Re: Variability in cache access times

Stephen C. Tweedie (sct@redhat.com)
Thu, 28 Jan 1999 16:31:04 GMT


Hi,

On Thu, 28 Jan 1999 00:20:07 -0800, "David S. Miller"
<davem@dm.cobaltmicro.com> said:

>> Add page coloring to the page allocator and your variability will
>> disappear.

> From: Richard Gooch <rgooch@atnf.csiro.au>
> I was wondering if that would be the response. What about using 4 MB
> pages? Would that help?

> It would, but I have no idea how feasible this is with the Intel page
> tables. Ingo could probably more quickly cook up a test patch than
> me.

As far as the Intel CPU is concerned, it is not particularly hard. The
real problem as far as Linux itself is concerned is twofold: first, it
requires contiguous 4MB pages, and secondly Linux assumes a uniform,
fixed-depth page table hierarchy. It is really not set up for looking
at ptes inside the page directory, although with a bit of tweaking it
would be possible to repeat the pmd folding trick and make the 4MB pte
appear as a single-entry page table.

--Stephen

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/