:: 26-Oct-2008 09:52 GMT (Sunday) ::
Dear friends,
It is with great excitement that we announce our upcoming project to
find the Optimal Golomb Ruler with 26 marks. Optimal Golomb Rulers
have many applications including sensor placements for X-ray
crystallography and radio astronomy. Golomb rulers can also play a
significant role in combinatorics, coding theory and
communications. Dr. Golomb was one of the first to analyze them for
use in these areas.
New clients and proxies will be required to support the new project,
which we are calling “OGR-NG”. The new clients will have version
2.9101.507 or higher. Clients with version numbers prior to this will
only be able to work on the RC5-72 project. Similarly, Personal Proxy
version 343 or higher is required.
Updated clients and proxies are being made available for early-adopter
testing at http://www.distributed.net/download/prerelease.php for many
major platforms. Builds for additional platforms are gradually being
produced and will be made available on that page as soon as
possible. If you do not feel comfortable helping with this
early-adopter testing and reporting bugs, then we recommend waiting
for a few days when we will move those clients to the official release
page. Not all of our public keyservers have finished upgrading to the
new version, so not all will have OGR-NG workunits yet. Stats for the
new project will also be launching in a couple of days. As with all of
our pre-release software, we encourage you to file bug reports for any
problems you find at http://bugs.distributed.net/
The OGR node rate for these new clients may appear to be slower than
those for the OGR-25p2 project. This isn’t a cause for concern. The
effective search rate for the overall project is actually more
efficient as these new clients are using an improved algorithm. The
new algorithm is named FLEGE (Feiri-Levet Enhanced GARSP Engine). It
has been developed by Didier Levet and Michael Feiri over the course
of the past few months. We are especially grateful to them for this
Herculean effort. In technical terms, the number of elements in the
‘choose lookup table’ has been increased from 48K to 2M
elements. Although this will slightly increase the size of the dnetc
binary, this optimization will significantly reduce the number of
nodes that we have to search through, sometimes a node improvement of
ten-fold or more.
We estimate that OGR-26 will take much less time to complete than
OGR-25 and will probably be more similar to OGR-24, in terms of
computational effort. Beyond that, we’re looking forward to OGR-27 and
OGR-28 because the current solutions to at least one of the two is
very likely not optimal. OGR-26 is on the way and must be done first
since higher-order rulers depend on the proven optimality of previous
ones.
We are grateful for your continuing support and enthusiasm.