Hi Brian, Thanks for the inputs. The packets transferred per second are as follows. ------------------------------------------------------------------------------------- Packet Size(Bytes) 2.4.31(Pkts/sec) 2.6.16(Pkt/sec) ------------------------------------------------------------------------------------- 64 3138 3255 128 3035 3167 256 2865 2935 512 2568 2623 1024 2094 2075 1280 1923 1907 1518 1795 1618 Regards, Bharath On 6/13/06, Brian F. G. Bidulock <bidulock@xxxxxxxxxxx> wrote:
Bharath, On Tue, 13 Jun 2006, Bharath wrote: > Packet Size(bytes) 2.4.31(Mbps) 2.6.16(Mbps) > ------------------------------------------------------------------------------------------ ... > 1518 22.09 19.91 The results are identical to within a few percent except for this last value. This is, I presume, being broken into to IP fragments? If so, the more sophisticated IP fragmentation logic in 2.6.16 necessitated by GigE might be responsible for 2.6.16's underperformance at this data point. Also, if so, you should investigate data points with higher degrees of fragmentation. Results are easier to analyze if they are given in number of frames transferred rather than bits per second.
- : send the line "unsubscribe linux-net" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
