Hardware Utilization

  2. /
  3. Our Activities
  4. /
  5. Services
  6. /
  7. Service Statistics
  8. /
  9. Hardware Utilization

Our Activities

HPC for science
HPC for Industry
HPC for Meteo
HPC for Bioinformatics
HPC for Cultural Heritage
Computer Graphics
Quantum Computing
Data Science
Digital Twins
High-level Support
User Support
Service Statistics

Statistic Cineca Hardware Utilization

In the year 2022, the Cineca HPC infrastructure was composed of four clusters: the Tier-0 systems Marconi and Marconi 100, the Tier-1 system Galileo 100 and the DGX system, an accelerated system for AI based on NVIDIA A100 GPUs.

Marconi 100 opened the way to the accelerated pre-exascale Leonardo supercomputer. In July 2023, Marconi 100 ends its production phase.

PRACE, ISCRA and ICEI are the three grant allocation frameworks with diffrenet calls per years. In addition computational resources are allocated through AGREEMENTS, as support to EU projects (PROJECT EU) or assigned to Industrial users (CONTRACTS) .

CINECA CLUSTERS UTILIZATION 2022/2023

In the following graphs, we show the general usage of the computing nodes (expressed in percentage) of the Cineca HPC systems in the last year (November 2022- November 2023)

 

Galileo100

Figure 1: Galileo100utilization from November 2022 until November 2023.

Marconi

Figure 2: Marconi utilization from November 2022 until November 2023.

 

Marconi100


Figure 3: Marconi100 utilization from November 2022 until August 2023. At the end of June 2023, Marconi100 ended its production phase.

CLUSTER ALLOCATION IN 2022

A significant part of the allocations on Cineca’s system are based on a peer review selection to ensure the highest scientific value of the selected projects. Italian and European researchers take advantage of the HPC clusters in Cineca for performing their computational research, applying for “computational projects”, and allowing access to the HPC facilities.

The following shows the allocation of the CINCECA infrastructure via call PRACE, ISCRA, ICEI and agreement and contracts/project EU in January – December 2022.
The systems have been overallocated at 117,9% due to the scientific community’s enormous request for computing resources.


Figure 5: Core Hours (expressed in percentage) allocated from January to December 2022.

 Marconi Allocation

The following shows the allocadet hours fo Marconi A3. The data refers only to the shared + test partition of 278 SKL nodes and is used almost exclusively by 4 conventions.

Figure 6: Marconi A3 Core Hours (expressed in percentage) allocated from January to December 2022.

PROJECTS DOMAINS

The following shows the utilization per scientific domain in 2022.

 

MENU