Contact us Heritage collections Image license terms
HOME ACL ACD ICF SUS DCS G&A STARLINK Literature
Further reading □ OverviewStaffingHead of ICFDEC10sChilton SystemUpgraded MinisMUM benchmarkingPDP11
C&A INF CCD CISD Archives Contact us Heritage archives Image license terms

Search

   
ACDICFPhase 1
ACDICFPhase 1
ACL ACD C&A INF CCD CISD Archives
Further reading

Overview
Staffing
Head of ICF
DEC10s
Chilton System
Upgraded Minis
MUM benchmarking
PDP11

Benchmarking Multi-User Mini Systems

The major activity in the first half of 1976 was upgrading existing systems and benchmarking for the Central Chilton system. On the new multi-user mini systems front, a specification was developed, based on the Rosenbrock Report recommendations, and after contacting many manufacturers, a total of 26 possible suppliers were found that could meet the hardware specification.

However, the multi-user software requirements were more difficult to meet and this cut the original list to under 10. A formal Operational Requirement was produced and this was used in producing a shortlist.

The design of a benchmark was a problem. The important measurements concerned the response times for the six interactive users which the minis had to support. However, unlike on the mainframes, no automatic means of running such a benchmark was available. The solution was to devise a synthetic benchmark and get real users to run through a well-defined script.

Multi-User Benchmark in Operation

Multi-User Benchmark in Operation
Full image ⇗
© UKRI Science and Technology Facilities Council

Here we see the benchmark being performed in earnest. The five terminals in the foreground are being run from the left by Graham Robinson, Eric Thomas, David Duce, Len Ford and Dave Toll. In the background to the right sits Bob Hopgood (facing away) on a Tektronix 4014. Dale Sutcliffe and Richard Hilken are standing at the back supervising the run.

1 2 3 4 5 6 USER 0 2 4 6 8 10 12 14 16 TIME - MINS Press EDITA RUNA COMPILE RUNB EDITB RUNC

A Very Good Benchmark Run

Some more information on the benchmark is given in the paper by Irene Buchanan and David Duce. Surprisingly consistent results were achieved across several runs. Each user edited two files, ran three jobs and did one compilation in a precise order. Each user started a different task in the cycle of six. On completion, each user started the cycle again until the last person had completed one cycle. That was taken as the time for one iteration of the benchmark. The benchmark was expected to complete in 30 minutes.

Some manufacturers declined to tender before getting to the benchmarking phase. A Prime 300 and Prime 400 were benchmarked at the Prime UK Headquarters in Bedford. Both passed, the Prime 400 with flying colours. A PDP11/70 was benchmarked at the DEC Offices in Reading on a Saturday morning. Although the results were extremely good the first time, it was discovered that the system was caching results and reusing them rather than performing the task again. It performed less well when benchmarked normally. DEC declined to benchmark the machine using Unix. An Interdata 8/32 at their Uxbridge headquarters gave the most unusual result. The second user hit the compilation phase before the first use completed it. The system thrashed between the two jobs. Eventually the other users all entered the compilation phase and got stuck. We all went to lunch and on return the machine was still attempting to perform the compilations. A GEC 4080 and a GEC 4070 passed the benchmarks. The GEC 4070 fitted into the price constraints and ran just inside the maximum time for the benchmark of 30 minutes. The GEC 4070 time was 27 minutes while the Prime 400 completed the task in 17 minutes.

The decision was made to purchase a Prime 400 and a GEC 4070 as the two evaluation systems with the companies having to repeat the benchmark within the time of 30 minutes on the delivered system as part of the acceptance tests.

GEC 4070 Benchmarking as part of the Acceptance Test, 14 April 1977

GEC 4070 Benchmarking as part of the Acceptance Test, 14 April 1977
Full image ⇗
© UKRI Science and Technology Facilities Council
⇑ 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