Dr. Prith Banerjee, senior VP of research at HP and director of HP Labs, discussed some promising areas for research while delivering his keynote on future research directions in EDA at the ongoing VLSID 2010.
He highlighted eight areas that HP Labs is currently working on. These are:
– Digital commercial print
– Content transformation
– Immersive interaction
– Information management
– Analytics
– Cloud
– Intelligent infrastructure
– Sustainability
EDA challenges
So, what are the EDA challenges? According to Dr. Banerjee:
* Today, EDA develops automated tools for designing ICs. However, there is a need to address automation for electronic systems at higher levels.
* There is a proliferation of new modes for communication and collaboration has resulted in the explosion of digital information.
* An intelligent IT infrastructure, which can deliver extremely high performance, adaptability and security — will be the backbone of these developments.
* In future, you need to look at design automation for entire systems.
– networks and data centers
– electronics and photonics
– performance and sustainability
Growth of new modes of communication and collaboration has led to an explosion of digital information. The IT industry would need to develop novel ways to acquire, store, process, and deliver information to customers. An intelligent IT infrastructure, which can deliver extremely high performance, adaptability and security, will be the backbone of these developments.
Intelligent infrastructure
This is required to capture more value via dramatic computing performance and cost improvement. HP Labs’ contribution has been on three big bet projects, namely, next generation data centers, networking and next generation scalable storage. HP Labs’ research contribution has been:
* Exascale
An Exascale data center that will provide 1000X performance while enhancing availability, manageability and reliability and reducing the power and cooling costs. HP Labs is working on the design of a sustainable data center that reduces total cost of operation (TCO) and carbon footprint, while meeting the current quality of service goals.
— Designed across components, interconnects, power and cooling, virtualization, management and software delivery.
* Photonics
– Replace copper with light to transmit data
* 1000X gain in performance
Dr. Banerjee said that there is a need to create brand new optical technologies that can work in exascale. Photonics interconnects make use of light for data communications. The transmit or receive optical bus is a simple modular system in four elements — transmitter media, optical tap, optical source, and optical receiver.
Non volatile memory and storage is another area that HP Labs is working on. Dr. Banerjee highlighted the memristor or a resistor with memory. In future, it has the potential to replace DRAM, hard drives and Flash memory.
“The memristor has the potential to revolutionize electronics,” he added. It is structurally simple and easy to fabricate. Also, it switches in nanoseconds and has many year lifetimes. HP Labs is said to be in discussions with leading memory makers for further.developing and licensing this technology.
Next-generation displays is yet another key research area at HP Labs. These are said to be unbreakable, conformable, ultra-thin and lightweight displays. Such displays have paper like qualities + video capabilities.
Technologies such as memristors, photonic interconnects, and sensors will likely revolutionize the way data is collected, stored and transmitted.
Sustainable data centers is another key research area. Among HP Labs’ research contribution, it is aiming to reduce TCO by 50 percent and carbon footprint by 75 percent.
Highlighting another industry challenge, Dr. Banerjee pointed out that the IT industry is only responsible for 2 percent of total carbon emissions. The global economny contributes the remaining 98 percent of total carbon emissions from other industries such as aviation, transportation, retail, etc. IT has a significant role to play here.
Naturally, all of the research areas would require sophisticated system-level design automation tools. Dr. Banerjee said: “In the past, the EDA research focused on chips. In the future, we need to look at entire systems.”
Thursday, January 7, 2010
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.