[daip] Y2K (the new DDT) - how quickly they forget

[Ketan Desai]

Opinions submitted by an outside observer:

> 
> A number of questions arise about the new DDT:
> 
> 1. What should the master architecture be?  The official lies were
> computed on Solaris for DDT93.

I would say still a Sun machine.  There are too many variants of Linux machines
to pick one as the master version.

> 
> 2. What machines should I test here?  I plan to run DDT large and then
> medium and large Y2K.
> 
> 3. Ketin's machine has done DDT in 341 sec - should I scale the new
> large to be ~10 X that?  More?

Is a target of an hour for the large DDT too short ?

[in a more theretical bent, could Moore's law be used to determine a
useable factor here?  Just put in the time difference between now and the
last time the DDT was scaled.  To be really fancy, produce a DDT that 
automatically increases the size of the problem as a function of time in
expectation of Moore's law. - okay, enough spam about a subject I know
nothing about...]






-- ketan
M.O.T.U.