[Info-vax] A few notes from VMS TUD days i Stockholm today.

[John Reagan]

"JF Mezei" <jfmezei.spamnot at vaxination.ca> wrote in message 
news:00350f2f$0$26761$c3e8da3 at news.astraweb.com...
> IanMiller wrote:
>
>> Tukwila - it's delayed so HP chose to separate Tukwila support from
>> VMS V8.4 so V8.4 is not delayed. Remember its not just the chip from
>> Intel but the whole system around it including firmware that has to be
>> completed then properly tested and each operating system qualified to
>> run on it. It's better this is done right than rushed.
>
> I realise that. But I was wondering if, at this stage of development of
> Tukwila, whether the feature set would be locked and thus, the OS code
> that provides chip specific support *could* also be fixed.
>
> Also, doesn't HP already have test Tukwila systems that they can test
> VMS on ?
>

Yes there are proto systems for testing.  They are probably using all sorts 
of prototype parts that aren't the final rev level (including the chip 
itself).  There has been a history of last minute firmware changes that have 
required changes to OpenVMS.  This isn't new with Itanium either.  There 
were Alpha systems that changed near the end of qual that caused changes in 
the OS.  Even something as simple as a change to the USB controller's rev 
level might interfere in the boot sequence for instance.  Protos often run a 
somewhat slower speeds so timing bugs don't really show up until the proto 
systems themselves get upgraded internally several times during the 
development and testing phases.  Again, not unique to Itanium.  Some early 
Alpha EV6 protos had floating point disabled for instance (if I remember 
correctly).

John