Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watnot!watmath!clyde!rutgers!ames!ucbcad!ucbvax!NRL-PPD.ARPA!mitch%vc.DECnet
From: mitch%vc.DECnet@NRL-PPD.ARPA.UUCP
Newsgroups: comp.sys.atari.st
Subject: Savage Benchmarks
Message-ID: <8704020356.AA25070@ucbvax.Berkeley.EDU>
Date: Wed, 1-Apr-87 21:36:00 EST
Article-I.D.: ucbvax.8704020356.AA25070
Posted: Wed Apr  1 21:36:00 1987
Date-Received: Sat, 4-Apr-87 10:45:23 EST
Sender: daemon@ucbvax.BERKELEY.EDU
Reply-To: "VC::MITCH" <mitch%vc.decnet@nrl-ppd.ARPA>
Distribution: world
Organization: The ARPA Internet
Lines: 16

Here are five quick additions to the Savage Benchmark list:

System         CPU / FPP    CLOCK        LANGUAGE          TIME     ERROR 
		   	    (MHz)                           (Sec)  Abs(a-2500)
Atari 520ST  (68000/-----)   8.00   GFA Basic Interpreter   15.7    3.7 E-05
Atari 520ST  (68000/-----)   8.00   ProFortran V1.1(single) 16.0    2.7 E+02
Atari 520ST  (68000/-----)   8.00   Absoft F77 V2.2(single) 24.0    3.9 E+02
Atari 520ST  (68000/-----)   8.00   ProFortran V1.1         52.0    3.1 E-07
Atari 520ST  (68000/-----)   8.00   Absoft F77 V2.2         70.0    1.6 E-07

The only surprising one is the GFA Basic interpreter, which has an unusual
6-byte accuracy for reals.  It seems that the 6-byte accuracy is far better
than the 4-byte results, which are unacceptable.  As a matter of fact, the
GFA Basic interpreter is marginally faster and 1.0e-7 more accurate than
the single precision results for both Absoft Fortran and ProFortran.
------