Eric Dittman wrote:
 There are some inaccuracies, though.  For instance, the /90 and /90A
 are shown to have the same TPS, but the /90A is faster.  The numbers
 for the /90 and /90A are also wrong as far as I can tell (the numbers
 more closely match what I first posted rather than what I determined
 experimentally). 
One remark here. The TPS numbers not always seem to be "logical". Please
notice,
that a lot of the benchmark is influenced by the I/o & disk system. So a
faster CPU not always
has a better TPS number. You can see it many times if you really look
into it.
And, the 4000/90 & /90a are workstations. So TPS is not of so much value
to this ...
  There's also a lot of holes in the list. 
So fill them ;-)
cheers