Aritz Bastida writes:
> I need to use pktgen for sending packets at very high speed to another
> machine, in order to test it under heavy network traffic. All my
> previous injection test were done with a dual Pentium III 800 MHz. As
> I needed a more powerful machine I got a Pentium 4 but the results are
> quite similar.
Also make sure HW_FLOW control from the receiver is not throttling the
sender.. Also adapters are different and also PCI-bridges.adds latency.
Bus latency and bus speed is the most important factors so PIII might
very well be faster then XEON/p4/Opteron.
> These results are similar to the ones I got with the Pentium III. I
> can't reach even 400kpps. With another machine, a dual AMD Opteron, I
> can send as fast as 650kpps but unfortunately that is the machine
> being tested. In the pktgen paper says "A single flow of 1.48Mpps is
> seen with a Xeon 2.67 GHz using a patched e1000 driver (64 byte
> packets)" [1]. Well, I don't know how much faster the Intel Xeon is,
> but I do have a fast machine, and I do have an e1000 NIC.
The 1.488 Mpps was with several a patches one with HW TX prefetching which
was catastrophic in production. It worked only in simplex mode.
Below is from our dual 2.67 GHz XEON w.serverworks chipset and Intel 82546GB
More or less vanilla 2.6.14
One
880039pps 422Mb/sec (422418720bps) errors: 0
Two concurrent
850419pps 408Mb/sec (408201120bps) errors: 0
850421pps 408Mb/sec (408202080bps) errors: 0
All with 64 byte pkts (60 in pktgen) you're using 100
Cheers.
--ro
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
