🚀 go-pugleaf

In article <1990Jan15.223217.6592@Solbourne.COM> nm@wcfields.Solbourne.COM (Nick Mati) writes:

>The Solbourne Series5 machine uses the Weitek 3171 FPU (Abacus) at 33.33 MHz. 
>The 3171 is similar to the 3170 except that it has a Cypress CY7C601 (IU) 
>interface wrapped around it instead of a Fuji/LSI Logic interface.  Running
>the same code, the performance of the 3171 is basically the same as
>theTI8847 which is used in the Sun 4/490, the only difference being
>the TI8847 requires    an extra interface chip to talk to the IU.
....

This is inconsistent, either the 3171 is like the 3170 or it isn't.

	3170          TI8847
fadds|d  5              8
fmults   5		8
fmultd   8	 	9
fdivs   38(?)          13
fdivd   66(?)          18
fsqrts                 15
fsqrtd		       22

My slide describing the 3170 has gotten a bit fuzzy (spilt coffee or
some such). I'm sure that the data sheets are publically available
from Weitek. The TI data is from the '89 Hot Chips Symposium TI
handout. My Weitek ref box only has the prelim sheet for the 3170/1
and lacks cycle counts (I suspect that the lit is around here
somewhere though... :>)

TI seems to have made an algorithmic advance for div/sqrt ... which
more than makes up for the 3 cycle lose for fadd,fmults, in most
applications that I've studied.

Cheers.

--
Keith H. Bierman    |*My thoughts are my own. !! kbierman@sun.com
It's Not My Fault   |	MTS --Only my work belongs to Sun* 
I Voted for Bill &  | Advanced Languages/Floating Point Group            
Opus                | "When the going gets Weird .. the Weird turn PRO"

"There is NO defense against the attack of the KILLER MICROS!"
			Eugene Brooks