Home » Services » High Performance Computing Facility (HPCF)

HPC Consulting Services

The HPC Facility provides consulting Services to NYULMC biomedical researchers. These services function within the integrated Best Practices Informatics Consulting (BPIC) model of NYULMC Informatics Center. The services are designed to provide a "one-stop shop" solution whereby researchers receive grant preparation assistance, project design help, cost effectiveness analysis of locally and externally applicable HPC solutions, and finally project execution services.

  • Schedule an HPC Consultation

To schedule an HPC BPIC consultion, please contact the Technical Director of the HPC Facility efstratios.efstathiadis@nyumc.org or InformaticsConsultation@nyumc.org. Please provide any appropriate background information for your project or question in advance of your consultation. You can email materials to the above email address. In addition, please bring to the consultation any information that can facilitate your discussion with BPIC faculty.

Consultations over WebEx are also possible (WebEx is maintained locally by MCIT. Charges apply).

  • Walk-In HPC Clinic

In addition to scheduled HPC BPIC consulting, HPC consultations are available on a "first-come first-served" walk-in clinic the 1st and 3rd Wednesday of each month from 2:00 to 3:00PM at 227 East 30th Street (7th Floor), New York, NY 10016. More information on Walk-In Informatics Consulting Clinics can be found here.

List of Services

The list of HPC services includes:

  • Dedicated Access to Powerful Compute Nodes

A number of Compute Nodes on the HPC Asclepius Compute Cluster can be reserved for the exclusive usage of a group's members. Exclusive access to cluster nodes insures dedicated, uninterrupted access, without having to compete with other users for resources. Each cluster node comes with 12 Intel processing cores, 48 Giga-Bytes of memory and has access to a large amount of data storage. In addition, several of the nodes are equipped with Graphics Processing Units (GPUs), while one of the nodes comes with a lot of RAM (512 Giga-Bytes). The nodes are managed by a Resource Manager (Sun Grid Engine) to ensure resources are efficiently, evenly and fairly used. Installed software is customized to fit the needs of the project. Additional services, such as databases, Code Version (CVS/subversion), etc. are available to facilitate the development, troubleshooting and execution of applications. The resources are monitored closely and usage accounting reports are produced every month.

  • Data Storage

Datasets stored on our Highly Scalable, High Performance Data Storage Cluster are available over the network to a number of NYULMC clients (windows, Linux, Mac OS X) via NFS/CIFS, providing unique opportunities for improved access to information and faciltating the simultaneous analysis/visualization of data on several of the reseachers' desktop workstations. In addition, our data storage cluster is available on the HPC Asclepius Compute Cluster making the stored data accessible to a large number of very powerful compute nodes equiped with the latest scientific tools and libraries, enabling Data-Intensive and Data-Parallel Computing. We consult on optimizing/tuning access to data, benchmarking IO throughput and monitoring of IO performance. Stored data are backed up with the user having fully access to the recovery software. We are currently in the process of proving a Disaster Recovery solution by deploying an identical storage cluster on a remote location. We have currently deployed 200 TeraBytes (200,000 Giga-Bytes) of raw storage on a three node Isilon storage cluster. In addition, 40TeraBytes of RAID6 storage is available for staging (cache) data on the HPC cluster. 

  • Cloud Computing

Cloud Computing, being an instantiation of Utility Computing, provides on-demand, scalable computing over the internet on a "pay-as-you-use" model, based on virtualization. Among the benefits of the cloud is being elastic (increase and decrease capacity as needed), easy to deploy, and being cost effective.

Our cloud services include: Deployment of analysis workflows on individual servers or on a cluster environment on the cloud, implementing a Proof of Concept (PoC) environment comparing metrics (cost, time to results, etc.) versus using in-house resources, data sharing amongst distributed collaboration members, and implementation of MapReduce services for analyzing large datasets.

Click here for a Case Study.

  • Analysis Workflows
  • System Administration
  • Code Parallelization
  • GPU programming (list of libraries ported to GPUs, list of GPU nodes on Asclepius cluster)
  • Access to Supercomputers