[Info-vax] Why it is a good idea that OpenVMS isn't on x86-64 just yet

[gcornelius at charter.net]

Johnny Billquist <bqt at softjar.se> wrote:
>Norm Raphael wrote:
>> TU58's used the DISK DRIVER, not the tape driver, so they blocked
>> disk I/O when in use.  Slow doesn't really convey it.
> 
> That was not necessary. We're talking about the ACP here, which would 
> only be involved if the TU58 was in Files-11 format. If it was in RT-11 
> format, you mounted it foreign and used EXCHANGE.
> Second, I'm sure you could use a separate, unique ACP for the TU58, just 
> like in RSX, to keep it separated from all other system activity.

 $ MOUNT/PROCESSOR=UNIQUE

Of course, most of us considered that an advanced qualifier way back
when and may not have taken the time to learn it.

Good point about $ EXCHANGE, though. I would not have thought of
that.

> Third, I'm sure that for plain operations reading/writing blocks, it did 
> not even block the ACP, unless VMS was more silly than RSX.

Well VMS eventually went to the XQP, the Files-11 ACP as part of the
executive.  It of course was/is multithreaded or would have become an
extraordinary bottleneck in multidrive configurations.

>        Johnny