Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/5/84; site psuvax1.UUCP
Path: utzoo!watmath!clyde!bonnie!akgua!whuxlm!harpo!decvax!genrad!mit-eddie!think!harvard!seismo!rochester!cmu-cs-pt!cadre!psuvax1!ian
From: ian@psuvax1.UUCP (Ian Parberry)
Newsgroups: net.research,net.math
Subject: optimal sorting networks
Message-ID: <1667@psuvax1.UUCP>
Date: Wed, 3-Jul-85 16:41:17 EDT
Article-I.D.: psuvax1.1667
Posted: Wed Jul  3 16:41:17 1985
Date-Received: Fri, 5-Jul-85 06:32:41 EDT
Distribution: net
Organization: Pennsylvania State Univ.
Lines: 13
Xref: watmath net.research:189 net.math:2123


In Knuth "The Art of Computer Programming", Volume 3, Section 5.3.4,
there is a list of the best known upper and lower-bounds (at the time of
publication) on the size and depth of sorting networks with up to 16 inputs.
The optimal size and optimal depth are known for n<=8.

Has any more progress been made for n>8 in the last 12 years?  I am
particularly interested in n=9,10,...16, but any "small" n will do.

P.S.  Yes, I do know about the Ajtai, Komlos and Szemeredi result, and the
subsequent improvements by Mike Paterson.  Even Batcher's sorting networks
beats the latter for "small" n (the last figure I heard bandied about was
2 to the power 1300).