[comp.sys.mac.misc] INIT Question: "SGIL Fast Math"

barvian@ece.cmu.edu (Scott Barvian) (04/28/91)

I recently downloaded the SGIL init from the umich archives.  It's supposed
to trap calls to Apple's "SANE" floating point ROM routines and speed things
up.  They claim that it works in general, but to *really* make things fly
you need to compile your software with their code library, bypassing
SANE entirely.

?? Any experiences?  Any evidence that it really helps?  If it works as
   advertised, it's one more reason to keep my Plus a bit longer.

--
==== Scott Barvian ============ Department of Electrical/Computer Eng. ===
==== barvian@ece.cmu.edu ====== Carnegie Mellon Univ., Pittsburgh, PA  ===

bskendig@dew.Princeton.EDU (Brian Kendig) (04/28/91)

In article <BARVIAN.91Apr27222148@dart.ece.cmu.edu> barvian@ece.cmu.edu (Scott Barvian) writes:
>I recently downloaded the SGIL init from the umich archives.  It's supposed
>to trap calls to Apple's "SANE" floating point ROM routines and speed things
>up.  They claim that it works in general, but to *really* make things fly
>you need to compile your software with their code library, bypassing
>SANE entirely.
>
>?? Any experiences?  Any evidence that it really helps?  If it works as
>   advertised, it's one more reason to keep my Plus a bit longer.

Yes!  Definitely keep it around!  On my Mac SE, unaccelerated, under
System 7.0b4, The "Speedometer" application tells me that the math
processing is almost four times faster than on a stock SE under 6.0,
when I have the SGIL INIT in my control panels folder.

I have "Use SGIL 32 bits" selected.

It would be kinda hard to compile, say, Excel with their code library,
but I imagine that if Microsoft did so (fat chance) their software
would speed up.  The code library is only for people who are
developing their own math-intensive applications.

SGIL definitely makes my machine feel peppy under 7.0, and that's a
feat... ;)

     << Brian >>

| Brian S. Kendig      \ Macintosh |   Engineering,   | bskendig             |
| Computer Engineering |\ Thought  |  USS Enterprise  | @phoenix.Princeton.EDU
| Princeton University |_\ Police  | -= NCC-1701-D =- | @PUCC.BITNET         |
"You gave your life to become the person you are right now.  Was it worth it?"

jh@ist.CO.UK (Jeremy Huxtable) (04/30/91)

From article <8811@idunno.Princeton.EDU>, by bskendig@dew.Princeton.EDU (Brian Kendig):
> In article <BARVIAN.91Apr27222148@dart.ece.cmu.edu> barvian@ece.cmu.edu (Scott Barvian) writes:
>>I recently downloaded the SGIL init from the umich archives.  It's supposed
>>to trap calls to Apple's "SANE" floating point ROM routines and speed things
>>up.  They claim that it works in general, but to *really* make things fly
>>you need to compile your software with their code library, bypassing
>>SANE entirely.
>>

I tried using a similar device called "Radius Math", but found some problems
with it. In particular, it was causing the "cos" function in THINK C 4.0 to
return values like -2.7! I never got round to testing it with "real"
applications like Excel or anything so I don't know how general the problem
was.

carter@cat23.cs.wisc.edu (Gregory Carter) (05/06/91)

In article <BARVIAN.91Apr27222148@dart.ece.cmu.edu> barvian@ece.cmu.edu (Scott Barvian) writes:
>I recently downloaded the SGIL init from the umich archives.  It's supposed
>to trap calls to Apple's "SANE" floating point ROM routines and speed things
>up.  They claim that it works in general, but to *really* make things fly
>you need to compile your software with their code library, bypassing
>SANE entirely.
>
>?? Any experiences?  Any evidence that it really helps?  If it works as
>   advertised, it's one more reason to keep my Plus a bit longer.
>
>--
>==== Scott Barvian ============ Department of Electrical/Computer Eng. ===
>==== barvian@ece.cmu.edu ====== Carnegie Mellon Univ., Pittsburgh, PA  ===

I know it really screwsup PM 3.0 when I use it.

Of course I am using this on my MEGA STE, I am not sure if this is do
to the emulator or if its a real problem.  The problem I usually have with
it is that it screws up PM's 3.01 pull down menu's.

--Gregory

-----------------------------------------------------------------------------
MEGA STE: 4megs, ST157N HD 750K/sec, built in Appletalk, 16Mhz 68000 $990!  
Anytime I show my friends my MAC, I show them a floppy disk and say
"The MAC is here on this little disk $1.50"
Mac Emulation Version 3.0, the ultimate in Mac emulation 99.999999% compatible
------------------------------------------------------------------------------