<tt><font size=2>>> Ok, this is a good complementary check. Here
is the result of the first<br>
>> test I do concerning service demand:<br>
>> single test:<br>
>> netperf -l 60 -H 10.0.17.200 -i 10,2 -I 99,5 -t TCP_STREAM -cC<br>
>> TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 10.0.17.200<br>
>> (10.0.17.200) port 0 AF_INET : +/-2.5% @ 99% conf. : demo<br>
>> Recv Send Send Utilization Service Demand<br>
>> Socket Socket Message Elapsed Send Recv Send Recv<br>
>> Size Size Size Time Throughput local remote local remote<br>
>> bytes bytes bytes secs. 10^6bits/s % S % S us/KB us/KB<br>
>><br>
>> 87380 65536 65536 60.01 2075.12 *5.19 9.32 0.409 3.027 *<br>
>><br>
>><br>
>> two-at-once test:<br>
>> netperf -l 60 -H 10.0.17.200 -i 10,2 -I 99,5 -t TCP_STREAM -cC
&<br>
>> netperf -l 60 -H 10.0.17.200 -i 10,2 -I 99,5 -t TCP_STREAM -cC<br>
>><br>
>> TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 10.0.17.200<br>
>> (10.0.17.200) port 0 AF_INET : +/-2.5% @ 99% conf. : demo<br>
>> TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 10.0.17.200<br>
>> (10.0.17.200) port 0 AF_INET : +/-2.5% @ 99% conf. : demo<br>
>><br>
>> Recv Send Send Utilization Service Demand<br>
>> Socket Socket Message Elapsed Send Recv Send Recv<br>
>> Size Size Size Time Throughput local remote local remote<br>
>> bytes bytes bytes secs. 10^6bits/s % S % S us/KB us/KB<br>
>><br>
>> 87380 65536 65536 60.06 1760.01 *9.21 13.11 0.858 4.801 *<br>
>><br>
>> Recv Send Send Utilization Service Demand<br>
>> Socket Socket Message Elapsed Send Recv Send Recv<br>
>> Size Size Size Time Throughput local remote local remote<br>
>> bytes bytes bytes secs. 10^6bits/s % S % S us/KB us/KB<br>
>><br>
>> 87380 65536 65536 60.06 1750.31 *9.21 13.14 0.861 4.730*<br>
>><br>
>> This looks good isn't it?<br>
><br>
>That definitely looks better. However, you need to change the
"-i 10,2" <br>
>to "-i 10" You must not let netperf finish "early"
if it hits the <br>
>confidence interval, so the min and max iterations must be the same.<br>
><br>
>rick<br>
</font></tt>
<br><tt><font size=2>Ok, now I understand!</font></tt>
<br><tt><font size=2>It seems more complicated for observing synchronization
with UDP stream test. Here is an example of what I try:</font></tt>
<br>
<br><font size=2 face="Courier New">netperf -l 60 -H 10.0.17.200 -t UDP_STREAM
-cC -i 10 -I 99,5 -- -m 8900 & </font>
<br><font size=2 face="Courier New">netperf -l 60 -H 10.0.17.200 -t UDP_STREAM
-cC -i 10 -I 99,5 -- -m 8900 & </font>
<br><font size=2 face="Courier New">netperf -l 60 -H 10.0.17.200 -t UDP_STREAM
-cC -i 10 -I 99,5 -- -m 8900</font>
<br>
<br><font size=2 face="Courier New">UDP UNIDIRECTIONAL SEND TEST from 0.0.0.0
(0.0.0.0) port 0 AF_INET to 10.0.17.200 (10.0.17.200) port 0 AF_INET :
+/-2.5% @ 99% conf. : demo</font>
<br><font size=2 face="Courier New">!!! WARNING</font>
<br><font size=2 face="Courier New">!!! Desired confidence was not achieved
within the specified iterations.</font>
<br><font size=2 face="Courier New">!!! This implies that there was variability
in the test environment that</font>
<br><font size=2 face="Courier New">!!! must be investigated before going
further.</font>
<br><font size=2 face="Courier New">!!! Confidence intervals: Throughput
: 2.3%</font>
<br><font size=2 face="Courier New">!!!
Local CPU util : 1.9%</font>
<br><font size=2 face="Courier New">!!!
