Alan> No the problem we have is that the time go in and out of irq
Alan> handlers and net_bh is a per packet constant. At 100Mbit small
Alan> packet routing we cant cope (nor can BSD). At 1Gbit 1500 byte
Alan> framing Linux routing is out (which in a way is fine since PCI
Alan> is too slow - but PCI wont be too slow next year). Jes has
Alan> accurate data on what 1500->8K packet size transitions do to our
Alan> 1Gbit performance.
Preliminary tests with a Pentium Pro/200 as the slowest of my two test
hosts showed a performance increase from about 28MB/sec to 40MB/sec
when increasing the packet size from 1500 bytes to 9K.
Jes
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/