Intel has announced a new 32-core server chip based on a new high-performance computing server architecture that mixes general x86 cores with specialised cores for faster processing of highly parallel scientific and commercial applications.
The chip, called Knights Ferry, is Intel's fastest processor ever and delivers more than 500 gigaflops of performance, says Kirk Skaugen, vice president and general manager of Intel's data center group. Skaugen mentioned the new chip in a speech at the International Supercomputing Conference in Hamburg, Germany.
The chip's cores run at 1.2GHz. It is the first in a new family of server chips called Knights, which the company describes as being based on a new "many integrated core" architecture.
The 32-core chip will be available in the second half of 2010, for development purposes.
The first commercial product will include more than 50 cores and be called Knights Corner. An Intel spokesman would not say when that chip will be available. However, the chip will be part of the Sandy Bridge chip architecture, manufactured using the 22-nanometer process, and those processors are due to reach laptops and servers in 2011.
The initial 32-core chip is made using the existing 45-nm process.
Knights Ferry includes 32 main Xeon chip cores with corresponding 512-bit vector processing units, and sits in the PCI-Express slot. The chip runs four threads per core and includes 8MB of shared cache, and up to 2GB of fast GDDR5 memory.
The company will merge the CPU cores and vector units into a single unit as chip development continues, Skaugen said.
The Knights architecture is the biggest server architecture shift since Intel launched Xeon chips, Skaugen said. The chip includes elements of the Larrabee chip, was characterised as a highly parallel, multicore x86 processor designed for graphics and high-performance computing. However, Intel last week said it had cancelled Larrabee for the short term, but said elements of the chip would first be used in server processors, and and later in laptops.
The new architecture could also fend off competition from Nvidia's Tesla and AMD's FireStream graphics processors, which pack hundreds of computing cores to boost application performance. The graphics processors are faster at executing certain specialized applications. The second fastest supercomputer in the world, Nebulae in China, combines CPUs with GPUs to boost application performance.
The chip will accelerate highly parallel applications, Skaugen said. It could also standardise software development platforms around the x86 architecture, making it easier to recompile programs, Skaugen said.
Intel has many new server chips in the pipeline. Earlier this month, the chip maker said it would release a successor to its eight-core Nehalem-EX chips next year, with more cores and faster speeds. The new chips, code-named Westmere-EX, will be for servers with four or more sockets. The company is also developing an experimental 48-core x86 chip with a mesh design, but has not announced plans to sell it.
Intel already has a significant lead in the HPC market, but the new chip and surrounding architecture could extend its presence. According to the Top500 list, 408 supercomputers — more than 80 percent of the list — use Intel's chips, giving it a significant lead over other chip companies such as AMD and IBM.