Of course it will be, Dave.

But there is no evidence at all that Windows (any form) pages out "stale" memory in any resident (as opposed to 'active') process, which it could do during wait states (of which there are plenty in these ultra-fast CPUs).

there is also no evidence to suggest that Windows 'anticipates' RAM shortages in the active task (again taking wait state cycles to prepare for the eventuality).

Finally, it **looks** like Windows *always* swaps only the active task.

There are plenty of opportunities in improving the swap mechanism of Windows, even with RAM getting cheaper and cheaper, because as it does we find more and more uses for it.
My take is that there is a virtual (pun intended) goldmine here. MS could drastically improve Windows performance and capabilities (for applications) by addressing this.

Cheers

JimN


>Jim,
>
>Disk is always going to be at least 3 orders of magnitude slower than RAM, until that changes VM will always be a significant performance hit.
>
>>Given my loooong ago experiences with mainframe paging/swapping there is plenty of scope for improvement in the way Windows does the job. Hopefully, some day.