On Mon, 2013-09-30 at 20:14 +0200, Peter Zijlstra wrote:On Mon, Sep 30, 2013 at 10:10:27AM -0700, Tim Chen wrote:I've tested with Waiman's qrwlock patches on top of Ingo's patches.Here's the exim workload data:What happens if you stuff Waiman's qrwlock patches on top of that?
rwsem improvment:
Waimain's patch: +2.0%
Alex+Tim's patchset: +4.8%
Waiman+Alex+Tim: +5.3%
convert rwsem to rwlock_t for root anon_vma lock
Ingo's patch +11.7%
admittedly and oft mentioned in this thread, our current rwlock_t is
somewhat suboptimal under a number of conditions.
It does not affect the throughput for exim and I still get
about +11.7% throughput change (same as with Ingo's patch only).
Tim