High performance computing and data centers

High performance computing and data centers

High performance computing

As computational efficiency of processor and co-processors increases beyond the linear, we at Orange Silicon Valley always strive to find the most efficient solution for the end users of AI and other high performance computing (HPC) workloads by prototyping systems in collaboration with our partners in the ecosystem.

Whether it is to maximize the speed of deep neural network training or to enable groundbreaking performance for seismic data analysis, the goal for us the remains the same: deliver rapid prototypes with the highest ROI for our end users while sourcing commercial off the shelf (COTS) components.

We are also looking at the next evolution of HPC infrastructure — software defined and manageable via programmable data center fabric to maximize the utilization of computational assets for large data center HPC deployments.

Data centers

With AI and Big Data Analytics driving the needs for computation and storage, there are needs for increased cooling efficiency for data centers which can be operated sustainably and deployed rapidly. With a ruggedized version, we can deploy supercomputing solutions for outdoors edge application under harsh environments.

We are working on high density data center designs for both air cooled and liquid immersion cooled systems. As industry is demanding lower PUE in air cooled systems to meet sustainability goals, with liquid immersion cooling we envision dramatically (10x) lower PUE with the possibility of a circular economy for energy via repurposing the heat rejected by the HPC data center.

Related work by Orange Silicon Valley

Lead analyst

Soumik Sinharoy
Principal, Technology Group

Please contact Soumik to explore opportunities for partnership involving high performance computing and data centers.

See our Who’s who for more information on all the members of our talented team.

Orange Silicon Valley posts on high performance computing and data centers