Cray-2

The Cray-2 was a supercomputer designed and manufactured by Cray Research. It was the successor to the Cray X-MP and was notable for its innovative design, particularly its immersion in a liquid coolant (Fluorinert) and its large shared memory.

The Cray-2 was announced in 1985 and first delivered that same year. It featured four background processors sharing a large (for the time) 256 million words (2 gigabytes) of central memory. This memory was designed for very fast access, helping to compensate for the relatively slow (for Cray) clock speed of 4.1 nanoseconds (244 MHz). While the clock speed was slower than its predecessor, the Cray X-MP, the machine's overall performance was enhanced by its large memory bandwidth and the four processors.

The machine’s architecture heavily relied on a shared memory model. The four processors were connected to the central memory via a complex switching network. This allowed any processor to access any part of the memory with relatively uniform latency.

One of the defining characteristics of the Cray-2 was its compact physical design, achieved through dense packaging of the integrated circuits. This dense packaging necessitated the use of liquid immersion cooling to manage the heat generated by the electronics. The entire computer was submerged in Fluorinert, an inert dielectric fluid, which circulated through the machine to remove heat.

The Cray-2 was used in a variety of scientific and engineering applications, including weather forecasting, computational fluid dynamics, and nuclear research. Although technically powerful for its time, its performance was not always dramatically superior to the Cray X-MP in all application areas, partly due to the challenges of effectively utilizing the large shared memory and the relatively slower clock speed. The subsequent Cray Y-MP addressed many of the limitations and became a more commercially successful machine.