multiplies by powers of two

[Henry Cejtin]

I did notice that there is a lot of funny arrangement of code which makes
the number of false positives not so surprising.  It was as if the generated
code was intermingling 2 unrelated tasks.  Mind you, on some machines this can
be a big win since it produces more instruction level parallelism.  That doesn't
happen so much on the Intel architecture because of the scarcity of register
names.  (It really is just the names that are scarce.  I just read that the
Pentium 4, which has a very long pipeline, really has 128 registers, used in
register renaming.)