slow last 128MB of RAM in a 2GB system?

[Dave Johnson]

Ben Scott writes:
> On 4/20/07, Bill McGonigle <bill at bfccomputing.com> wrote:
> > Ouch - is that simply a matter of cache impact on
> > performance?  I wouldn't have guessed it would be so high.
> 
>   It could be (although it could be something else I'm just not aware of).

50x isn't too bad. I'd expect you would end up with different results
depending on which memory happends to end up in that last 32MB.  The
highmem area can be disk cache, userspace code, and userspace data,
but not kernel code or kernel lowmem allocations. If you were to run a
bunch of programs first then the test program you may end up with the
test program in cached memory, but if some critical disk blocks end
up uncached (say the test program's code segment, or even worse an
important part of libc) then it'll crawl.

-- 
Dave