The network infrastructure is the same( default one )
All 3 computers ( madrid, dublin, prague ) are running kernel 2.4.10

eth0:3Com      eth1:tulip1	eth2:tulip2

tulip driver are compiled as modules in the kernel

I am before madrid. 
one telnet session to prague.
one telnet to lava
one xserver holding netscap session from lava. NOT active while experimenting.



experiment 1:

it's very very slow??? 

[lin@madrid lin]$ ttcp/ttcp -t 192.168.3.2 < ncurses-5.2.tar.gz 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.3.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 1743818 bytes in 100.81 real seconds = 16.89 KB/sec +++
ttcp-t: 213 I/O calls, msec/call = 484.65, calls/sec = 2.11
ttcp-t: 0.0user 0.0sys 1:40real 0% 0i+0d 0maxrss 0+2pf 0+0csw

[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.3.1
ttcp-r: 1743818 bytes in 101.23 real seconds = 16.82 KB/sec +++
ttcp-r: 724 I/O calls, msec/call = 143.18, calls/sec = 7.15
ttcp-r: 0.0user 0.0sys 1:41real 0% 0i+0d 0maxrss 0+1pf 0+0csw


Is it because we compile the tulip driver as module.?????
Also I found on the Two tulip NIC cards involved in this connection, 
the FDX( full duplex ) LED is not on

Redo experiment 1 for a few times

[lin@madrid lin]$ ttcp/ttcp -t  192.168.3.2 < iperf-1.2.tar.gz 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.3.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 177099 bytes in 9.67 real seconds = 17.88 KB/sec +++
ttcp-t: 22 I/O calls, msec/call = 450.26, calls/sec = 2.27
ttcp-t: 0.0user 0.0sys 0:09real 0% 0i+0d 0maxrss 0+2pf 0+0csw

[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.3.1
ttcp-r: 177099 bytes in 10.30 real seconds = 16.79 KB/sec +++
ttcp-r: 75 I/O calls, msec/call = 140.68, calls/sec = 7.28
ttcp-r: 0.0user 0.0sys 0:10real 0% 0i+0d 0maxrss 0+1pf 0+0csw


experiment 1


[lin@madrid lin]$ ttcp/ttcp -t  192.168.3.2 < iperf-1.2.tar.gz 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.3.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 177099 bytes in 23.73 real seconds = 7.29 KB/sec +++
ttcp-t: 22 I/O calls, msec/call = 1104.35, calls/sec = 0.93
ttcp-t: 0.0user 0.0sys 0:23real 0% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.3.1
ttcp-r: 177099 bytes in 24.78 real seconds = 6.98 KB/sec +++
ttcp-r: 121 I/O calls, msec/call = 209.68, calls/sec = 4.88
ttcp-r: 0.0user 0.0sys 0:24real 0% 0i+0d 0maxrss 0+2pf 0+0csw


IT"S VERY VERY SLOW. What's going on here????
To test whether it's because the tulip driver is compiled as module.

reboot madrid and prague to kernel 2.2.10 
in which tulip driver is in kernel directly.

experiment 1

It was very slow. Maybe the problem is : the crossover cable is not plugged well
because after I disconnect the cable and plug back. The FDX( full duplex ) LED
is on again. And the speed becomes very fast. Normal

experiment 1

[lin@madrid lin]$ ttcp/ttcp -t -s 192.168.3.2 < linux-2.4.10.tar  
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.3.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 16777216 bytes in 1.43 real seconds = 11458.22 KB/sec +++
ttcp-t: 2048 I/O calls, msec/call = 0.71, calls/sec = 1432.28
ttcp-t: 0.0user 0.1sys 0:01real 10% 0i+0d 0maxrss 0+2pf 0+0csw

[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.3.1
ttcp-r: 16777216 bytes in 1.44 real seconds = 11375.74 KB/sec +++
ttcp-r: 12285 I/O calls, msec/call = 0.12, calls/sec = 8529.72
ttcp-r: 0.0user 0.1sys 0:01real 7% 0i+0d 0maxrss 0+1pf 0+0csw


experiment 1

very large amount of traffic:  124MB . Do the same experiment for a few times

[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.3.1
ttcp-r: 124835840 bytes in 36.74 real seconds = 3318.31 KB/sec +++
ttcp-r: 91418 I/O calls, msec/call = 0.41, calls/sec = 2488.34
ttcp-r: 0.0user 0.3sys 0:36real 1% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@madrid lin]$ ttcp/ttcp -t 192.168.3.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.3.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 124835840 bytes in 36.73 real seconds = 3319.23 KB/sec +++
ttcp-t: 15239 I/O calls, msec/call = 2.47, calls/sec = 414.91
ttcp-t: 0.0user 28.6sys 0:36real 77% 0i+0d 0maxrss 30478+2pf 0+0csw


experiment 1

[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.3.1
ttcp-r: 124835840 bytes in 10.66 real seconds = 11435.82 KB/sec +++
ttcp-r: 91364 I/O calls, msec/call = 0.12, calls/sec = 8570.44
ttcp-r: 0.0user 0.5sys 0:10real 5% 0i+0d 0maxrss 0+2pf 0+0csw

