On Demand HPC Infrastructure

GHS provides traditional high performance computing (HPC) architectures on x86 infrastructure as well as GPU-accelerated computational systems.  High performance storage architectures combines flash, disk, tape, and cloud storage with IBM Spectrum Scale to deliver a unified single namespace that is higher performing, limitless in scale, and lower cost than traditional storage solutions.  GHS HPC systems allow research organizations to store their valuable data more intelligently and economically.

Full service

End-to-end managed HPC services from GHS include workflow-specific consulting (which HPC components are best suited to my workflows), up-front planning services (how do I submit my HPC jobs?), ongoing implementation services (how are my HPC jobs doing?), and ongoing management services (should I adjust my HPC jobs?)  Unlike generic as-a-service (aaS) cloud services providers, GHS offers real HPC expertise to help you get the most bang-for-buck out of your HPC resources.

Utility based pricing

Traditional HPC clusters are highly capital-intensive.  Hardware acquisition costs, ongoing datacenter real estate costs, power and cooling costs, administration expenses, etc.  There is a better way.  With GHS’ Utility Based Pricing it becomes very simple to forecast costs and set budgetary limits on usage to ensure you always know what your HPC jobs are going to cost and you never go over your budget.  Pay for the resources you use with no hidden costs or extra fees.

• Compute, Network and Storage 
• Spectrum Scale experts. We help you implement Scale on any HPC storage platform
• Access to more than 58.5 PF/s on demand and over 148PB of online HPC storage, scratch and HSM

• Planning
• Implementation
• Management

• Only pay for what you use
• Subscription models