes Yes (4 cores and 50 GB)
"Home" storage Cygnus
"Project" Storage Cygnus

CPU

Intel nodes (4th generation)

  • Architecture: x86_64
  • Intel Xeon E5-2630 v3 Huawei XH620 V3 nodes
  • Advanced Vector Extensions 2 (AVX2)
  • CPU op-mode(s): 32-bit, 64-bit
  • Byte Order: Little Endian
  • CPU(s): 16
  • Thread(s) per core: 1
  • Core(s) per socket: 8
  • Socket(s): 2
  • NUMA node(s): 2
  • Model name: Intel Core Processor (Haswell, no TSX, IBRS)
  • CPU MHz: 2.4 GHz
  • For more info, type: lscpu in the terminal

AMD nodes (Zen4)

Zen4 type compute nodes

Settings on Bianca (virtualisation)

  • Physical nodes divided into 3 sections, each with:
    • 16 cores 256 GB RAM

GPUs

  • 10 compute nodes each equipped with 2xNVIDIA A100 (40 GB) GPUs
  • 17 compute nodes each equipped with 1xNVIDIA T4 (16 GB) GPU

Network

Dual 10 Gigabit Ethernet for all nodes

Storage

Cygnus

Security

Since Bianca is designed to handle sensitive personal data security is a key aspect of the configuration. In order to ensure that the data is safe we have implemented a series of security measures including, but not limited to:

  • One virtualised cluster per project, no resources are shared between projects.
  • Separate storage volumes per project.
  • Detailed logging of file transfers in and out of the cluster.
  • Two factor authentication
  • No internet access inside the clusters.
  • Locked racks for the hardware
  • Destruction of broken hard drives

Uppsala University has decided on the following KRT classifications for Bianca:

  • 321 for project directories
  • 322 for home directories