Boston, MA, 11/15/2013 (wallstreetpr) – Micron Technology, Inc. (NASDAQ:MU) has announced that it is planning to adopt the Hybrid Memory Cube (HMC) for petscale supercomputers which marks a huge step for the company. The new technology is planned for applications that require low energy, high bandwidth access to memory, which is an important requirement for supercomputers.

The company and Fujitsu which is a global leader in the supercomputing industry will both exhibit a display board which will feature HMC devices in a prototype which will be a next generation supercomputer at the Supercomputing ’13 Conference which will be held in Denver between 19-21 Novembers.

Supercomputing is an important technology which allows scientists and engineers to tackle complex simulations that drive research and development and helps them to find out important questions about the organization of our world. Tacking such difficult questions need excellent data movement capability. Using the new device to meet the ability of the supercomputer’s multicore processor architecture will allow the supercomputer to function efficiently.

Brian Shirley, vice president of Micron’s DRAM Solutions Group said that the designers and engineers at Fujitsu had recognized the potential of the HMC device early and it expected that the company would be able to help the latter company to grow its capabilities in the sector.

Yuji Oinaga, head of Fujitsu’s Next Generation Technical Computing Unit said that his company’s designers were very happy with the HMC design.

HMC which is the latest breakthrough uses the latest through-silicon vias (TSVs) that connect chips stacked on top of each other electrically. It gives 160 GB of memory bandwidth while using only 70% less energy per bit than the technologies, already existing thus reducing the clients total cost of ownership (TCO).

The device has won recognition from leaders in the industry as the long waited reply to meet the growing gap between the DRAM performance rate and the consumption rate of processor data.