[lin@madrid lin]$ ttcp/ttcp -t 192.168.3.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.3.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 124835840 bytes in 10.65 real seconds = 11441.87 KB/sec +++
ttcp-t: 15239 I/O calls, msec/call = 0.72, calls/sec = 1430.26
ttcp-t: 0.0user 0.4sys 0:10real 4% 0i+0d 0maxrss 0+2pf 0+0csw


Experiment 1

[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.3.1
ttcp-r: 124835840 bytes in 10.66 real seconds = 11440.71 KB/sec +++
ttcp-r: 91373 I/O calls, msec/call = 0.12, calls/sec = 8574.95
ttcp-r: 0.0user 0.4sys 0:10real 4% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@madrid lin]$ ttcp/ttcp -t 192.168.3.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.3.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 124835840 bytes in 10.65 real seconds = 11445.72 KB/sec +++
ttcp-t: 15239 I/O calls, msec/call = 0.72, calls/sec = 1430.74
ttcp-t: 0.0user 1.0sys 0:10real 10% 0i+0d 0maxrss 0+2pf 0+0csw

Experiment 1

[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.3.1
ttcp-r: 124835840 bytes in 10.66 real seconds = 11440.75 KB/sec +++
ttcp-r: 91338 I/O calls, msec/call = 0.12, calls/sec = 8571.70
ttcp-r: 0.0user 0.4sys 0:10real 4% 0i+0d 0maxrss 0+2pf 0+0csw

[lin@madrid lin]$ ttcp/ttcp -t 192.168.3.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.3.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 124835840 bytes in 10.65 real seconds = 11445.68 KB/sec +++
ttcp-t: 15239 I/O calls, msec/call = 0.72, calls/sec = 1430.73
ttcp-t: 0.0user 0.4sys 0:10real 3% 0i+0d 0maxrss 0+2pf 0+0csw


Experiment 1

[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.3.1
ttcp-r: 124835840 bytes in 10.66 real seconds = 11438.61 KB/sec +++
ttcp-r: 91362 I/O calls, msec/call = 0.12, calls/sec = 8572.34
ttcp-r: 0.0user 0.5sys 0:10real 5% 0i+0d 0maxrss 0+2pf 0+0csw

[lin@madrid lin]$ ttcp/ttcp -t 192.168.3.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.3.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 124835840 bytes in 10.65 real seconds = 11444.08 KB/sec +++
ttcp-t: 15239 I/O calls, msec/call = 0.72, calls/sec = 1430.53
ttcp-t: 0.0user 1.4sys 0:10real 13% 0i+0d 0maxrss 0+2pf 0+0csw

From the above 5 experiment 1, we see the speed is normal about 11400KB/sec


NOW reboot prague and madrid to kernel 2.4.10 with tulip driver as module in it

and redo the experiment 1 for a few times

experiment 1

[lin@madrid lin]$ ttcp/ttcp -t -s 192.168.3.2 < nttcp-1.47.tar.gz 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.3.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 16777216 bytes in 1.42 real seconds = 11537.44 KB/sec +++
ttcp-t: 2048 I/O calls, msec/call = 0.71, calls/sec = 1442.18
ttcp-t: 0.0user 0.0sys 0:01real 3% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.3.1
ttcp-r: 16777216 bytes in 1.43 real seconds = 11482.41 KB/sec +++
ttcp-r: 11587 I/O calls, msec/call = 0.13, calls/sec = 8120.53
ttcp-r: 0.0user 0.0sys 0:01real 5% 0i+0d 0maxrss 0+1pf 0+0csw




experiment 1


[lin@madrid lin]$ ttcp/ttcp -t -s 192.168.3.2 < nttcp-1.47.tar.gz 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.3.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 16777216 bytes in 1.42 real seconds = 11538.40 KB/sec +++
ttcp-t: 2048 I/O calls, msec/call = 0.71, calls/sec = 1442.30
ttcp-t: 0.0user 0.0sys 0:01real 2% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.3.1
ttcp-r: 16777216 bytes in 1.43 real seconds = 11482.94 KB/sec +++
ttcp-r: 11587 I/O calls, msec/call = 0.13, calls/sec = 8120.90
ttcp-r: 0.0user 0.0sys 0:01real 3% 0i+0d 0maxrss 0+1pf 0+0csw


experiment 1

[lin@madrid lin]$ ttcp/ttcp -t -s 192.168.3.2 < nttcp-1.47.tar.gz 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.3.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 16777216 bytes in 1.42 real seconds = 11547.45 KB/sec +++
ttcp-t: 2048 I/O calls, msec/call = 0.71, calls/sec = 1443.43
ttcp-t: 0.0user 0.0sys 0:01real 1% 0i+0d 0maxrss 0+2pf 0+0csw

