Re: Avoid treating the stack as an array [Re: "Back & Forth" is back!]

In article <90389fea385c08c72f39d4fdef04d076@www.novabbs.com>,
mhx <mhx@iae.nl> wrote:

On Sat, 14 Sep 2024 17:41:23 +0000, Ahmed wrote:
>

On Sat, 14 Sep 2024 17:13:51 +0000, Ahmed wrote:
>

utime 0.1e neg_big utime d- dnegate d.
with locals:    about 19 ms
without locals: about 18 ms
>
Ahmed

Oops.
>
Please read micro seconds (us) instead of milli seconds (ms).
>
with locals:    about 19 us
without locals: about 18 us

>
That can't be correct.
>
In iForth I used dfloats instead of floats
( 4.9ns instead of 7.3ns).
Using structs is not a great idea in this case.
>
anew -testlocals
>
: tri_mf:  ( f: a b c )
    create frot df, fswap df, df,
    does> ( F: x -- y )
                     ( ad_a)           ( f: x)
      dup fdup        ( ad_a ad_a)      ( f: x x)
      df@             ( ad_a)           ( f: x x a)
      f>=             ( ad_a -1|0)      ( f: x)
      over dfloat+    ( ad_a -1|0 ad_b) ( f: x)
      fdup df@        ( ad_a -1|0)      ( f: x x b)
      f< and if       ( ad_a)           ( f: x)
        dup df@ f-    ( ad_a)           ( f: x-a)
        dup df@       ( ad_a)           ( f: x-a a)
        dfloat+       ( ad_b)           ( f: x-a a)
        f@ fswap f-                     ( f: x-a b-a)
        f/                              ( f: [x-a]/[b-a])
        exit
      then
      dfloat+         ( ad_b)           ( f: x)
      dup fdup        ( ad_b ad_b)      ( f: x x)
      df@             ( ad_b)           ( f: x x b)
      f>=             ( ad_b -1|0)      ( f: x)
      over dfloat+    ( ad_b -1|0 ad_c) ( f: x)
      fdup df@        ( ad_b -1|0)      ( f: x x c)
      f< and if       ( ad_b)           ( f: x)
        dup dfloat+ df@ ( ad_b)         ( f: x c)
        f-            ( ad_b)           ( f: x-c)
        dup dfloat+   ( ad_b ad_c)      ( f: x-c)
        swap df@ df@ f-                 ( f: x-c b-c)
        f/                              ( f: [x-c]/[b-c])
        exit
      then
      drop fdrop
      0e
;
>
-1e309    -1e     0e         tri_mf: nol_neg_big
>
: (tri_mf) ( f: x a b c -- mv)
      FLOCALS| c b a x |
      x a f>= x b f< and if x a f- b a f- f/ exit then
      x b f>= x c f< and if c x f- c b f- f/ exit then
      0e ;
>
: loc_neg_big -1e309 -1e   0e   (tri_mf) ;
: .timing MS? S>F 1e-3 F*  1e7 F/ F.N2 ." s/call." ;
>
: tnb  CR ." \ no locals: " TIMER-RESET #10000000  ( 1e7 times )
         0 DO  -10e   nol_neg_big FDROP  LOOP .timing
      CR ." \    locals: " TIMER-RESET #10000000  ( 1e7 times )
         0 DO  -10e   loc_neg_big FDROP  LOOP .timing ;

This captures the meaning of the problem not good.
Anton Ertl is right that you have to bound a b c
into something, that is more than its parts.

0E0 FDUP FDUP class triangle-function
    M: a F@ M;   F,
    M: b F@ M;   F,
    M: c F@ M;   F,

    M: fx   ( f1 -- f1 )
      FDUP a f>= FDUP b f< and if a f- b a f- f/ exit then
      FDUP b f>= FDUP c f< and if c FSWAP  f- c b f- f/ exit then
      0e M;
endclass

5E0 3E0 1E0 triangle-function orang-utan

orang-utan
2E0 fx F.
4E0 fx F.

Note that I have not introduced anything special, only classes
that you need anyway. These classes are straightforward
generalisation of the CREATE DOES> construct,minus the
awkward syntax.
Note that x is passed as it should, volatile in Forth fashion.
Passing 4 parameters is c-style.

NOTE:
These are presentation of ideas, nothing is tested.

>
FORTH> tnb
\ no locals: 4.9ns/call.
\    locals: 21.3ns/call. ok
>
-marcel

--
Temu exploits Christians: (Disclaimer, only 10 apostles)
Last Supper Acrylic Suncatcher - 15Cm Round Stained Glass- Style Wall
Art For Home, Office And Garden Decor - Perfect For Windows, Bars,
And Gifts For Friends Family And Colleagues.