Global Sources
EE Times-India
 
EE Times-India > EDA/IP
 
 
EDA/IP  

Future research avenues in EDA

Posted: 04 Feb 2010     Print Version  Bookmark and Share

Keywords:EDA  research  VLSI  data centre  processor 

Banerjee: Automation is required for electronic systems at higher levels.

EDA vendors develop automated tools for designing ICs but there is a need to address automation for electronic systems at higher levels. Prith Banerjee, senior VP of research and director, HP Labs highlights a few such promising areas with live examples.

HPLabs' current high-impact research areas are: Digital Commercial Print, Intelligent Infrastructure, Content Transformation, Immersive Interaction, Information management, Analytics, Cloud and Sustainability.

The proliferation of new modes of communication and collaboration has resulted in an explosion of digital information. To turn this challenge into an opportunity, the IT industry will have to develop novel ways to acquire, store, process and deliver information to customers. Computers not only used for computing but also for other applications like web chat, picture sharing sites etc. In the next five years, more information will be created than ever in the history of this planet. Intelligent IP structure for the information created is required. Thus design automation is required in these areas. To address this, HP Labs is building an exascale data centre that will provide 1000X performance while enhancing availability, manageability and reliability. The date centre will be built on a cloud-scale, intelligent storage system that is massively scalable, resilient to failures, self-managed and enterprise grade.

Photonics in data centres
Data centre processors may currently have a dual-core or quad-core processor but these will grow to billion cores in the future. These are currently connected with storage blades or electronics with speeds not more than 10Gbps. Eventually the scale will be so huge that it will be unmanageable and optics will eventually replace them. To make optic connectors commercially viable, they need to be cost effective. Currently, a simple process involving two optical connectors can cost around Rs.2.31 lakh ($5000), hence new technology has to be developed for cost effective optic usage. The use of photonics to replace copper will result in 1000x gain in performance.

Further elaborating on the role of optics, Banerjee illustrated the same with an example where photonic interconnects can replace the backpane of servers — this is HP Labs' optical crossbar. Currently, eight cores on a chip are supported by copper. But, in the future a thousand cores cannot be supported. HP Labs has developed a corona architecture that is a ring-based topology within the chip, and they are partnering with chip manufacturers to extend this architecture.

Additionally, optical bus needs to be made of a partially transmittive and partially retentive material. HP Labs designed the same with a hollow metal wave, optical tap (reflector), optical source and optical receiver. Aligning of the components is very important and can prove to be an expensive process. Automation can help reduce cost in this area. By using plastic moulding, there is a significant reduction in cost; thereby ultimately enabling this optical bus to replace any bus in a data centre server.

Efforts are also underway to transform the way the commercial print industry delivers data—a transformation from physical to digital and back to physical where the management of data is critical which in turn will lessen the carbon footprint.


1 • 2 Next Page Last Page



Comment on "Future research avenues in EDA"
Comments:  
*  You can enter [0] more charecters.
*Verify code:
 
 
Webinars

Seminars

Visit Asia Webinars to learn about the latest in technology and get practical design tips.

 

Go to top             Connect on Facebook      Follow us on Twitter      Follow us on Orkut

 
Back to Top