Path: utzoo!utgpu!attcan!uunet!husc6!bu-cs!ngeow
From: ngeow@bu-cs.BU.EDU (Yee Ngeow)
Newsgroups: comp.sys.ibm.pc
Subject: Re: Is the XT dead?
Message-ID: <39315@bu-cs.BU.EDU>
Date: 1 Oct 89 18:50:43 GMT
References: <1989Sep27.104957.24581@cs.dal.ca> <340006@hpsgpa.HP.COM> <2576@netcom.UUCP> <39221@bu-cs.BU.EDU> <7719@cg-atla.UUCP>
Reply-To: ngeow@cs.bu.edu (Yee Ngeow)
Followup-To: comp.sys.ibm.pc
Organization: Boston University
Lines: 30

In article <7719@cg-atla.UUCP> fredex@cg-atla.UUCP (Fred Smith) writes:
>In article <39221@bu-cs.BU.EDU> ngeow@bucsd.bu.edu (Yee Ngeow) writes:
>>
>>Does anybody know if the V20 is compatible with 80186 instructions, using
>>the Turbo C Compiler options for 80186/286 instructions? 
>>
>I don't knwo about Turbo C, but I use the -G2 switch on Microsoft C
>all the time with my V20. -G2 causes MSC to generate 80286 code. I have
>built a number of programs this way, including STEVIE, MicroEmacs, etc., and
>they all work just fine. Therefore I would expect 80186 code to also work (but
>haven't actually tried it.).
>
>
>The only gotcha I know of is that if you are also using floating point anywhere
>in your program that the compiler attempts to generate 287 code.

Hmmm.. Since V-20 runs 286 instructions, does that mean it can use the
protected mode to run OS/2 also? Probably not, but I am really curious.

Furthur benchmark on my 10 MHz: 875 Dhrystone, 2705 chars/second for character
output, 10K math performance. Pretty good vs. unmodified 10 MHz XT with
735 Dhrysones, 1650 chars/second and 9K math performance. 

Overall speed is about 2.95 times 4.77 MHz XT with a 8088 :-)

Oh, forgot to say I furthur improved speed by decreasing the RAM refresh
rate. That gives aother 7% overall boost to the speed. All benchmarks were
ran using this utility.

Kwong