32bit PAE v.s. non-PAE performance

[Joseph Salisbury]

On Sun, Jan 3, 2010 at 9:42 AM, Tim Gardner <tim.gardner at canonical.com> wrote:
> Gleaned from a discussion on LKML entitled 'Ubuntu 32-bit, 32-bit PAE,
> 64-bit Kernel Benchmarks'
>
> http://www.phoronix.com/scan.php?page=article&item=ubuntu_32_pae&num=1
>
> It appears that the performance impact of PAE is not as great as was
> originally feared (given a normal workload).
>
> rtg
> --
> Tim Gardner tim.gardner at canonical.com
>
> --
> kernel-team mailing list
> kernel-team at lists.ubuntu.com
> https://lists.ubuntu.com/mailman/listinfo/kernel-team
>

I performed some testing a few years ago comparing 32bit PAE vs
non-PAE kernels.  Performance was within 10% or less using micro
benchmarks(lmbench, tiotest, etc).  The real benefit to the PAE Kernel
was seen running database benchmarks such as DB2 or TPCC(Comparing PAE
to non-PAE - excluding 64bit).  For example Oracle can address up to
62GB of memory versus 1.7GB.  So any applications that will benefit
from the ability to address addition memory will have a tremendous
performance boots versus non-PAE kernels.

Thanks,

Joe