What question did this study set out to answer?

To establish a communication link between OMERO and high-performance computing resources for improved segmentation.

April 25, 2026Open Access

OMERO-to-HPC Segmentation Connector

Key Points

To establish a communication link between OMERO and high-performance computing resources for improved segmentation.
Utilized computational artifacts on UltraViolet HPC system with AMD and Intel processors,
Incorporated GPUs for enhanced processing power,
Designed for scalable data movement and storage management.
Leveraged A100 GPU nodes to accelerate segmentation processes,
Achieved efficient data transfer with 503Gi RAM and high-speed networking standards,
Ensured robust performance using SLURM scheduler for task management.

Abstract

OMERO-to-HPC Segmentation Connector, Computational Artifacts 𝐴1 UltraViolet (HPC system) login node: 2 x Intel(R) Xeon(R) Gold 6148 CPU, 376Gi of RAM 𝐴2 UltraViolet Scheduler: slurm 25.05.3, HA, SchedulerType: backfill 𝐴3 UltraViolet Scheduler Type: backfill 𝐴4 UltraViolet compute node: A100 nodes : 4 x a100 GPU cards, 2 x Intel(R) Xeon(R) Gold 6342 CPU @ 2.80GHz, 503Gi RAM, /tmp on 6 TB of NVME. 500 GB HDD. 𝐴5 UltraViolet data mover node: 2 x Intel(R) Xeon(R) Gold 6148 CPU, 376Gi of RAM 𝐴6 UltraViolet Container: Singularity version 3.9.8 𝐴7 UltraViolet Networking standard: Infiniban, HDR, Switch Mellanox QM8790, Interface: Mellanox Technologies MT28908 Family ConnectX-6 𝐴8 UltraViolet operating system: Red Hat Enterprise Linux 8.8 𝐴9 UltraViolet storage scale: IBM E7300C, 111 GB RAM, Red Hat Enterprise Linux 9.4 𝐴10 OMERO production server 1 specs: 79 GB RAM, 16 vCPUs, Red Hat Enterprise Linux 9.3 𝐴11 OMERO production server 2 specs: 32 GB RAM, 16 vCPUs, Red Hat Enterprise Linux 9.3 𝐴12 OMERO storage scale server: IBM MTM 5149-091, System 7.0, 1.91 PB

Read Full Paperexternally

Bookmark

View Full Paper

Cite This Study

Adam Rubens (Thu,) studied this question.

synapsesocial.com/papers/69ec5b6088ba6daa22dace60 https://doi.org/https://doi.org/10.5281/zenodo.19713708

Bookmark

View Full Paper