MCNC benchmarks (was: Simple way to generate random netlists of ALUcells)

Hi,

I'm looking for a way to get or generate small netlists of high level
cells to test a placement algorithm. I don't need timing or
functional information about the netslist. I posted a description of
this in the original thread.

Many of the papers I've read refer to the MCNC testbenches at
http://www.cbl.ncsu.edu. Many of the links on their benchmark page
seem to be obsolete. In particular, most links that can provide
background to navigate the site (and its various test suites) are no
longer available. These include the technical reports, the
Benchmark-review, and the ftp address.

Is there some kind of guide information that helps visitors to the
site choose a test suite? MCNC seems to be very prevalent, so I
assume this "high-level view" information exists somewhere (maybe in
the technical reports?). A paper I found that seemed like it might
help is

ACM/SIGDA Design Automation Benchmarks: Catalyst or Anathema
Franc Brglez
IEEE Design & Test of Computers (magazine)
10(3), Sept., 1993, pp. 87--91.

Strangely, this paper seems to have disappeared from the web. It
doesn't show up in that issue in IEEE Explore. Our library also has
access to ACM, which merely points to the IEEE society website.
There, a user login is needed to get the article. So despite my
personally subscribing to D. and T., and the library having access to
IEEE and ACM, I can't get it.

Thanks for any pointers to where the technical reports might still be
available. Or, if there is some kind of summary/overview that can
help in determining which of the test suites to start looking at
first, I'd appreciate a referral to that, too. How do other newcomers
go about getting this "roadmap"?

Fred