Re: Syslets, Threadlets, generic AIO support, v6
[Posted May 30, 2007 by corbet]
| From: |
| Ingo Molnar <mingo-AT-elte.hu> |
| To: |
| Jeff Garzik <jeff-AT-garzik.org> |
| Subject: |
| Re: Syslets, Threadlets, generic AIO support, v6 |
| Date: |
| Wed, 30 May 2007 09:20:55 +0200 |
| Cc: |
| Zach Brown <zach.brown-AT-oracle.com>, linux-kernel-AT-vger.kernel.org,
Linus Torvalds <torvalds-AT-linux-foundation.org>,
Arjan van de Ven <arjan-AT-infradead.org>,
Christoph Hellwig <hch-AT-infradead.org>,
Andrew Morton <akpm-AT-zip.com.au>,
Alan Cox <alan-AT-lxorguk.ukuu.org.uk>,
Ulrich Drepper <drepper-AT-redhat.com>,
Evgeniy Polyakov <johnpol-AT-2ka.mipt.ru>,
"David S. Miller" <davem-AT-davemloft.net>,
Suparna Bhattacharya <suparna-AT-in.ibm.com>,
Davide Libenzi <davidel-AT-xmailserver.org>,
Jens Axboe <jens.axboe-AT-oracle.com>,
Thomas Gleixner <tglx-AT-linutronix.de> |
* Jeff Garzik <jeff@garzik.org> wrote:
> You should pick up the kevent work :)
3 months ago i verified the published kevent vs. epoll benchmark and
found that benchmark to be fatally flawed. When i redid it properly
kevent showed no significant advantage over epoll. Note that i did those
measurements _before_ the recent round of epoll speedups. So unless
someone does believable benchmarks i consider kevent an over-hyped,
mis-benchmarked complication to do something that epoll is perfectly
capable of doing.
Ingo
