good@fortune.UUCP (Roy Good) (08/03/85)
In late '84, Jon Retsky published a very detailed Motorola Application Note on how to use a platform board to replace a 68000 by a 68020 plus some PALs. His benchmarks showed without fail that running the 68020 with cache disabled, at up to 10MHz., invariably gave inferior performance over the original 68000 based system. With cache enabled, however, performance was significantly improved in all benchmarks except two, namely "Bit test-set-clr" and the Berkeley Benche string concatenation. Has anyone tried this concept on a UNIX^tm (AT&T etc) based system. If so what performance conclusions were reached? It is assumed that no hardware changes would be made to the existing 68000 system, other than to install this platform adapter. We'd be interested to hear of your experiences. -- +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Roy J. Good, Fortune Systems Corp., 101 Twin Dolphin Drive, Redwood Shores, CA 94065 The contents of this message do not necessarily reflect the views of Fortune Systems, or anyone else's for that matter. +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++