[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.3.1
ttcp-r: 16777216 bytes in 1.43 real seconds = 11480.62 KB/sec +++
ttcp-r: 11588 I/O calls, msec/call = 0.13, calls/sec = 8119.96
ttcp-r: 0.0user 0.2sys 0:01real 21% 0i+0d 0maxrss 0+1pf 0+0csw


experiment 1

USING very large traffic: 124MB, do it for a few times

experiment 1

[lin@madrid lin]$ ttcp/ttcp -t 192.168.3.2 < linux-2.4.10.tar     
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.3.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 124835840 bytes in 10.68 real seconds = 11420.05 KB/sec +++
ttcp-t: 15239 I/O calls, msec/call = 0.72, calls/sec = 1427.53
ttcp-t: 0.0user 1.1sys 0:10real 10% 0i+0d 0maxrss 0+2pf 0+0csw

[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.3.1
ttcp-r: 124835840 bytes in 10.68 real seconds = 11411.63 KB/sec +++
ttcp-r: 86363 I/O calls, msec/call = 0.13, calls/sec = 8084.18
ttcp-r: 0.0user 1.3sys 0:10real 12% 0i+0d 0maxrss 0+1pf 0+0csw


experiment 1

[lin@madrid lin]$ ttcp/ttcp -t 192.168.3.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.3.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 124835840 bytes in 10.61 real seconds = 11494.00 KB/sec +++
ttcp-t: 15239 I/O calls, msec/call = 0.71, calls/sec = 1436.77
ttcp-t: 0.0user 0.9sys 0:10real 8% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.3.1
ttcp-r: 124835840 bytes in 10.61 real seconds = 11486.18 KB/sec +++
ttcp-r: 86214 I/O calls, msec/call = 0.13, calls/sec = 8122.95
ttcp-r: 0.0user 1.4sys 0:10real 14% 0i+0d 0maxrss 0+1pf 0+0csw


experiment 1

[lin@madrid lin]$ ttcp/ttcp -t 192.168.3.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.3.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 124835840 bytes in 10.61 real seconds = 11495.34 KB/sec +++
ttcp-t: 15239 I/O calls, msec/call = 0.71, calls/sec = 1436.94
ttcp-t: 0.0user 0.8sys 0:10real 8% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.3.1
ttcp-r: 124835840 bytes in 10.61 real seconds = 11488.15 KB/sec +++
ttcp-r: 86212 I/O calls, msec/call = 0.13, calls/sec = 8124.16
ttcp-r: 0.0user 1.2sys 0:10real 12% 0i+0d 0maxrss 0+1pf 0+0csw


From the above experiments on experiment 1, we see that 
kernel 2.4.10 with tulip as module compiled it runs normal







experiment 2:

[lin@madrid lin]$ ttcp/ttcp -t 192.168.2.2 < ncurses-5.2.tar.gz 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 1743818 bytes in 0.54 real seconds = 3135.94 KB/sec +++
ttcp-t: 213 I/O calls, msec/call = 2.61, calls/sec = 392.23
ttcp-t: 0.0user 0.0sys 0:00real 1% 0i+0d 0maxrss 0+2pf 0+0cs

[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 1743818 bytes in 0.75 real seconds = 2258.26 KB/sec +++
ttcp-r: 726 I/O calls, msec/call = 1.06, calls/sec = 962.74
ttcp-r: 0.0user 0.0sys 0:00real 2% 0i+0d 0maxrss 0+2pf 0+0csw

WEIRD enough. the routed connection speed is much much faster than directly connected with crossover cable??????


