small clean up

math32.tal

@@ -111,17 +111,16 @@
     COMPLEMENT32 RTN
 
 ( temporary registers )
-( used by all operations except div32 )
-@sh [
-  &r $1
-  &x0 $1 &x1 $1 &x2 $1 &x3 $1
-  &y0 $1 &y1 $1 &y2 $1 &y3 $1
-  &z0 $1 &z1 $1 &z2 $1 &z3 $1
-  &a0 $1 &a1 $1 &a2 $2
-]
+( used by most operations, except mul32 and div32 )
+@sh [ &r $1
+      &x0 $1 &x1 $1 &x2 $1 &x3 $1
+      &y0 $1 &y1 $1 &y2 $1 &y3 $1
+      &z0 $1 &z1 $1 &z2 $1 &z3 $1
+      &a0 $1 &a1 $1 &a2 $2 ]
 
 ( bit shifting )
 
+( shift right, i.e. >> )
 @right-shift ( x** n^ -> x<<n )
     DUP #08 LTH ;right-shift0 JCN2 ( x n )
     DUP #10 LTH ;right-shift1 JCN2 ( x n )
@@ -168,6 +167,7 @@
     ;sh/r LDA SFT
     RTN
 
+( shift left, i.e. << )
 @left-shift ( x** n^ -> x<<n )
     DUP #08 LTH ;left-shift0 JCN2 ( x n )
     DUP #10 LTH ;left-shift1 JCN2 ( x n )
@@ -235,6 +235,7 @@
 
 ( arithmetic )
 
+( addition, i.e. + )
 @add32 ( xhi* xlo* yhi* ylo* -> zhi* zlo* )
     ;sh/y2 STA2 ;sh/y0 STA2 ( save ylo, yhi )
     ;sh/x2 STA2 ;sh/x0 STA2 ( save xlo, xhi )
@@ -259,6 +260,7 @@
     ;sh/z0 LDA2 ;sh/z2 LDA2
     RTN
 
+( negation, i.e. unary - )
 @negate32 ( x** -> -x** )
     COMPLEMENT32
     INC2 ( ~xhi -xlo )
@@ -268,9 +270,11 @@
     &done
     RTN
 
+( subtraction, i.e. binary - )
 @sub32 ( x** y** -> z** )
     ;negate32 JSR2 ;add32 JSR2 RTN
 
+( 16-bit multiplication )
 @mul16 ( x* y* -> z** )
     ;sh/y1 STA ;sh/y0 STA ( save ylo, yhi )
     ;sh/x1 STA ;sh/x0 STA ( save xlo, xhi )
@@ -295,22 +299,24 @@
     ;add32 JSR2
     RTN
 
-@mul32 ( x** y** -> z** )
-    ;sh/y2 STA2 ;sh/y0 STA2 ( save ylo, yhi )
-    ;sh/x2 STA2 ;sh/x0 STA2 ( save xlo, xhi )
-    ;sh/y2 LDA2 ;sh/x2 LDA2 ;mul16 JSR2 ( [x2*y2] )
-    ;sh/z2 STA2 ;sh/z0 STA2 ( sum = x2*y2, save zlo, zhi )
+( multiplication, i.e. * )
+@mul32 ( x** y** -> z** ) 
+    ,&y1 STR2 ,&y0 STR2 ( save ylo, yhi )
+    ,&x1 STR2 ,&x0 STR2 ( save xlo, xhi )
+    ,&y1 LDR2 ,&x1 LDR2 ;mul16 JSR2 ( [x1*y1] )
+    ,&z1 STR2 ,&z0 STR2 ( sum = x1*y1, save zlo, zhi )
 
-    ;sh/y2 LDA2 ;sh/x0 LDA2 MUL2 ( [x0*y2]<<16 )
-    ;sh/y0 LDA2 ;sh/x2 LDA2 MUL2 ( [x2*y0]<<16 )
+    ,&y1 LDR2 ,&x0 LDR2 MUL2 ( [x0*y1]<<16 )
+    ,&y0 LDR2 ,&x1 LDR2 MUL2 ( [x1*y0]<<16 )
     ( [x0*y0]<<32 will completely overflow )
-    ADD2 ;sh/z0 LDA2 ADD2 ( sum += x0*y2<<16 + x2*y0<<16 )
-    ;sh/z2 LDA2
+    ADD2 ,&z0 LDR2 ADD2 ( sum += x0*y1<<16 + x1*y0<<16 )
+    ,&z1 LDR2
     RTN
-[ &x0 $2 &x2 $2 ]
-[ &y0 $2 &y2 $2 ]
-[ &z0 $2 &z2 $2 ]
+[ &x0 $2 &x1 $2
+  &y0 $2 &y1 $2
+  &z0 $2 &z1 $2 ]
 
+( division, i.e. / )
 @div32 ( x** y** -> q** )
     ( store y and x for repeated use )
     ;div32/div1 STA2 ;div32/div0 STA2 ( y -> div )

tester.py

@@ -56,7 +56,7 @@ def bitcount(x):
     return floor(log(x, 2)) + 1
         
 def main():
-    trials = 1000
+    trials = 100
     run(['uxnasm', 'test-math32.tal', 'run.rom'])
     p = pipe()
     test(p, trials, b'+', [('x', u32), ('y', u32)], u32, lambda x, y: x + y)