Contact us Heritage collections Image license terms
HOME ACL ACD ICF SUS DCS G&A STARLINK Literature
Further reading □ Overview □ 33. Start of year □ 34. Hardware □ 35. Communications □ 36. UNIX □ 37. ACCENT UNIX □ 38. Dalkeith closure □ 39. User Support □ 40. Software □ 41. Assessment □ 42. SUSSG □ 43. PERQ - DAP □ 44. PERQ orders □ 45. Critique of 1983
C&A INF CCD CISD Archives Contact us Heritage archives Image license terms

Search

   
⟶ ACD ⟶ Single User Systems ⟶ PERQ History ⟶ Part VII
⟶ ACD ⟶ Single User Systems ⟶ PERQ History ⟶ Part VII
ACL ACD C&A INF CCD CISD Archives
Further reading ⇓

□ Overview
□ 33. Start of year
□ 34. Hardware
□ 35. Communications
□ 36. UNIX
□ 37. ACCENT UNIX
□ 38. Dalkeith closure
□ 39. User Support
□ 40. Software
□ 41. Assessment
□ 42. SUSSG
□ 43. PERQ - DAP
□ 44. PERQ orders
□ 45. Critique of 1983

1983

37. ACCENT-UNIX

As ACCENT UNIX was still the long-term proposal for an operating system on the PERQ, the first quarter of 1983 was spent by RAL assessing its performance and pinpointing problems.

CMU had by now signed a contract with IBM to provide a distributed personal computing network within the university. The decision was made to base it on 68000 processors and the SUN workstation was chosen as the vehicle. Time was spent at the start of 1983 assessing how easily Accent and UNIX could be ported to the SUN.

By February 1983, Three Rivers had decided to back out of any new operating system development with ICL but to concentrate on a version of UNIX based on the Accent work done at RAL. Thus Three Rivers never marketed PNX in the USA.

SERC's evaluation of ACCENT UNIX was slow as about 50% of the people with expertise were helping ICL with utilities and compiler developments.

The major points that came out of the assessment were:

  1. ACCENT UNIX was twice the performance of PNX on all but the procedure call overhead which was between 10 and 500 times worse. Thus there were significant performance advantages in compilation and program execution but UNIX commands were very slow.
  2. No one item was to blame in the procedure call overhead. The RAL approach was sound and to improve the performance would require changes to ACCENT as well as some redesign of the SERC code. A major problem was the high cost of message passing in ACCENT. Either this had to be reduced or a redesign undertaken to reduce the number of messages sent.

By this time, progress on PNX was such that the decision was made to document ACCENT UNIX and freeze it. Program load time had been reduced to two seconds but no further improvement was obtained before it was frozen. Although a failure in as far as it never became a product, a small team had made a major software development in a short period of time. The amount of time taken by a much larger ICL team to achieve the same results confirms this assessment. It was unfortunate that the basic speed of ACCENT at that time proved insufficient to create a viable product.

⇑ Top of page
© Chilton Computing and UKRI Science and Technology Facilities Council webmaster@chilton-computing.org.uk
Our thanks to UKRI Science and Technology Facilities Council for hosting this site