Intel Demonstrates Teraflop Processor

Link: Intel Demonstrates Teraflop Processor
Filed under: Hardware News, Technology News

Intel demonstrated its 80-core processor in San Francisco last week and will present a paper on the research behind it during the International Solid State Circuits Conference in the city this week.

Intel’s super-fast chip can produce 1 trillion floating-point operations per second - a speed known as a teraflop. Only a decade ago it would have taken 2,500 square feet of large computers to produce this performance.

Intel revealed a prototype 80-core processor during the Intel Developer Forum in autumn 2006 and said it would deliver the chip within five years.

There is a long way to go before PCs and servers come with 80-core processors. Researchers still have to find a way to connect the chip to memory. Intel is developing a stacked memory chip that could be placed on top of the research chip. It is also collaborating with memory designers about next-generation designs for memory chips.

A major challenge will be writing software which can take advantage of the chip’s phenomenal speed.

Each of the 80 cores, or “tiles” on Intel’s chip has a computing element and a router, allowing it to process data individually and transport that data to neighboring tiles.

Intel used 100 million transistors on the chip, which measures 275 millimeters squared. Although the chip was built using Intel’s 65-nanometer manufacturing technology, a product based on the design would probably use a future process based on smaller transistors. A chip the size of the current research chip is considered too large to be manufactured cost-effectively.

The company expects products based on the chip to be available in 5 years time.