On Wednesday 07 August 2002 16:19, you wrote:
> are you talking about JUMBO frames? or am I missing something.
> Increasing the size of the MTU in a gigabit network, you might be
> suprised with the results.
Probably. The time time between 1500 byte packets is ridiculously small at
Gigabit speeds.
> Also, not all cards are created equal,
> look for ones with tcp/ip checksum offloading n stuff. The less
> interupts the better, as it takes so so so long for the processer to
> handle the request, it's a shame you can't get networks cards with
> direct dma transfers...,like scsi.... ;-).
AFAIK the checksum calculation with modern CPUs has almost zero overhead, a
lot of work was done on the algorithm, and permitting the calculations to run
in parrallel with other networking code. Remember modern CPUs have pipelines
and try to do more than one thing at a time, including out of order execution.
This has nothing to do with DMA, AFAIK all decent PCI cards will use DMA for
transfers.
Even with DMA, the card generally has to generate an Interrupt to say a
packet needs servicing. There are cards which permit polling rather than
interrupts, and a driver was under development to use that feature. If
you're interested a search of Kernel Traffic or the LWN kernel pages, might
show up an explanation how a non-interrupt driven driver would work.
> Lee
please don't top post and learn to snip!
Rob
--------------------------------------------------------------------
http://www.lug.org.uk http://www.linuxportal.co.uk
http://www.linuxjob.co.uk http://www.linuxshop.co.uk
--------------------------------------------------------------------
This archive was generated by hypermail 2.1.3 : Wed 07 Aug 2002 - 18:02:14 BST