Re: How to verify RX checksum offload??
Re: How to verify RX checksum offload??
- Subject: Re: How to verify RX checksum offload??
- From: Andrew Gallatin <email@hidden>
- Date: Tue, 07 Jul 2009 19:32:16 -0400
Terry Lambert wrote:
Presumably if you are getting good inbound packets, they're going to
have to go to memory sooner or later, so it might as well be sooner, and
you will presumably be reading their payload in order to act on it, so
your cache is no worse off than otherwise.
People keep saying this, but empirical evidence keeps proving them
wrong ;). Software rx checksums hurt performance 10 years ago, when
I did a lot of performance analysis for
http://www.usenix.org/events/usenix99/full_papers/gallatin/gallatin.pdf
and the situation is even worse today. This is mainly because now some
OSes use non-temporal (eg, uncached) copy routines for bulk data, or
they use the IOAT DMA engines (which are present on Apple machines,
but which so far seem to be ignored by MacOSX).
Drew
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Darwin-kernel mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden