Path: utzoo!mnetor!uunet!lll-winken!lll-lcc!ames!pasteur!ucbvax!RAND-UNIX.ARPA!bridger%rcc
From: bridger%rcc@RAND-UNIX.ARPA (Bridger Mitchell)
Newsgroups: comp.os.cpm
Subject: Sieve benchmark for Z280
Message-ID: <8801301850.AA19041@newton.arpa>
Date: 30 Jan 88 18:50:52 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The Internet
Lines: 16

A year ago, curious about the speed of compilers vs. hand-optimized
Z80 code, I wrote JETPRIME.  It proved about 140% faster that
Turbo-Modula 2, which itself was much faster than Turbo Pascal, BDS-C,
and a number of 8088 compiled versions.  JETPRIME, on 4Mhz Ampro
Littleboard, was 4.1 sec for 10 iterations; TM2 was about 10.

JETPRIME.LBR or JPRIME.LBR is available on ZNode-Central,
408-432-0821.  It may also be on other boards, and simtel20.

I'd be curious to hear how it holds up on the 280.

My third-hand understanding is that the current z280 mask doesn't
correctly handle code (and data?) changes that occur in the
associative cache memory.  Self-modifying code is therefore
unreliable, and possibly there are problems with data values also.

--bridger