Story image

HPE and TACC build the world’s first solar-powered supercomputer

Visitors arriving at the Texas Advanced Computing Center at the University of Texas immediately notice large solar panels covering the parking lot. These panels power Hikari, the world’s first solar-powered supercomputer, which is currently calculating biology applications to help solve the Zika virus crisis.

Supercomputers have the power to transform science and innovations, but the amount of energy they require has been a limiting factor towards their expansion. Because traditional supercomputers consume vast amounts of electricity and produce a lot of heat, larger cooling facilities must be constructed to ensure proper operation. To overcome this challenge, the Texas Advanced Computing Center partnered with the New Energy and Industrial Technology Development Organization (NEDO), a Japanese government agency, NTT Facilities Inc. and HPE to develop Hikari.

Energy-Efficient Supercomputers

Hikari leverages the Apollo 8000 warm water cooling system that eliminates the need for Computer Room Air Handlers (CRAH) and, by using dry-coolers instead of chillers, significantly reduces the cooling energy that would normally be required. In addition, the Apollo 8000 system was designed to minimize the supporting data center infrastructure to a minimum, driving savings not only on the operating expenses but on the upfront capital expenses too.

Transmitting from the photovoltaic (PV) array high-voltage DC source straight to a high-voltage DC computer is truly a game changer.

The Data Center of the Future

As data centers are rapidly growing in size to meet data and computing demands, the amount of power and density required to run the operations has increased. Since we lose energy efficiency every time AC converts to DC, Dan Stanzione, executive director at the Texas Advanced Computing Center, says Hikari is a large-scale experiment to run a production-scale data center entirely on DC power. At the core of the project is the HPE 8000 server. The supercomputer runs on about 200,000 watts of DC power. Other components of Hikari include DC battery systems, DC air-conditioning, DC lighting, DC supply and solar panels.

Since the Hikari team launched the measurement phase at the end of August, more than 30% of the total power used by the supercomputer has been supplied by renewable energy sources.  During some daylight hours, the system was operating at 100% from renewable sources. The team is now working on measuring DC power and the amount of power saved from HDVC distribution alone.

Partnering for the Future of Supercomputers

Stanzione says Hewlett-Packard Enterprise has been a key partner on this project. Texas Advanced Computing Center selected the Apollo 8000 for the core of the Hikari because HPE was the only manufacturer that could deliver an HVDC-supplied cluster with the required size and speed.

Telco central offices are being transformed into the edge
Schneider Electric's Steven Carlini discusses how the old central offices are now ideal to enable the telco edge.
Pliops raises US$30m to ‘transform’ data centre infrastructure
With backing including the likes of Intel and SoftBank Ventures Asia, the company intends to expand its operations in the US, China, and Israel.
Bill advancing in Indiana to provide tax breaks for data centers
This week a bill has progressed to make life easier for data centers in Indiana and promote a digital economy.
Q&A: Aruba manager on imminent data centre challenges
Aruba's Alessandro Bruschini shares his thoughts on booming demand for data centres and the growing obstacles bolstered by regulation and energy efficiency requirements.
Interview: Next steps needed in data centre energy efficiency
SPIE UK's Peter Westwood shares what needs to be done to make the vision of a data-driven world more sustainable.
Gartner names ExtraHop leader in network performance monitoring
ExtraHop provides enterprise cyber analytics that deliver security and performance from the inside out.
APAC holds largest installed base of storage capacity
"The Global StorageSphere is large and diverse, encompassing many different storage technologies, and growing rapidly."
IBM opens up Watson to run on anything, anywhere
Big Blue has made Watson portable across any cloud with the goal to empower businesses to prevent vendor lock-in and start deploying AI wherever their data resides.