Mercury Computer Systems and IBM said Tuesday that they would integrate the Cell microprocessor technology into computer systems for data-intensive applications. Mercury is the first company outside of the consumer industry to use Cell microprocessor technology design services from IBM.

“The tremendous performance advantages afforded by the Cell processor will enable Mercury to address an even broader range of compute-intensive challenges for our customers,” said Jay Bertelli, president and chief executive officer, Mercury Computer Systems.

Mercury has initial plans to integrate the Cell microprocessor technology into a wide variety of future products, with the aim of boosting computing performance in Mercury’s customer applications. By incorporating Cell technology, Mercury intends to take demanding applications such as radar, sonar, MRI, CT, digital X-ray, and others to a new level of sophistication and performance, the company said.

Mercury and engineers from IBM’s Engineering and Technology Services unit will collaborate to develop products with improved performance for graphic-intensive workloads and computationally intensive applications. The optimized Cell-based products will be targeted at current and new applications in medical imaging, industrial inspection, aerospace and defense, seismic processing, and telecommunications. New sensor capabilities in these markets are dramatically increasing the volumes of available data to be processed. For example, applying the processing power of Cell technology in medical imaging can yield significant improvements in image quality - enabling earlier detection of diseases and potentially saving lives.

The prototype Cell microprocessor produced using 90nm process technology incorporates one dual-threaded PowerPC core and eight so-called synergistic processing units (SPEs) intended for floating-point calculations, the most demanding tasks in entertainment, workstation and server systems. The PowerPC core is projected to have 32KB L1 cache and 512KB L2 cache, while each of the SPEs will have 256KB of cache.

The processor code-named Cell has built-in Rambus XDR memory interface, capable of data rates of 3.20GHz to 8.0GHz. The chip also uses FlexIO processor buses, formerly codenamed Redwood that are capable of running up to 6.40GHz data rates providing bandwidth more than four times faster than best-of-class processor buses available today. Sony and IBM claim their Cell processor will be able to run at speeds exceeding 4.0GHz, even at 90nm process technology, while Intel Corp. is expected to push its dual-core central processing units to 3.20GHz this year.

The Cell processor has peak performance in excess of 200GFLOPS - which equates to 200 billion floating-point operations per second – as measured during initial hardware testing.