Number of threads should not be larger than the number of cores (for now don't use taskset and set the number of threads to a number smaller than the number of idle real cores). Also make sure the cpus you are running on are not hyperthreaded.
Amin On Mar 7, 2013 1:39 AM, "Mona Sinha" <[email protected]> wrote: > Hi, > It didnt helped. > Firstly I installed NOX varity through following steps: > > sudo apt-get update > wget -c ' > http://sourceforge.net/projects/boost/files/boost/1.50.0/boost_1_50_0.tar.bz2/download > ' > tar xf download > cd boost_1_50_0 > ./bootstrap.sh > ./b2 install > ./configure --prefix=/usr --with-boost-libdir=/usr/local/lib > --enable-ndebug > make clean > make -j > Run nox through > taskset -c 0-2 ./nox_core -i ptcp:0.0.0.0:6633 -v switch -t 15 > > Cbench installtion > > Downloaded cbench > ./configure > make > make install > Run through : > cbench -c localhost -M 10000 -m 1000 -s 16 -l 10 -t > > Output: > > root@mona:/home/mona/cbench/openflow# cbench -c localhost -M 10000 -m > 1000 -s 16 -l 10 -t > cbench: controller benchmarking tool > running in mode 'throughput' > connecting to controller at localhost:6633 > faking 16 switches offset 1 :: 10 tests each; 1000 ms per test > with 10000 unique source MACs per switch > learning destination mac addresses before the test > starting test with 0 ms delay after features_reply > ignoring first 1 "warmup" and last 0 "cooldown" loops > connection delay of 0ms per 1 switch(es) > debugging info is off > 15:02:47.237 16 switches: flows/sec: 199 199 1 199 199 199 199 > 199 199 199 199 199 199 199 199 199 total = 1.930342 per ms > 15:02:48.377 16 switches: flows/sec: 800 800 798 800 800 819 800 > 819 0 819 0 819 0 0 0 800 total = 8.541053 per ms > 15:02:49.602 16 switches: flows/sec: 0 819 0 0 0 0 819 0 819 0 > 819 0 0 0 0 0 total = 2.913248 per ms > 15:02:50.912 16 switches: flows/sec: 819 0 819 819 819 119 0 119 > 0 119 0 119 0 819 119 938 total = 4.648458 per ms > 15:02:52.622 16 switches: flows/sec: 0 0 0 0 0 0 0 0 119 0 0 > 0 0 119 0 0 total = 0.147932 per ms > 15:02:53.723 16 switches: flows/sec: 119 0 0 0 119 0 0 0 0 0 0 > 0 0 0 0 0 total = 0.237737 per ms > 15:02:54.897 16 switches: flows/sec: 0 119 119 119 0 0 119 1843 > 0 1843 0 0 0 0 0 0 total = 3.874086 per ms > 15:02:56.041 16 switches: flows/sec: 0 0 0 0 0 1843 0 0 0 0 > 701 1843 0 0 1024 1843 total = 6.948589 per ms > 15:02:57.680 16 switches: flows/sec: 0 0 0 0 0 0 0 0 1843 0 0 > 0 1844 0 0 0 total = 2.397420 per ms > 15:02:59.235 16 switches: flows/sec: 2048 0 0 0 2048 0 0 0 0 0 > 0 0 0 0 0 0 total = 2.814428 per ms > RESULT: 16 switches 9 tests min/max/avg/stdev = > 147.93/8541.05/3613.66/2636.42 responses/s > > Please suggest some solution. > > > > On Thu, Mar 7, 2013 at 1:51 PM, Amin Tootoonchian <[email protected]>wrote: > >> Most probably you are benchmarking a debug build. Pass --enable-ndebug >> to the configure script. >> >> Amin >> >> On Wed, Mar 6, 2013 at 10:56 PM, Mona Sinha <[email protected]> >> wrote: >> > Hi, >> > Nox-destiny throughput is too low in my setup.Can you please >> suggest >> > me how to enhance it? >> > It is showing just 400-500 flows/sec on cbench. >> > >
