> For what it is worth I did some trials in matching page size to the
> underlying virtual memory page size and was surprised to find that I did
> not measure any significant performance change. My guess is that it is
> the structure of the data in your application which would be sensitive
> t
For what it is worth I did some trials in matching page size to the
underlying virtual memory page size and was surprised to find that I did
not measure any significant performance change. My guess is that it is
the structure of the data in your application which would be sensitive
to page siz
* Chris Schirlinger:
> 4096 in Win32 machines and 1024 on nix ones (I think, I am no expert
> on Unix style OS's)
This depends on the file system. On Linux on x86, it's typically 4096
bytes.
According to my performance measurements, switching from page size of
1024 bytes to 4096 bytes gives a
There doesn't appear to be any real documentation over what page size to use. I
think it is more of a case of experimenting and determining which is best for
your system/application.
In the archive I found an article stating that for optimum performance on Win32
to match the page size with that
We found the best setting for page size was to match the size of the
cluster size of the drive the database was expected to run on
4096 in Win32 machines and 1024 on nix ones (I think, I am no expert
on Unix style OS's)
I would think, if you set it to the exact size of a number of records
all
Hi Paolo,
i met the same problem.
I have experimented with the page size and found that the cluster size
of the file system brought the best results. However, the gain in speed
was not very big (5%)
Martin
Zibetti Paolo schrieb:
I could not find a document explaining how to find the optima
6 matches
Mail list logo
