Discussion:
measuring throughput with netperf
(too old to reply)
Mark
2015-01-23 11:57:22 UTC
Permalink
Hello,

I'm doing benchmark test of Linux server using netperf. According to
information at
http://wiki.networksecuritytoolkit.org/nstwiki/index.php/LAN_Ethernet_Maximum_Rates,_Generation,_Capturing_%26_Monitoring#Gigabit_Ethernet_Using_UDP
the theoretical max throughput rate for 1514-bytes packets is 117.35 Mbp/s,
however I'm
getting 957 Mbp/s for UDP traffic of 1514 packets. I generate the traffic
with netperf.

I am not sure how this can be possible. Does netperf return udp/tcp or
Ethernet throughput?
--
Mark



---
This email has been checked for viruses by Avast antivirus software.
http://www.avast.com
Stephen
2015-01-23 20:38:49 UTC
Permalink
On Fri, 23 Jan 2015 06:57:22 -0500, "Mark"
Post by Mark
Hello,
I'm doing benchmark test of Linux server using netperf. According to
information at
http://wiki.networksecuritytoolkit.org/nstwiki/index.php/LAN_Ethernet_Maximum_Rates,_Generation,_Capturing_%26_Monitoring#Gigabit_Ethernet_Using_UDP
the theoretical max throughput rate for 1514-bytes packets is 117.35 Mbp/s,
Bytes / sec not bits.........
Post by Mark
however I'm
getting 957 Mbp/s for UDP traffic of 1514 packets. I generate the traffic
with netperf.
I am not sure how this can be possible. Does netperf return udp/tcp or
Ethernet throughput?
Stephen Hope ***@xyzworld.com
Replace xyz with ntl to reply
glen herrmannsfeldt
2015-01-23 23:22:24 UTC
Permalink
Post by Stephen
On Fri, 23 Jan 2015 06:57:22 -0500, "Mark"
(snip)
Post by Stephen
Post by Mark
the theoretical max throughput rate for 1514-bytes packets
is 117.35 Mbp/s,
Bytes / sec not bits.........
Post by Mark
getting 957 Mbp/s for UDP traffic of 1514 packets. I generate the traffic
with netperf.
OK, but 117.35*8 is 938.8, so 957 is still more.

But you have to send a really large amount of data to get good
statistics on this, to avoid problems at the beginning and end.

If you start the clock when the first packet is received, and end
with the last one, you are one packet off.

(note the rates that ftp has reported for years.)

-- glen
Rick Jones
2015-01-23 18:24:29 UTC
Permalink
There can be bits versus Bytes confusion when information sources are
not consistently using 'b' for bits and 'B' for bytes.

rick jones
--
No need to believe in either side, or any side. There is no cause.
There's only yourself. The belief is in your own precision. - Joubert
these opinions are mine, all mine; HP might not want them anyway... :)
feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH...
Rick Jones
2015-01-23 16:59:56 UTC
Permalink
Post by Mark
I'm doing benchmark test of Linux server using netperf. According to
information at
http://wiki.networksecuritytoolkit.org/nstwiki/index.php/LAN_Ethernet_Maximum_Rates,_Generation,_Capturing_%26_Monitoring#Gigabit_Ethernet_Using_UDP
the theoretical max throughput rate for 1514-bytes packets is 117.35
Mbp/s, however I'm getting 957 Mbp/s for UDP traffic of 1514
packets. I generate the traffic with netperf.
I am not sure how this can be possible. Does netperf return udp/tcp
or Ethernet throughput?
Netperf returns "to the user" throughput. By default it reports
megabits per second (10^6 bits per second).

Following the link, I see them talking about 119,635,891 Bytes/s,
which would be 957,087,128 bits per second or 957.1 megabits/s.

happy benchmarking,

rick jones
--
It is not a question of half full or empty - the glass has a leak.
The real question is "Can it be patched?"
these opinions are mine, all mine; HP might not want them anyway... :)
feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH...
Loading...