Redo experiment 2 for 3 times, using different amount of traffic

experiment 2:

[lin@madrid lin]$ ttcp/ttcp -t -s 192.168.2.2 < ncurses-5.2.tar.gz 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 16777216 bytes in 3.23 real seconds = 5069.56 KB/sec +++
ttcp-t: 2048 I/O calls, msec/call = 1.62, calls/sec = 633.70
ttcp-t: 0.0user 0.0sys 0:03real 0% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 16777216 bytes in 3.24 real seconds = 5063.46 KB/sec +++
ttcp-r: 6956 I/O calls, msec/call = 0.48, calls/sec = 2149.75
ttcp-r: 0.0user 0.2sys 0:03real 9% 0i+0d 0maxrss 0+2pf 0+0csw



experiment 2:


[lin@madrid lin]$ ttcp/ttcp -t -s 192.168.2.2 < ncurses-5.2.tar.gz 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 16777216 bytes in 3.43 real seconds = 4781.05 KB/sec +++
ttcp-t: 2048 I/O calls, msec/call = 1.71, calls/sec = 597.63
ttcp-t: 0.0user 0.1sys 0:03real 4% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 16777216 bytes in 3.43 real seconds = 4776.18 KB/sec +++
ttcp-r: 6965 I/O calls, msec/call = 0.50, calls/sec = 2030.40
ttcp-r: 0.0user 0.2sys 0:03real 5% 0i+0d 0maxrss 0+2pf 0+0csw


experiment 2:


[lin@madrid lin]$ ttcp/ttcp -t  192.168.2.2 < linux-2.4.10.tar.gz 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 28338216 bytes in 5.74 real seconds = 4817.11 KB/sec +++
ttcp-t: 3460 I/O calls, msec/call = 1.70, calls/sec = 602.27
ttcp-t: 0.0user 0.1sys 0:05real 3% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 28338216 bytes in 5.75 real seconds = 4813.63 KB/sec +++
ttcp-r: 11833 I/O calls, msec/call = 0.50, calls/sec = 2058.23
ttcp-r: 0.0user 0.2sys 0:05real 4% 0i+0d 0maxrss 0+2pf 0+0csw


experiment 2
[lin@madrid lin]$ ttcp/ttcp -t  192.168.2.2 < linux-2.4.10.tar.gz 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 28338216 bytes in 5.44 real seconds = 5083.01 KB/sec +++
ttcp-t: 3460 I/O calls, msec/call = 1.61, calls/sec = 635.51
ttcp-t: 0.0user 0.1sys 0:05real 3% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 28338216 bytes in 5.45 real seconds = 5079.36 KB/sec +++
ttcp-r: 11748 I/O calls, msec/call = 0.47, calls/sec = 2156.26
ttcp-r: 0.0user 0.3sys 0:05real 6% 0i+0d 0maxrss 0+2pf 0+0csw




RE do experiment 2. dublin, prague, madrid runs kernel 2.4.10 with tulip driver
compiled as module.


experiment 2


[lin@madrid lin]$ ttcp/ttcp -t -s 192.168.2.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 16777216 bytes in 3.42 real seconds = 4787.79 KB/sec +++
ttcp-t: 2048 I/O calls, msec/call = 1.71, calls/sec = 598.47
ttcp-t: 0.0user 0.0sys 0:03real 0% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 16777216 bytes in 3.43 real seconds = 4779.23 KB/sec +++
ttcp-r: 6967 I/O calls, msec/call = 0.50, calls/sec = 2032.28
ttcp-r: 0.0user 0.4sys 0:03real 14% 0i+0d 0maxrss 0+2pf 0+0csw



experiment 2

[lin@madrid lin]$ ttcp/ttcp -t -s 192.168.2.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 16777216 bytes in 3.23 real seconds = 5072.74 KB/sec +++
ttcp-t: 2048 I/O calls, msec/call = 1.61, calls/sec = 634.09
ttcp-t: 0.0user 0.0sys 0:03real 1% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 16777216 bytes in 3.24 real seconds = 5064.24 KB/sec +++
ttcp-r: 6953 I/O calls, msec/call = 0.48, calls/sec = 2149.15
ttcp-r: 0.0user 0.3sys 0:03real 12% 0i+0d 0maxrss 0+2pf 0+0csw


experiment 2

[lin@madrid lin]$ ttcp/ttcp -t -s 192.168.2.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 16777216 bytes in 3.22 real seconds = 5086.84 KB/sec +++
ttcp-t: 2048 I/O calls, msec/call = 1.61, calls/sec = 635.86
ttcp-t: 0.0user 0.0sys 0:03real 0% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 16777216 bytes in 3.22 real seconds = 5084.59 KB/sec +++
ttcp-r: 6960 I/O calls, msec/call = 0.47, calls/sec = 2159.96
ttcp-r: 0.0user 0.0sys 0:03real 2% 0i+0d 0maxrss 0+2pf 0+0csw



experiment 2

[lin@madrid lin]$ ttcp/ttcp -t -s 192.168.2.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 16777216 bytes in 3.23 real seconds = 5070.57 KB/sec +++
ttcp-t: 2048 I/O calls, msec/call = 1.62, calls/sec = 633.82
ttcp-t: 0.0user 0.0sys 0:03real 0% 0i+0d 0maxrss 0+2pf 0+0csw

