kernel performance degradation.

C. Scott Ananian (
Sat, 2 May 1998 18:13:57 -0400 (EDT)

Off and on throughout the 2.1.x kernel series we've heard various reports
of kernel performance degradation of one form or another: "low memory
systems work better with 2.0.x", "networking slowed down when I upgraded",
even a report (from some time back, fixed since, I think) that disk
benchmarks were better on 2.0.x, despite the new dcache code.

I'd like to put these bugaboos to rest. Regrettably I don't have the time
to undertake this myself right now, but I'd like to see someone start a
large-scale benchmarking effort comparing 2.1.latest to 2.0.X.
It would be nice to quote numbers to *show* that daveM's networking fixes
really do work, and that the new memory manager really is better at
defragmenting memory, and that the new dcache code really does work.
I know various people have been doing bits and pieces of this
benchmarking, but I'd like to see it all in one place, and against the
most recent 2.1.x kernel (so we can be sure that speedups in the early
2.1.x series didn't get rendered moot by bugfixes later on, or some such).

And, of course, if we do since significant slowdowns in 2.1.x, we'll know
what's left to fix before 2.2.

Any volunteers? Any previous benchmark results to report?
I'll serve as coordinator and post the results on the web and to
linux-kernel from time to time if other people can do the benchmarking.
@ @
C. Scott Ananian: / Declare the Truth boldly and
Laboratory for Computer Science/Crypto / without hindrance.
Massachusetts Institute of Technology /META-PARRESIAS AKOLUTOS:Acts 28:31
-.-. .-.. .. ..-. ..-. --- .-. -.. ... -.-. --- - - .- -. .- -. .. .- -.
PGP key available via finger and from

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to