View Single Post
Old 05-27-2009   #6 (permalink)
darkassain


Server 2008 R2 , Windows 7 6801 + 6936 x64,6801 6956 x32, Windows Home Premium Sp1 x64, Linux
 
 

Re: Why is Hardware Assisted Virtualisation now mandatory?

while one the reasons you are needing Virtualization support is because less needs to be done by the program to run the vm....

Virtualization support makes it possible to be able to run more "native commands" on the processor and not have to make the software do more work...

i believe they are just testing this feature in the new VPC as most of the people running XPM mode are the people running 64 bit processors that are not able to run 16 bit process in a 64 bit system (and those cryptic applications that just refuse to run because of the author for what ever reason did not support it)

i said most not all so while they will eventually support non AMD-v or Intel-VTx processors (when it goes the software suite goes "gold") or any that are motherboards that are "locked" in a sense and not be able to bring Virtualization Support on to the OS or software...

and remember this is a beta not even a RC of the product...


Quote  Quote: Originally Posted by nak View Post
Hi Jakob,
Quote:

> Yes, it is stupid. Microsoft does not (officially at least) make any
> money from the sale of hardware upgrades (although they often use the
> sale of a new PC as an opportunity to sell another copy of Windows at
> the discounted OEM price).
>
> Technically, the deepest core function in any VPC product is the
> "hypervisor" which is responsible for making the Virtual machine think
> it owns its own CPU, even though it really doesn't. This hypervisor
> needs to be implemented differently (and in a more difficult way) on a
> CPU without enabled hardware VT, than when much of the work can be done
> by the VT feature of the CPU.
>
> So if Microsoft was starting from scratch with no older VPC product to
> upgrade, implementing VPC to run only with hardware VT would be the
> easiest/cheapest solution. But Microsoft is not starting from scratch,
> they have a functioning hypervisor for non-VT machines, which they could
> continue to maintain, with higher level code in VPC automatically
> loading either the VT or the non-VT hypervisor depending on machine
> capabilities.
Well you certainly have allot of knowledge on the subject, and
understand what's going on under the hood. So it seems that technically
Microsoft are doing this abandon legacy code? Rather than just optimising it
and bringing it up to new standards, same old same old.

This is a big shame and I don't see how big companies will be able to
adobt this technology without performing the upgrade. And durning a
worldwide recession this is extremely blinkered of Microsoft.
Quote:

> Note that large amounts of "auxiliary" code such as .vhd support, device
> simulation, virtual BIOS, user interface, VPC additions etc. etc. could
> be shared by the two hypervisor implementations, keeping the total cost
> down and allowing VPC images to be shared amongst different physical
> machines. It would also allow VPCs with saved state to survive a change
> in the VT BIOS setting on the Host PC.
Well I just hope one of the devs/pms might take these points on board
and reconsider their architechture. Not meaning they abandon what they have
done, just bring the legacy stuff back even stronger than it was before,
knowing Microsoft I very much doubt the code was immaculate as it was.

*fingers crossed*

Nick.
My System SpecsSystem Spec