[Virtualacorn-list] Virtual Acorn Setup Problem

[Nick Roberts]

In message <c1fea2fa4f.Alan.Adams at orchard-way.freeserve.co.uk>
       Alan Adams <alan at adamshome.org.uk> wrote:

> In message <mpro.k9xrgm0006uxb01n8.groups at planet3.freeuk.co.uk>
>           Brian D <groups at planet3.freeuk.co.uk> wrote:
> 
> > Nick Roberts <tigger at orpheusinternet.co.uk> wrote:
> 
> > > I've found that a more optimal setting is to select "Custom"
> > > pagefile size settings, with the "Initial" and "Maximum" settings
> > > the same. Given the size of current hard discs, I tend to be
> > > fairly generous with the pagefile size - it's set at about twice
> > > the size that Windows recommends.
> > > 
> > > If you go this route, it's worthwhile doing it as soon as
> > > possible, before Windows has a chance to fragment the pagefile
> > > too badly.
> 
> > What I do on established systems is to remove the (variable size)
> > pagefile (and if available put a fixed one on a different disc) and
> > re-boot. This gets rid of the original (fragmented) pagefile (and
> > if there is no pagefile on an alternative drive, a very slow
> > machine.
> 
> > I /then/ configure a large fixed size pagefile, at least twice the
> > memory size on the original disc.
> 
> Interesting that the advice is always to use twice the memory size.

My recommendation, and that of Brian D, were to use twice the amount
of page file that Windows suggests as the recommendation, not twice the
amount of physical RAM.

> In  fact, the size of pagefile needed reduces as memory size
> increases -  the total of memory and pagefile need to be big enough
> to accommodate  the largest concurrent workload expected. It an ideal
> world, the  memory would be big enough not to need a pagefile - just
> as RISC OS  works.

Even when you throw a lot of RAM at Windows, it wouldn't work the
same way as RISC OS does. For example, once you have run a program
once, it keeps it around in memory after you quit, unless and until the
memory neds to be reclaimed for another task - it has the benefit of
hugely speeding up subsequent loads of the same app.

And unless you are running 64 bit Windows XP, it can only use about
3.25GB of physical RAM (4GB RAM address space, less about 0.75 GB for
memory-mapping graphics cards, sound cards, etc). I'm not sure whether
Vista is any better in this regard.

> The only reason for increasing pagefile when increasing memory is
> that  increasing memory is usually done as a result of increasing
> workload,  e.g. Vista instead of XP, Office instead of Works.

If you've ever tried using Omnipage to OCR 300-page PDFs you would
realise why I feel the need to be generous with the pagefile.

And given the current cost and capacity of hard discs, I don't think
that losing a few GB of disc space is much of an issue.

-- 
Nick Roberts           tigger @ orpheusinternet.co.uk           

Hanlon's Razor: Never attribute to malice that which
can be adequately explained by stupidity.