Messages in this thread | | | Date | Tue, 24 Sep 2002 15:01:23 -0700 | From | William Lee Irwin III <> | Subject | Re: on 2.5.38-mm2 tbench 64 smptimers shows 30% improvement |
| |
William Lee Irwin III wrote: >> 2.5.38-mm2-smptimers: >> c01053dc 30936965 41.2616 poll_idle >> c020ee62 30635964 40.8601 .text.lock.dev <- what's this? >> c0114c08 2499541 3.33371 load_balance >> c01175db 2141278 2.85589 .text.lock.sched >> c020ce40 2141045 2.85558 dev_queue_xmit >> c013a47e 932681 1.24394 .text.lock.page_alloc
On Tue, Sep 24, 2002 at 10:40:08AM -0700, Andrew Morton wrote: > queue lock for the loopback device?
Yeah. I did some fishing around last time I ran smptimers benchmarks and found (to everyone's surprise) it was the queue lock and not the xmit lock, or at least as far as my reading of the disassembly went.
Busywait time was reduced from 80% to 45%, which is good.
Cheers, Bill - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |