This is the mail archive of the mailing list for the glibc project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH 3/3] Remove x86 assembler rwlock code

On Wed, Jun 18, 2014 at 09:00:28PM +0200, Andi Kleen wrote:
> On Wed, Jun 18, 2014 at 02:41:05PM +0530, Siddhesh Poyarekar wrote:
> > On Wed, Jun 11, 2014 at 05:23:41PM -0700, Andi Kleen wrote:
> > > From: Andi Kleen <>
> > > 
> > > With the recent tuning the C version of rwlocks is basically the same
> > > performance as the x86 assembler version for uncontended locks (with a
> > > a few cycles near the run-to-run variability). For others it should not
> > > matter anyways.
> > > 
> > > So remove the assembler code and use the C version like other
> > > architectures.
> > 
> > Benchmark results?

I was referring to either the benchmark I wrote or the one that Ondrej
wrote, since my concern was to get a benchmark into glibc.  If you can
either validate using either of the benchmarks or confirm that your
benchmark has the same characteristics (which it seems to based on
your description) then we can include the microbenchmark for it.


Attachment: pgp3AerQjw0RE.pgp
Description: PGP signature

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]