Started exploring options in 1966

1967: Harwell IBM 360/75, Rutherford IBM 360/65

CDC 7600

• Visited CDC, 1967, for pre-launch presentation of CDC 7600 (20 times Atlas power)
• DTI keen to support ICT/ICL so ICL 1906A-CDC 7600 proposed (22 times Atlas)

P51

• DTI supports ICT's new multi-processor system called the P51 later the 1908A, 7-10 times Atlas per processor
• ICL initially proposed a triple-processor ICL 1908A (30 times Atlas in speed) expandable to a 4-processor system
• Later, ICL propose ICL 1906A-1908A, then add second 1908A (22 times Atlas)
• 1968: Laboratory order ICL 1906A - ICL 1908A (12 times Atlas)
• New very large computer block designed to house 1906A-1908A
• 1969: ICL abandon ICL 1908A leaving ACL with replacement machine just twice Atlas in speed
• 1971: New smaller computer block erected

CDC STAR-100

• Propose CDC Star-100 vector processor front-ended by an ICL 1906A (102 times Atlas in speed)
• Benchmarked Star-1 using mainly Quantum chemistry applications

Finally:

• July 1971: ICL 1906A delivered
• November 1971: Rutherford IBM 360/195 delivered (10 times Atlas power)
• 1906A + 20% 360/195 to provide a large-scale computer service just to SRC Grant Holders