Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watnot!watmath!clyde!rutgers!mit-eddie!uw-beaver!uw-june!entropy!dataio!bright
From: bright@dataio.UUCP
Newsgroups: comp.arch,comp.lang.c
Subject: Re: String Handling -- Incompetence of run-time libraries
Message-ID: <1281@dataio.Data-IO.COM>
Date: Tue, 31-Mar-87 13:38:34 EST
Article-I.D.: dataio.1281
Posted: Tue Mar 31 13:38:34 1987
Date-Received: Thu, 2-Apr-87 03:28:52 EST
References: <15292@amdcad.UUCP> <978@ames.UUCP> <15694@sun.uucp>
Reply-To: bright@dataio.UUCP (Walter Bright)
Organization: Data I/O - FutureNet Corp., Redmond, WA
Lines: 19
Keywords: instruction set architectures, strcpy
Xref: utgpu comp.arch:727 comp.lang.c:1401

In article <5@wb1.cs.cmu.edu] avie@wb1.cs.cmu.edu (Avadis Tevanian) writes:
] In article <1531@husc6.UUCP] reiter@harvard.UUCP (Ehud Reiter) writes:
]] As a brief addition to my previous posting, when I used my own strcpy and strcmp
]] routines, the Dhrystone rating of my microVAX II (cc -O, 4.3BSD) jumped from
]] 1000 (library strcpy, strcmp) to 1400 (my strcpy, strcmp).
] You've got to be careful when you run benchmarks.  If you are indeed running
] 4.3, then you are most likely using the 4.3 libc, which has been carefully
] optimized to use the fancy VAX instructions for the string routines.
] Unfortunately, some of these instructions are not implemented by the
] MicroVAX-II hardware.  As it turns out, what is happening is that your tests
] (including Dhrystone) are causing kernel traps to emulate those
] instructions!

The fix, of course, is:
	o Have the startup module determine what version of the CPU is running
	  and set a global.
	o Have the string functions use different instruction sequences,
	  depending on which CPU is being used.
This is similar to how the 8087 is used on the PC.