Reduces LUT needed by mul extension

bench/tests/op/divu.s

@@ -1,14 +1,55 @@
-# expected: ffffffffffffffff, dead0, 1
     .text
 _start:
-    li x2, 0xdead0
-    li x3, 0x0
-    li x4, 0x1
-    li x6, 0xdead0
+    li x2,  0xdead0        # 912080
+    li x3,  0x1234         # 4660
+    li x4,  1              # 1
+    li x5, 123456          # 123456
+    li x6,  0xdead0        # duplicate of x2
+    li x7, 765432          # 765432
+    li x8,  123456         # 123456
+    li x9, 234567          # 234567
+    li x10, 0x7fffffffffffffff  # max positive 64b
+    li x11, 0x8000000000000000  # min negative 64b (treated as unsigned = 0x8000...)
 
-    # Div by 0
-    divu x31, x2, x3
-
-    # Basic divisions
+    # 1. DIVU : a / 1
     divu x31, x2, x4
+    # expected: 00000000000dead0
+
+    # 2. DIVU : a / a
     divu x31, x2, x6
+    # expected: 0000000000000001
+
+    # 3. DIVU : smaller / larger
+    divu x31, x5, x8
+    # expected: 0000000000000001
+
+    # 4. DIVU : larger / smaller
+    divu x31, x7, x8
+    # expected: 0000000000000006
+
+    # 5. DIVU : any / any
+    divu x31, x8, x7
+    # expected: 0000000000000000
+
+    # 6. DIVU : larger / larger
+    divu x31, x7, x9
+    # expected: 0000000000000003
+
+    # 7. DIVU : max / 2
+    li x12, 2
+    divu x31, x10, x12
+    # expected: 3fffffffffffffff
+
+    # 8. DIVU : min / 1
+    divu x31, x11, x4
+    # expected: 8000000000000000
+
+    # 9. DIVU : 0 / any
+    divu x31, x3, x2
+    # expected: 0000000000000000
+
+    # 10. DIVU : any / 0
+    divu x31, x2, x0
+    # expected: ffffffffffffffff
+    divu x31, x7, x0
+    # expected: ffffffffffffffff