[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 16777216 bytes in 3.44 real seconds = 4764.81 KB/sec +++
ttcp-r: 6959 I/O calls, msec/call = 0.51, calls/sec = 2023.82
ttcp-r: 0.0user 0.1sys 0:03real 4% 0i+0d 0maxrss 0+2pf 0+0csw


experiment 2

USING very larg amount of traffic: 124MB


experiment 2

[lin@madrid lin]$ ttcp/ttcp -t 192.168.2.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 124835840 bytes in 23.98 real seconds = 5084.83 KB/sec +++
ttcp-t: 15239 I/O calls, msec/call = 1.61, calls/sec = 635.61
ttcp-t: 0.0user 0.7sys 0:23real 3% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 124835840 bytes in 24.18 real seconds = 5041.02 KB/sec +++
ttcp-r: 51792 I/O calls, msec/call = 0.48, calls/sec = 2141.62
ttcp-r: 0.0user 1.7sys 0:24real 7% 0i+0d 0maxrss 0+2pf 0+0csw



experiment 2

[lin@madrid lin]$ ttcp/ttcp -t 192.168.2.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 124835840 bytes in 24.17 real seconds = 5044.24 KB/sec +++
ttcp-t: 15239 I/O calls, msec/call = 1.62, calls/sec = 630.54
ttcp-t: 0.0user 0.8sys 0:24real 3% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 124835840 bytes in 24.17 real seconds = 5043.17 KB/sec +++
ttcp-r: 51789 I/O calls, msec/call = 0.48, calls/sec = 2142.41
ttcp-r: 0.0user 2.1sys 0:24real 9% 0i+0d 0maxrss 0+2pf 0+0csw


experiment 2

[lin@madrid lin]$ ttcp/ttcp -t 192.168.2.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 124835840 bytes in 24.18 real seconds = 5042.22 KB/sec +++
ttcp-t: 15239 I/O calls, msec/call = 1.62, calls/sec = 630.29
ttcp-t: 0.0user 0.5sys 0:24real 2% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 124835840 bytes in 24.39 real seconds = 4998.50 KB/sec +++
ttcp-r: 51767 I/O calls, msec/call = 0.48, calls/sec = 2122.53
ttcp-r: 0.0user 1.6sys 0:24real 6% 0i+0d 0maxrss 0+2pf 0+0csw


From the above experiments of experiment 2. we can see that the speed of 
routed connection is stable. about 5000KB/sec  NORMAL


NOW reboot dublin to kernel 2.4.10 with Tulip driver directly compiled in it

And redo the experment 2 for a few times

Experiment 2

[lin@madrid lin]$ ttcp/ttcp -t -s 192.168.2.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 16777216 bytes in 1.42 real seconds = 11549.29 KB/sec +++
ttcp-t: 2048 I/O calls, msec/call = 0.71, calls/sec = 1443.66
ttcp-t: 0.0user 0.1sys 0:01real 7% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 16777216 bytes in 1.43 real seconds = 11476.33 KB/sec +++
ttcp-r: 11588 I/O calls, msec/call = 0.13, calls/sec = 8116.93
ttcp-r: 0.0user 0.2sys 0:01real 20% 0i+0d 0maxrss 0+1pf 0+0csw



Experiment 2

[lin@madrid lin]$ ttcp/ttcp -t -s 192.168.2.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 16777216 bytes in 1.42 real seconds = 11529.73 KB/sec +++
ttcp-t: 2048 I/O calls, msec/call = 0.71, calls/sec = 1441.22
ttcp-t: 0.0user 0.0sys 0:01real 2% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 16777216 bytes in 1.43 real seconds = 11478.58 KB/sec +++
ttcp-r: 11588 I/O calls, msec/call = 0.13, calls/sec = 8118.52
ttcp-r: 0.0user 0.0sys 0:01real 5% 0i+0d 0maxrss 0+1pf 0+0csw


Experiment 2


[lin@madrid lin]$ ttcp/ttcp -t -s 192.168.2.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 16777216 bytes in 1.42 real seconds = 11536.11 KB/sec +++
ttcp-t: 2048 I/O calls, msec/call = 0.71, calls/sec = 1442.01
ttcp-t: 0.0user 0.0sys 0:01real 0% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 16777216 bytes in 1.43 real seconds = 11474.88 KB/sec +++
ttcp-r: 11588 I/O calls, msec/call = 0.13, calls/sec = 8115.90
ttcp-r: 0.0user 0.3sys 0:01real 22% 0i+0d 0maxrss 0+1pf 0+0csw


