>>>>Benefit of buffering on host gave a stronger boost than piping the data through
>>>TCP/IP stack without physical cables added as overhead....
>>>
>>>But, IIRC, you had to employ the TCP/IP to access host machine from inside the VM, cableless and simulated probably, but still used. The overhead you avoided seems to be the emulated disk access inside the VM and its memory management.
>>
>>Yupp. But the speed benefit resulted IMO from option to pool disk caching on host for all VMs.
>
>I did pretty much the same for about three years, when I ran a virtual W7 on Ubuntu host, doing almost all of everything on an NT partition of the host, mapping it as a network drive. Didn't notice any speed degradation. In the long run, it saved me a lot of time, because I didn't waste it on M$-induced overhead - turned updates off in the VM, so no new devils in the dll hell, no functionality gone because of a security patch, no new option with wrong default which takes few hours to find out and set to normal...

The benefit was marked in my case as the code was really skating on the platters AND 2 or 3 cores were accessing HD (each a different ones) at the same time. So even if 1 core was only working on a set of small tables, physical memory could be used by other core massaging large table.