[Info-vax] Real time performance, was: Re: VSI: "Official 8.4-1H1 Launch"

[David Froble]

Stephen Hoffman wrote:
> On 2015-06-08 03:47:24 +0000, Robert A. Brooks said:
> 
>> On 6/7/2015 6:25 PM, David Froble wrote:
>>
>>> If I had to guess, I'd say the native stuff will exist before they 
>>> add stuff for running in a VM.  It still will be executing on x86, 
>>> not in emulation.  You still need something that runs on x86 before 
>>> you can run it on x86 in a VM.
>>
>> It's not as clear-cut as you think . . .
> 
> I'd be betting that the VM path well might "boot" before the hardware, 
> just because interfacing with software virtual devices is usually 
> somewhat easier and more predictable than with physical hardware, which 
> is a boon particularly in the boot path.  You also get reasonable 
> visibility into the carcass of a failed boot, using the VM framework 
> that's running and stable, too.
> 
> 

Ok, yeah, I think I'm understanding a bit more.  Having to support HW 
devices can be interesting.  I'd guess that since the VM supplied 
devices are software, they might be easier to work with.  Maybe a lot 
less flaky?

Or am I still clueless?