diff --git a/math32.tal b/math32.tal
index 17744e4..523bd77 100644
--- a/math32.tal
+++ b/math32.tal
@@ -22,6 +22,8 @@
 
 ( program )
 |0100
+(    #1234 #5678 ;emit-long JSR2 NEWLINE
+    #0000 #0009 #0000 #0003 ;div32 JSR2 ;emit-long JSR2 NEWLINE )
     ;test-interact .Console/vector DEO2 BRK
 
 ( bitcount: number of bits needed to represent number )
@@ -339,37 +341,55 @@ RTN
 [ &y0 $2 &y1 $2 ]
 [ &z0 $2 &z1 $2 ]
 
+@div32 ( x** y** -> q** )
+    ( store y and x for repeated use )
+    ;div32/div1 STA2 ;div32/div0 STA2 ( y -> div )
+    ;div32/rem1 STA2 ;div32/rem0 STA2 ( x -> rem )
 
-@divmod32-by-32 ( x** y** -> q** r** )
-    ,&div1 STR2 ,&div0 STR2 ( y -> div )
-    ,&rem1 STR2 ,&rem0 STR2 ( x -> rem )
-    #0000 ,&quo1 STR2  #0000 ,&quo0 STR2 ( 0 -> quo )
-    ,&rem0 LDR2 ,&rem1 LDR2 ;bitcount32 JSR2 ( rbits^ )
-    ,&div0 LDR2 ,&div1 LDR2 ;bitcount32 JSR2 ( rbits^ dbits^ )
-    SUBk ,&shift STR ( rbits-dbits -> shift )
-    ,&div0 LDR2 ,&div1 LDR2 ,&shift LDR ;left-shift JSR2 ( div<<shift )
-    ,&div1 STR2 ,&div0 STR2
+    ( if x < y then the answer is 0 )
+    ;div32/rem0 LDA2 ;div32/rem1 LDA2
+    ;div32/div0 LDA2 ;div32/div1 LDA2
+    ;lt32 JSR2 ,&is-zero JCN ,&not-zero JMP
+    &is-zero
+    #0000 #0000 RTN
+
+    ( x >= y so the answer is >= 1 )
+    &not-zero
+    #0000 ;div32/quo0 STA2 #0000 ;div32/quo1 STA2 ( 0 -> quo )
+
+    ( bitcount[x] - bitcount[y] determines the largest multiple of y to try )
+    ;div32/rem0 LDA2 ;div32/rem1 LDA2 ;bitcount32 JSR2 ( rbits^ )
+    ;div32/div0 LDA2 ;div32/div1 LDA2 ;bitcount32 JSR2 ( rbits^ dbits^ )
+    SUB ( shift=rbits-dits )
+    #00 DUP2 ( shift 0 shift 0 )
+
+    ( 1<<shift -> cur )
+    #0000 #0001 ROT2 POP
+    ;left-shift JSR2 ;div32/cur1 STA2 ;div32/cur0 STA2
+    
+    ( div<<shift -> div )
+    ;div32/div0 LDA2 ;div32/div1 LDA2 ROT2 POP
+    ;left-shift JSR2 ;div32/div1 STA2 ;div32/div0 STA2 
+
+     &loop
+    ( if rem >= the current divisor, we can subtract it and add to quotient )
+    ,&rem0 LDR2 ,&rem1 LDR2 ,&div0 LDR2 ,&div1 LDR2 ;lt32 JSR2 ( rem<div? )
+    ,&rem-lt JCN ( if rem < div skip this iteration )
+
+    ( if rem >= div, then we have found a multiple of y that divides x )
+    ,&rem0 LDR2 ,&rem1 LDR2 ,&div0 LDR2 ,&div1 LDR2 ;sub32 JSR2 ,&rem1 STR2 ,&rem0 STR2 ( rem -= div )
+    ,&quo0 LDR2 ,&quo1 LDR2 ,&cur0 LDR2 ,&cur1 LDR2 ;add32 JSR2 ,&quo1 STR2 ,&quo0 STR2 ( quo += cur )
+
+    &rem-lt
+    ,&div0 LDR2 ,&div1 LDR2 #01 ;right-shift JSR2 ,&div1 STR2 ,&div0 STR2 ( div >>= 1 )
+    ,&cur0 LDR2 ,&cur1 LDR2 #01 ;right-shift JSR2 ,&cur1 STR2 ,&cur0 STR2 ( cur >>= 1 )
+    ,&cur0 LDR2 ,&cur1 LDR2 ;non-zero32 JSR2 ,&loop JCN ( if cur>0, loop. else we're done )
+    ,&quo0 LDR2 ,&quo1 LDR2 ( TODO: consider making this divmod32 )
 RTN
 [ &div0 $2 &div1 $2
   &rem0 $2 &rem1 $2
   &quo0 $2 &quo1 $2
-  &shift $1 ]
-
-@divmod16-by-8 ( x* y^ -> q* r^ )
-    DUP ,&y STR
-    ROT SWP ( x1 x0 y )
-    DIVk ( x1 x0 y q0 )
-    DUP ,&q0 STR ( x1 x0 y q0 )
-    MUL SUB ( x1 r=x0-y*q0 )
-    SWP #00 ,&y LDR ( r x1 00 y )
-    DIV2k ( rx1 00y q1 )
-    DUP2 ROT2 MUL2 ( rx1 q1 y*q1 )
-    ROT2 SWP2 SUB2 ( q1 rr=rx1-y*q1 )
-    SWP POP ( q1 rrlo )
-    ROT POP ( q1lo rrlo )
-    ,&q0 LDR TOR
-RTN
-[ &y $1 &q0 $1 ]
+  &cur0 $2 &cur1 $2 ]
 
