B
BSchnur
No problem -- based on what Dale said, I was actually expecting a
statistically significant change -- especially on a Vista 64 system.
I do expect to see a change on the DDR2 based system, which is what I
plan on testing when I have the parts to do it, if only because I'll be
installing faster memory (667 versus 533).
What I need to do on the first system is see what Sandra reports for
memory performance as well.
That being said, even if one doesn't see a significant change in the
memory speed tests, I *would* expect an actual performance increase
especially if one is running multiple applications -- memory caching
being much faster than disk caching.
I suppose the thing here would be for there to be two performance tests
related to memory, one for actual memory benchmark performance (which
should not change appreciably with more memory) and one which reports
performance that relies on memory cache being larger -- which should
report major performance gains.
The thing with Vista compared to XP is that this performance does scale
much further. From what I've seen in the past extra memory performance
gains get pushed further and further up. With Win 95 memory above 128M
didn't matter that much. With Windows 98, that got stretched to 256M.
With Windows 2000, 512M (or even up to 1G). With XP moving from 512M
to 1G made a significant difference for many users. With Vista (even
Vista 32), improvement shows up going from 1G to 2G. (with Vista 64,
that number gets stretched further).
The thing is, Microsoft calls their benchmark as one based memory
operations per second. Perhaps they are including some other component
in a more composite test, but are not providing the technical
information behind that so as let some hold to their cherished beliefs.
statistically significant change -- especially on a Vista 64 system.
I do expect to see a change on the DDR2 based system, which is what I
plan on testing when I have the parts to do it, if only because I'll be
installing faster memory (667 versus 533).
What I need to do on the first system is see what Sandra reports for
memory performance as well.
That being said, even if one doesn't see a significant change in the
memory speed tests, I *would* expect an actual performance increase
especially if one is running multiple applications -- memory caching
being much faster than disk caching.
I suppose the thing here would be for there to be two performance tests
related to memory, one for actual memory benchmark performance (which
should not change appreciably with more memory) and one which reports
performance that relies on memory cache being larger -- which should
report major performance gains.
The thing with Vista compared to XP is that this performance does scale
much further. From what I've seen in the past extra memory performance
gains get pushed further and further up. With Win 95 memory above 128M
didn't matter that much. With Windows 98, that got stretched to 256M.
With Windows 2000, 512M (or even up to 1G). With XP moving from 512M
to 1G made a significant difference for many users. With Vista (even
Vista 32), improvement shows up going from 1G to 2G. (with Vista 64,
that number gets stretched further).
The thing is, Microsoft calls their benchmark as one based memory
operations per second. Perhaps they are including some other component
in a more composite test, but are not providing the technical
information behind that so as let some hold to their cherished beliefs.