an experiment in pipe bandwidth improvement

[Zack Weinberg]

On Tue, Jan 25, 2000 at 09:40:36PM -0800, Gregory P. Smith wrote:
> On Sun, Jan 09, 2000 at 07:32:30PM -0800, Zack Weinberg wrote:
> > I attempted to improve the bandwidth of a pipe by enlarging the kernelside
> > buffer if a writer offers more data than there's room.  Currently a pipe
> > write is broken up into one-page chunks.  The theory was that this would
> > reduce the number of context switches and system calls, and therefore improve
> > performance.
> > 
> > The results are quite surprising: bandwidth is reduced by about 100 MB/sec.
> 
> Quick summary from the rest of this thread: it sounds like the reason
> this is slower is due to killing the CPUs L1 cache as well as the memory
> allocation being relatively slow.  That makes sense.
> 
> Without looking at the linux pipe code first (a sin, I know), does it
> always copy the data twice for pipes?  I recall that in FreeBSD 2.2.x it
> does something similar to this:  The pipe code will "simply" <evil_grin>
> remap the page to the reader's target buffer if it is properly aligned
> (actually, it would have to map it as the kernel pipe buffer first;
> remapping it into the reader only when it actually calls read).

Yes, it always copies the data twice.  A socket does the same thing.
Linux may or may not be capable of the remapping trick but I couldn't
find a way to do it.  In general, FreeBSD seems to assume that page table
manipulation is cheap, and Linux that it is expensive (this from not very
thorough reading of both kernels).

I bet there'd be some point, probably different for each platform, at
which page flipping became a win.  I also bet it'd be much bigger than
just one page.

zw

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