I very much doubt it. Its best to ask your local Sybase sales rep.
    But, if its one physical CPU, I don't see how they can charge you
    a license for two (assuming your licensing is CPU based). Regardless,
    an Internet Access license isn't dependant on the number of CPUs
    (or is it?). The real issue would be - how many Sybase ASE engines
    can you run? If its only one, then its only the one CPU to ASE.

    No idea. Its best to ask that in the Solaris or Sun hardware newsgroup.
    Sun will have info online as well.

    Hard to say. Sybase ASE has its own threading model so it doesn't
    make use of an OS's own threading. Whether its on or off makes no
    difference to ASE.

-am    © 2004

It would depend on Sybase's licensing model.  They might choose to
make you purchase two licenses for an UltraSPARC IV, since it is
roughly twice as fast as an UltraSPARC III....

Well, you can disable a processor core using 'psrset' but then you
aren't using half of the machine that you bought.

No, it's quite different.

A hyperthreaded Intel CPU has a single processor core which
alternately issues instructions from two threads (the details are more
complex, but that's the rough idea).  Once the instructions from the
two threads have been issued, the single processor core executes them
together, possibly simultaneously if enough resources are available.

An UltraSPARC IV, like a POWER 4, has two separate processor cores.
They each run one thread.  In many ways, it's like having two entirely
independent processors, just glued together into one chip.

No.

On the Xeon, you want hyperthreading disabled because most of the
threads are doing database work, and would be competing for the same
set of processor resources -- having two database threads in the
processor at once would just slow both of them down.

On the SPARC, you want both processor cores enabled because they each
have a complete set of resources [*] and would both be doing useful
work on a database thread.  Given two threads, they'll run very nearly
twice as fast as a single thread.

-- Anton

(Disclaimer: I work for Sun, albeit not in the processor or systems group.)

[*] There are a few resources shared, notably the system bus interface
and the cache.  However, the system bus interface supports several
outstanding transactions, so it's unlikely to be a bottleneck.  The
cache could be, but two database threads are likely to be accessing
many of the same areas in memory, so the shared cache might actually
be a net win here, or at least a wash.