Can somebody recommend a freebie tool that will give an
indication of how fast my system is doing office rat things like
opening MS Office documents, saving same ..... and maybe playing
media files?
I'd like to fool around with a few different memory configs and
need something to give me an idea of whether things are getting
better or worse with a given change.
--
PeteCresswell
(PeteCresswell) wrote:
> Can somebody recommend a freebie tool that will give an
> indication of how fast my system is doing office rat things like
> opening MS Office documents, saving same ..... and maybe playing
> media files?
>
> I'd like to fool around with a few different memory configs and
> need something to give me an idea of whether things are getting
> better or worse with a given change.
For memory changes, I'd use a tool like SuperPI. SuperPI uses
some memory, and depending on the number of digits computed,
should "bust out of the cache" if enough digits are selected.
And since it is not measuring memory bandwidth as such, it
might reflect how a real application behaves to the memory
change. (It is sensitive to both CPU speed, and memory speed,
sort of like Photoshop might be.)
Everyone has their favorite benchmark, and that one is good enough
for me. I like a benchmark that doesn't use the disk, because if your
disk accesses are part of the time interval measured in the
benchmark, the changes made by tuning the memory, might be
all but invisible :-) An application like SuperPI, might make
you feel better.
My system varies from 45 seconds (best result) to my result
today of 50 seconds. If your OS has a lot of "junk in the trunk",
the SuperPI results might suffer a bit, just like your real
apps might.
Per Paul:
> I like a benchmark that doesn't use the disk, because if your
>disk accesses are part of the time interval measured in the
>benchmark, the changes made by tuning the memory, might be
>all but invisible :-) An application like SuperPI, might make
>you feel better.
Unencumbered by any knowledge at all, I was thinking maybe if XP
had enough excess memory, additional speed bco caching disk I/O
in said memory might be a factor.
(PeteCresswell) wrote:
> Per Paul:
>> I like a benchmark that doesn't use the disk, because if your
>> disk accesses are part of the time interval measured in the
>> benchmark, the changes made by tuning the memory, might be
>> all but invisible :-) An application like SuperPI, might make
>> you feel better.
>
> Unencumbered by any knowledge at all, I was thinking maybe if XP
> had enough excess memory, additional speed bco caching disk I/O
> in said memory might be a factor.
>
> Or am I just wishing?
But what are the odds, that the thing you want, is in the cache ?
My experience with Win2K, is file caching does seem to work.
The second time I execute a checksum program on a file, the
execution time is reduced, and the disk light doesn't flash.
So that works.
But I find directory caching doesn't seem to work, and I don't know
if that is related to my antivirus software or not. My metric
there, is the behavior of the search command.
In any case, in practical terms, I don't seem to be seeing a
significant effect here (little command to command load time
variation), and I currently have 2GB installed, of which only
a small percentage is occupied.
Benchmark Tool? 
Linear Mode
