First, sorry for the name confusion. The real name is Leonid.

Second, since I don't see any abs suitable transforms or vops
in the SBCL source, I assume there isn't one.
In what other ways CMUCL is ahead of SBCL
(in terms of optimizations)?

On 8/18/06, Raymond Toy <> wrote:
>>>>> "Alexander" == Alexander Slobodov <> writes:

    Alexander> While reading  this book it occurred to me that it would be interesting
    Alexander> to compare the shown algorithms with the ones SBCL uses.

I have implemented some of these for CMUCL.  Some of the algorithms in
the book were not faster than existing algorithms, but abs for
signed-byte 32 was implemented.