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

[Johnny Billquist]

On 2016-01-18 18:23, abrsvc wrote:
> On Monday, January 18, 2016 at 12:00:11 PM UTC-5, Johnny Billquist wrote:
>> On 2016-01-18 15:00, Kerry Main wrote:
>>> Most system vendors systems are driven by marketing and positioning
>>> to target markets.
>>>
>>> One possible example is things like removing NOOP's in ucode, adding a
>>> few tweaks & selling the new ucode as an enhancement to an existing
>>> system. Yes, they might throw in a new label to attach to the system, add
>>> new p/n's etc so it has the appearances and justification for the $'s spent.
>>>
>>> They might also do things like purposely slowing down benchmarks on
>>> one system so that a more favored system from the marketing groups
>>> is easier to position as "better".
>>>
>>> Pure speculation of course ..
>>>
>>> :-)
>>
>> Oh, I know the stories. There have been plenty over the years. How many
>> of them actually are true is a different story, though...
>>
>> For the 11/750, I seriously doubt this would have been the case, but I
>> obviously don't know for sure.
>> But let's say the VAX-11/750 could perform on par with the 11/780.
>> Wouldn't have made much more sense to just price the 11/750 a little
>> cheaper than the 11/780, and just stop producing the 11/780. The margins
>> on the 11/750 would have been much greater, since it would cost way less
>> than the 11/780 to produce.
>> So customers would have gotten the same performance at a lower price,
>> and DEC would have made much more money... Can't see how they could have
>> refused that.
>>
>> 	Johnny
>
> IIRC the 750 was about 80-90% of the capacity of the 780 in terms of CPU speed.  The benefit was both the smaller footprint as well as not requiring anywhere near the same AC.  The 750 was also limited in its peripherals.  The 780 had both MassBus and UniBus. The 750 was restricted to the Unibus.

I think the official numbers are 0.6 VUPS, while the 11/780 by 
definition is 1.
But the 11/750 can take both Massbus and CI. So that would not be a 
differentiating factor. However, you limited to (I think) one massbus 
and three unibuses on the 11/750. (Or maybe it was one of each?)

> The bottom line was "similar" performance at less cost both of the actual machine as well as ownership (AC, power etc.) which mafde it much more affordable for the masses.

Well, buying and running an 11/750 was way cheaper. But I do remember 
the feeling that they were slow. However, I mostly would have compared 
that with an 11/785... Which obviously widened the gap.

	Johnny