Experiment 2


[lin@madrid lin]$ ttcp/ttcp -t -s 192.168.2.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 16777216 bytes in 1.42 real seconds = 11535.76 KB/sec +++
ttcp-t: 2048 I/O calls, msec/call = 0.71, calls/sec = 1441.97
ttcp-t: 0.0user 0.0sys 0:01real 0% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 16777216 bytes in 1.43 real seconds = 11479.03 KB/sec +++
ttcp-r: 11590 I/O calls, msec/call = 0.13, calls/sec = 8120.24
ttcp-r: 0.0user 0.0sys 0:01real 6% 0i+0d 0maxrss 0+1pf 0+0csw


Experiment 2

[lin@madrid lin]$ ttcp/ttcp -t -s 192.168.2.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 16777216 bytes in 1.42 real seconds = 11524.24 KB/sec +++
ttcp-t: 2048 I/O calls, msec/call = 0.71, calls/sec = 1440.53
ttcp-t: 0.0user 0.0sys 0:01real 2% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 16777216 bytes in 1.43 real seconds = 11479.54 KB/sec +++
ttcp-r: 11588 I/O calls, msec/call = 0.13, calls/sec = 8119.20
ttcp-r: 0.0user 0.2sys 0:01real 16% 0i+0d 0maxrss 0+1pf 0+0csw


Experiment 2.

USING very large amount of traffic: 124MB

re-do experiments for a few times

experiment 2

[lin@madrid lin]$ ttcp/ttcp -t  192.168.2.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 124835840 bytes in 10.60 real seconds = 11497.34 KB/sec +++
ttcp-t: 15239 I/O calls, msec/call = 0.71, calls/sec = 1437.19
ttcp-t: 0.0user 0.8sys 0:10real 7% 0i+0d 0maxrss 0+2pf 0+0csw

[lin@prague ttcp]$ ttcp -r -s 
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 124835840 bytes in 10.61 real seconds = 11488.55 KB/sec +++
ttcp-r: 86210 I/O calls, msec/call = 0.13, calls/sec = 8124.26
ttcp-r: 0.0user 1.3sys 0:10real 12% 0i+0d 0maxrss 0+2pf 0+0csw



Experiment 2

[lin@madrid lin]$ ttcp/ttcp -t  192.168.2.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 124835840 bytes in 10.61 real seconds = 11494.47 KB/sec +++
ttcp-t: 15239 I/O calls, msec/call = 0.71, calls/sec = 1436.83
ttcp-t: 0.0user 0.8sys 0:10real 8% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 124835840 bytes in 10.61 real seconds = 11485.59 KB/sec +++
ttcp-r: 86212 I/O calls, msec/call = 0.13, calls/sec = 8122.35
ttcp-r: 0.0user 1.3sys 0:10real 13% 0i+0d 0maxrss 0+1pf 0+0csw


Experiment 2

[lin@madrid lin]$ ttcp/ttcp -t  192.168.2.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 124835840 bytes in 10.60 real seconds = 11497.35 KB/sec +++
ttcp-t: 15239 I/O calls, msec/call = 0.71, calls/sec = 1437.19
ttcp-t: 0.0user 0.8sys 0:10real 8% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 124835840 bytes in 10.61 real seconds = 11488.09 KB/sec +++
ttcp-r: 86214 I/O calls, msec/call = 0.13, calls/sec = 8124.31
ttcp-r: 0.0user 1.2sys 0:10real 12% 0i+0d 0maxrss 0+1pf 0+0csw



Experiment 2

[lin@madrid lin]$ ttcp/ttcp -t  192.168.2.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 124835840 bytes in 10.60 real seconds = 11496.64 KB/sec +++
ttcp-t: 15239 I/O calls, msec/call = 0.71, calls/sec = 1437.10
ttcp-t: 0.0user 0.8sys 0:10real 8% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 124835840 bytes in 10.61 real seconds = 11487.90 KB/sec +++
ttcp-r: 86213 I/O calls, msec/call = 0.13, calls/sec = 8124.08
ttcp-r: 0.0user 1.3sys 0:10real 13% 0i+0d 0maxrss 0+1pf 0+0csw





From the above experiments on experiment 2, we achieve a full throughput 
on routed connection. 11500KB/sec.  This is what we expect.

