[Info-vax] clock problems with OpenVMS x86 on VirtualBox

[Arne Vajhøj]

On 5/10/2023 11:11 AM, bill wrote:
> On 5/9/2023 11:19 PM, Dave Froble wrote:
>> On 5/9/2023 6:56 PM, Arne Vajhøj wrote:
>>> On 5/9/2023 12:44 PM, Johnny Billquist wrote:
>>>> On 2023-05-09 17:52, Dave Froble wrote:
>>>>> I can guess that perhaps the biggest technical reason is the lack 
>>>>> of old VAX
>>>>> hardware to test against.  A few edits and firing up the build 
>>>>> procedure
>>>>> might be possible.  But on what HW do you run the build procedure?
>>>>
>>>> simh would be perfectly reasonable.
>>>
>>> What about build time?
>>
>>  From some past posts, the VAX/VMS build was done on one or more 
>> clustered systems.  We don't know of the requirements for an OS build.
>>
>> As far as time, who cares.  If it is to be a one time build for a VSI 
>> release, who is going to care if it takes a month or two?  A one time 
>> build is all VSI would need to allow many things, such as hobbyist 
>> use, and others.
>>
>> But, I'm guessing it would not be so simple.  None of us knows the 
>> requirements of the build procedure.
> 
> I'll bet at least one person here knows exactly what the requirements
> are and how long it is likely to take.  :-)

Everyone employed in the VMS group at DEC before Alpha arrived
would know the VAX build process.

Probably more than one. John R, Hoff etc..

I expect it to have taken a long time even on a 8xxx cluster.

I remember late 80's building a single (pretty big) application
EXE taking like an hour. VMS is a lot of EXE files.

Obviously the real HW will matter as well. An 8 core i9 and
SIMH emulating a VAX 6660 and SSD disks may help.

Arne