Remote CPU util : 46.7%</font>
<br>
<br><font size=2 face="Courier New">Socket Message Elapsed
Messages
CPU Service</font>
<br><font size=2 face="Courier New">Size Size
Time Okay Errors Throughput Util
Demand</font>
<br><font size=2 face="Courier New">bytes bytes secs
# #
10^6bits/sec % SS us/KB</font>
<br>
<br><font size=2 face="Courier New">112640 8900 60.00
2766708 0 3283.2
64.21 3.411 </font>
<br><font size=2 face="Courier New">108544
60.00 2599612 3084.9
64.83 13.768</font>
<br>
<br><font size=2 face="Courier New">UDP UNIDIRECTIONAL SEND TEST from 0.0.0.0
(0.0.0.0) port 0 AF_INET to 10.0.17.200 (10.0.17.200) port 0 AF_INET :
+/-2.5% @ 99% conf. : demo</font>
<br><font size=2 face="Courier New">!!! WARNING</font>
<br><font size=2 face="Courier New">!!! Desired confidence was not achieved
within the specified iterations.</font>
<br><font size=2 face="Courier New">!!! This implies that there was variability
in the test environment that</font>
<br><font size=2 face="Courier New">!!! must be investigated before going
further.</font>
<br><font size=2 face="Courier New">!!! Confidence intervals: Throughput
: 1.9%</font>
<br><font size=2 face="Courier New">!!!
Local CPU util : 1.8%</font>
<br><font size=2 face="Courier New">!!!
Remote CPU util : 53.1%</font>
<br>
<br><font size=2 face="Courier New">Socket Message Elapsed
Messages
CPU Service</font>
<br><font size=2 face="Courier New">Size Size
Time Okay Errors Throughput Util
Demand</font>
<br><font size=2 face="Courier New">bytes bytes secs
# #
10^6bits/sec % SS us/KB</font>
<br>
<br><font size=2 face="Courier New">112640 8900 60.00
2809476 0 3333.9
64.24 3.384 </font>
<br><font size=2 face="Courier New">108544
60.00 2620965 3110.2
60.55 12.770</font>
<br>
<br><font size=2 face="Courier New">UDP UNIDIRECTIONAL SEND TEST from 0.0.0.0
(0.0.0.0) port 0 AF_INET to 10.0.17.200 (10.0.17.200) port 0 AF_INET :
+/-2.5% @ 99% conf. : demo</font>
<br><font size=2 face="Courier New">!!! WARNING</font>
<br><font size=2 face="Courier New">!!! Desired confidence was not achieved
within the specified iterations.</font>
<br><font size=2 face="Courier New">!!! This implies that there was variability
in the test environment that</font>
<br><font size=2 face="Courier New">!!! must be investigated before going
further.</font>
<br><font size=2 face="Courier New">!!! Confidence intervals: Throughput
: 1.9%</font>
<br><font size=2 face="Courier New">!!!
Local CPU util : 1.8%</font>
<br><font size=2 face="Courier New">!!!
Remote CPU util : 48.5%</font>
<br>
<br><font size=2 face="Courier New">Socket Message Elapsed
Messages
CPU Service</font>
<br><font size=2 face="Courier New">Size Size
Time Okay Errors Throughput Util
Demand</font>
<br><font size=2 face="Courier New">bytes bytes secs
# #
10^6bits/sec % SS us/KB</font>
<br>
<br><font size=2 face="Courier New">112640 8900 60.00
2790054 0 3310.9
64.23 3.399 </font>
<br><font size=2 face="Courier New">108544
60.00 2609576 3096.7
59.35 12.560</font>
<br>
<br><tt><font size=2>Due to the protocol, the CPU utilization varies a
lot. Could I only check the local CPU utilization in order to make sure
that tests are running simultaneously? Do I need both to be near to consider
UDP test as synchronized? I also consider the service demand which is clearly
higher than with a single instance test.</font></tt>
<br>
<br><tt><font size=2>I plan to try TCP sendfile also to compare results
with TCP. Does the size of the sent file matter?</font></tt>
<br>
<br><tt><font size=2>A not related question: I try to use the -n option
to choose the number of core used by my remote computer but this seems
not to be considered.</font></tt>
<br><tt><font size=2>Does Netperf/Netserver consider this options or does
it ignores it if it can determine itself the number of cpu?</font></tt>
<br><tt><font size=2>So, I will use CPU bind in order to do it manually.</font></tt>
<br>
<br><tt><font size=2>Thank you and have a nice day.</font></tt>
<br>
<br><tt><font size=2>Simon Duboué</font></tt>
<br>
<br>
<br>