We can conclude that:

If 3Com card is not significantly faster than Tulip chip
	( NOTE: with Tulip driver directly compiled to kernel 2.4.10. the order
		of eth? is changed. tulip1:eth0 tulip2:eth1  3Com:eth2 )
with NIC driver directly compiled in kernel, the routing speed improves 
significantly (about 2:1)

To remove the possibility that the improvement of routing speed is caused 3Com
card, we configure the NIC card so that the NIC cards involved in routing are
2 tulip chips.

Do the experiments of experiment 2 again:

experiment 2 

[lin@madrid lin]$ ttcp/ttcp -t  -s 192.168.2.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 16777216 bytes in 1.42 real seconds = 11530.66 KB/sec +++
ttcp-t: 2048 I/O calls, msec/call = 0.71, calls/sec = 1441.33
ttcp-t: 0.0user 0.0sys 0:01real 0% 0i+0d 0maxrss 0+2pf 0+0csw

[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 16777216 bytes in 1.43 real seconds = 11474.57 KB/sec +++
ttcp-r: 11590 I/O calls, msec/call = 0.13, calls/sec = 8117.08
ttcp-r: 0.0user 0.1sys 0:01real 12% 0i+0d 0maxrss 0+1pf 0+0csw


experiment 2


[lin@madrid lin]$ ttcp/ttcp -t  -s 192.168.2.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 16777216 bytes in 1.42 real seconds = 11531.83 KB/sec +++
ttcp-t: 2048 I/O calls, msec/call = 0.71, calls/sec = 1441.48
ttcp-t: 0.0user 0.0sys 0:01real 2% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 16777216 bytes in 1.43 real seconds = 11478.03 KB/sec +++
ttcp-r: 11588 I/O calls, msec/call = 0.13, calls/sec = 8118.13
ttcp-r: 0.0user 0.1sys 0:01real 9% 0i+0d 0maxrss 0+1pf 0+0csw



Experiment 2


[lin@madrid lin]$ ttcp/ttcp -t  -s 192.168.2.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 16777216 bytes in 1.42 real seconds = 11528.91 KB/sec +++
ttcp-t: 2048 I/O calls, msec/call = 0.71, calls/sec = 1441.11
ttcp-t: 0.0user 0.0sys 0:01real 2% 0i+0d 0maxrss 0+2pf 0+0csw

[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 16777216 bytes in 1.43 real seconds = 11481.41 KB/sec +++
ttcp-r: 11588 I/O calls, msec/call = 0.13, calls/sec = 8120.52
ttcp-r: 0.0user 0.1sys 0:01real 10% 0i+0d 0maxrss 0+1pf 0+0csw


Experiment 2


[lin@madrid lin]$ ttcp/ttcp -t  -s 192.168.2.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 16777216 bytes in 1.42 real seconds = 11535.87 KB/sec +++
ttcp-t: 2048 I/O calls, msec/call = 0.71, calls/sec = 1441.98
ttcp-t: 0.0user 0.0sys 0:01real 4% 0i+0d 0maxrss 0+2pf 0+0csw

[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 16777216 bytes in 1.43 real seconds = 11480.12 KB/sec +++
ttcp-r: 11587 I/O calls, msec/call = 0.13, calls/sec = 8118.90
ttcp-r: 0.0user 0.1sys 0:01real 7% 0i+0d 0maxrss 0+1pf 0+0csw



Experiment 2

USING very large amount of traffic: 124MB


experiment 2

[lin@madrid lin]$ ttcp/ttcp -t  192.168.2.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 124835840 bytes in 10.61 real seconds = 11494.03 KB/sec +++
ttcp-t: 15239 I/O calls, msec/call = 0.71, calls/sec = 1436.78
ttcp-t: 0.0user 0.8sys 0:10real 8% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 124835840 bytes in 10.61 real seconds = 11487.71 KB/sec +++
ttcp-r: 86212 I/O calls, msec/call = 0.13, calls/sec = 8123.85
ttcp-r: 0.0user 1.3sys 0:10real 13% 0i+0d 0maxrss 0+1pf 0+0csw


Experiment 2

[lin@madrid lin]$ ttcp/ttcp -t  192.168.2.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 124835840 bytes in 10.60 real seconds = 11495.83 KB/sec +++
ttcp-t: 15239 I/O calls, msec/call = 0.71, calls/sec = 1437.00
ttcp-t: 0.0user 0.9sys 0:10real 8% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 124835840 bytes in 10.61 real seconds = 11488.64 KB/sec +++
ttcp-r: 86204 I/O calls, msec/call = 0.13, calls/sec = 8123.75
ttcp-r: 0.0user 1.2sys 0:10real 12% 0i+0d 0maxrss 0+2pf 0+0csw



