update some timings

mandel.s

@@ -461,7 +461,7 @@ input_max:
     sub 4, dest, arg1, arg2
 .endmacro
 
-; 3 + 5 * bytes cycles
+; 3 + 5 * (bytes - 1) cycles
 .macro shl bytes, arg
     asl arg              ; 3 cyc
     .repeat bytes-1, i
@@ -469,17 +469,17 @@ input_max:
     .endrepeat
 .endmacro
 
-; 13 cycles
+; 8 cycles
 .macro shl16 arg
     shl 2, arg
 .endmacro
 
-; 18 cycles
+; 13 cycles
 .macro shl24 arg
     shl 3, arg
 .endmacro
 
-; 23 cycles
+; 18 cycles
 .macro shl32 arg
     shl 4, arg
 .endmacro
@@ -529,11 +529,11 @@ input_max:
     neg 4, arg
 .endmacro
 
-; 11-27 + 23 * shift cycles
-; 103-119 cycles for shift=4
+; 11-27 + 18 * shift cycles
+; 65-81 cycles for shift=3
 .macro shift_round_16 arg, shift
     .repeat shift
-        shl32 arg ; 23 cycles
+        shl32 arg ; 18 cycles
     .endrepeat
     round16 arg ; 11-27 cycles
 .endmacro