VAX (Was: Cromemco landmarks)

From: Allison <ajp166_at_bellatlantic.net>
Date: Mon Nov 12 07:20:40 2001

Carlos,

Both have the FPU. It was optional only in the sense that it was not
required for a minimal cpu.

Allison

-----Original Message-----
From: Carlos Murillo <cmurillo_at_emtelsa.multi.net.co>
To: classiccmp_at_classiccmp.org <classiccmp_at_classiccmp.org>
Date: Sunday, November 11, 2001 11:16 PM
Subject: Re: VAX (Was: Cromemco landmarks)


>At 03:02 AM 11/12/01 +0100, Iggy wrote:
>>Carlos Murillo skrev:
>>>Didn't the uVax II implement some of the original VAX instructions
>>>with emulation? I always wondered what the VUP rating would have
>>>been were they not emulated.
>>
>>in order to reduce the architecture to a single (integer) chip, only 175
of
>>the 304 instructions (and 6 of 14 native data types) were implemented
>(through
>>microcode), while the rest were emulated - this subset included 98% of
>>instructions in a typical program. The optional FPU implemented 70
>>instructions and 3 VAX data types, which was another 1.7% of VAX
>instructions.
>>All remaining VAX instructions were only used 0.2% of the time, and this
>>allowed MicroVAX designs to eventually exceed the speed of full VAX
>>implementations
>
>Aha! So, if I were doing numerical linear algebra in a uVaxII w/o the
optional
>FPU, I could expect performance to be badly hit with respect to the 0.9 VUP
>rating.
>
>(FYI, I always tend to benchmark machines using float performance,
>because that's the kind of thing that I do for a living. I look at integer
>performance only as something that has an impact on sparse linear algebra
>blocks, as opposed to dense system methods).
>
>How do I know if my Vaxstation 2000 has the optional FPU? What is
>its model number?
>
>What about the Vaxstation 4000/60? Does it have a built-in FPU?
>
>carlos.
>
>--------------------------------------------------------------
>Carlos E. Murillo-Sanchez carlos_murillo_at_nospammers.ieee.org
>
>
Received on Mon Nov 12 2001 - 07:20:40 GMT