Experiment 2


[lin@madrid lin]$ ttcp/ttcp -t  192.168.2.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 124835840 bytes in 10.60 real seconds = 11497.26 KB/sec +++
ttcp-t: 15239 I/O calls, msec/call = 0.71, calls/sec = 1437.18
ttcp-t: 0.0user 0.8sys 0:10real 8% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 124835840 bytes in 10.61 real seconds = 11488.69 KB/sec +++
ttcp-r: 86206 I/O calls, msec/call = 0.13, calls/sec = 8123.97
ttcp-r: 0.0user 1.3sys 0:10real 13% 0i+0d 0maxrss 0+2pf 0+0csw


Experiment 2


[lin@madrid lin]$ ttcp/ttcp -t  192.168.2.2 < linux-2.4.10.tar 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 192.168.2.2
ttcp-t: socket
ttcp-t: connect
ttcp-t: 124835840 bytes in 10.60 real seconds = 11495.93 KB/sec +++
ttcp-t: 15239 I/O calls, msec/call = 0.71, calls/sec = 1437.01
ttcp-t: 0.0user 0.9sys 0:10real 8% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 192.168.1.2
ttcp-r: 124835840 bytes in 10.61 real seconds = 11488.18 KB/sec +++
ttcp-r: 86212 I/O calls, msec/call = 0.13, calls/sec = 8124.18
ttcp-r: 0.0user 1.2sys 0:10real 11% 0i+0d 0maxrss 0+1pf 0+0csw



From the above experiments we do see that the speed is still about 11500KB/sec

So the improvement is not significantly because of NIC speed difference between
3Com card and tulip chips

So, we can draw the conclusion:

With NIC card driver directly compiled into kernel, the routing speed improve 
a lot, compared the case that NIC card driver compiled and loaded as module.







Experiment 3

[lin@madrid lin]$ ttcp/ttcp -t -s 132.177.8.29 < iperf-1.2.tar.gz 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 132.177.8.29
ttcp-t: socket
ttcp-t: connect
ttcp-t: 16777216 bytes in 1.42 real seconds = 11549.41 KB/sec +++
ttcp-t: 2048 I/O calls, msec/call = 0.71, calls/sec = 1443.68
ttcp-t: 0.0user 0.0sys 0:01real 2% 0i+0d 0maxrss 0+2pf 0+0csw

[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 132.177.8.27
ttcp-r: 16777216 bytes in 1.43 real seconds = 11484.46 KB/sec +++
ttcp-r: 11588 I/O calls, msec/call = 0.13, calls/sec = 8122.68
ttcp-r: 0.0user 0.2sys 0:01real 15% 0i+0d 0maxrss 0+1pf 0+0csw


as fast as expected, this connection use 3COM cards and the driver is build in 
kernel directly

test experiment 3 for a few times


experiment 3


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 132.177.8.27
ttcp-r: 28338216 bytes in 2.41 real seconds = 11492.06 KB/sec +++
ttcp-r: 19572 I/O calls, msec/call = 0.13, calls/sec = 8127.57
ttcp-r: 0.0user 0.3sys 0:02real 14% 0i+0d 0maxrss 0+1pf 0+0csw


[lin@madrid lin]$ ttcp/ttcp -t 132.177.8.29 < linux-2.4.10.tar.gz 
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 132.177.8.29
ttcp-t: socket
ttcp-t: connect
ttcp-t: 28338216 bytes in 2.40 real seconds = 11523.10 KB/sec +++
ttcp-t: 3460 I/O calls, msec/call = 0.71, calls/sec = 1440.70
ttcp-t: 0.0user 0.3sys 0:02real 12% 0i+0d 0maxrss 0+2pf 0+0csw


experiment 3


[lin@madrid lin]$ ttcp/ttcp -t 132.177.8.29 < linux-2.4.10.tar    
ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> 132.177.8.29
ttcp-t: socket
ttcp-t: connect
ttcp-t: 124835840 bytes in 10.61 real seconds = 11492.66 KB/sec +++
ttcp-t: 15239 I/O calls, msec/call = 0.71, calls/sec = 1436.61
ttcp-t: 0.0user 1.1sys 0:10real 10% 0i+0d 0maxrss 0+2pf 0+0csw


[lin@prague ttcp]$ ttcp -r -s
ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
ttcp-r: socket
ttcp-r: accept from 132.177.8.27
ttcp-r: 124835840 bytes in 10.61 real seconds = 11486.64 KB/sec +++
ttcp-r: 86215 I/O calls, msec/call = 0.13, calls/sec = 8123.37
ttcp-r: 0.0user 1.6sys 0:10real 15% 0i+0d 0maxrss 0+1pf 0+0csw


experiment 3 is NORMAL