[Info-vax] x86 Update 01/20/19

[clair.grant@vmssoftware.com]

On Wednesday, March 13, 2019 at 5:30:27 AM UTC-4, mcle... at gmail.com wrote:
> Clair,
> 
> I appreciate it might be vary greatly between different VMS components but how different is the code for X86.  Is it 5%, 95% or some point in between?
> 
> Also, what have been the major causes of delays?  What's taken up the time?  Is it difficult pieces of code? A larger task than anticipated? A shortage of tools?  A shortage of staff?
> 
> I'm just interested in the background to the work.

I still like my 5% (or less) number knowing about the underlying architecture. Hard to really measure but it is a very small percentage for sure and it  is concentrated in just a few areas. You do run into most of the new code very early in bringing the system up: paging, privilege access, mode changes, stack management, time keeping, calling standard mandates....you encounter them all right off the bat, and of course new compilers and linker to make it all work.

As I have said from the very beginning, we are doing this with far fewer people than what is needed to make it a predictable process. Are there difficult things to overcome? Sure, but that's not new and we have good people doing the work, many of whom have done this before.

Historical perspective: When VMS Engineering was 200 to 300 people it took 3 years to port to Alpha and then 3 years again to port to IA64.

Tools? We had XDELTA way earlier than in either of the two previous ports and Virtual Box has made things orders of magnitude faster than dealing only with real HW.