by Anton Shilov
05/24/2006 | 03:31 PM
Intel Corp. has recently disclosed certain peculiarities of its quad-core microprocessors for server computers. Apparently, the chips code-named Clovertown central processing units for 2-way enterprise-class machine will be available in three thermal envelopes: typical, low-voltage and “performance-optimized”.
Trevor Lawless, a representative for Intel’s server product group, demonstrated in
An Intel’s slide demonstrated to journalists claimed that Clovertown “delivers significant increase in processing capabilities on Bensley and Glidewell” platforms, which are designed for servers and workstations, respectively. Even though, it is currently unknown, whether the quad-core chips are drop-in compatible with servers based on the aforementioned platforms which ship today.
Intel’s Clovertown processor for volume servers will have 80W thermal design power, low-voltage incarnation of the product will sport 40W TDP, whereas “performance optimized” flavour will consume more than 120W, according to Intel.
Earlier the company demonstrated a sample of the quad-core server IDF Taipei 2006. The chip worked at 2.0GHz, whereas dual-core server processors code-named Woodcrest, which are due to be out in a few weeks from now, will be able to operate at up to 3.33GHz, significantly higher speed bin compared to the current quad-core prototype.
Intel’s code-named Clovertown chip is the company’s first microprocessor to feature four execution engines, or cores. The product is designed as two separate 65nm chips on a single piece of substrate and is intended for dual-socket servers. Intel is also working on Tigerton, quad-core processor for multi-processor servers. Additionally, the company is designing quad-core chips for desktops and notebooks.