 ( testing )
 
@@ -403,6 +423,7 @@ RTN
     ;buf LDA LIT '+ EQU ;test-add32 JCN2
     ;buf LDA LIT '* EQU ;test-mul32 JCN2
     ;buf LDA LIT '- EQU ;test-sub32 JCN2
+    ;buf LDA LIT '/ EQU ;test-div32 JCN2
     ;buf LDA LIT 'L EQU ;test-left-shift JCN2
     ;buf LDA LIT 'R EQU ;test-right-shift JCN2
     ;buf LDA LIT 'B EQU ;test-bitcount32 JCN2
@@ -473,6 +494,7 @@ RTN
 @test-add32 ;add32 ;binary-32-test JMP2
 @test-mul32 ;mul32 ;binary-32-test JMP2
 @test-sub32 ;sub32 ;binary-32-test JMP2
+@test-div32 ;div32 ;binary-32-test JMP2
 @test-left-shift ;left-shift ;binary-32-8-32-test JMP2
 @test-right-shift ;right-shift ;binary-32-8-32-test JMP2
 @test-bitcount32 ;bitcount32 ;unary-32-8-test JMP2
@@ -496,9 +518,9 @@ RTN
     SWP EMIT-BYTE EMIT-BYTE
 RTN
 
-@emit-short ( x* -> )
-    SWP EMIT-BYTE EMIT-BYTE
-RTN
+( @emit-short ( x* -> )
+(    SWP EMIT-BYTE EMIT-BYTE
+RTN )
 
 @emit-byte ( x^ -> )
     EMIT-BYTE
diff --git a/tester.py b/tester.py
index 67d362a..1174fd6 100644
--- a/tester.py
+++ b/tester.py
@@ -62,6 +62,7 @@ def main():
     test(p, trials, b'+', [('x', u32), ('y', u32)], u32, lambda x, y: x + y)
     test(p, trials, b'-', [('x', u32), ('y', u32)], u32, lambda x, y: x - y)
     test(p, trials, b'*', [('x', u32), ('y', u32)], u32, lambda x, y: x * y)
+    test(p, trials, b'/', [('x', u32), ('y', u32)], u32, lambda x, y: x // y)
     test(p, trials, b'L', [('x', u32), ('y', u5)], u32, lambda x, y: x << y)
     test(p, trials, b'R', [('x', u32), ('y', u5)], u32, lambda x, y: x >> y)
     test(p, trials, b'B', [('x', u32)], u8, bitcount)