performance - how to compare different versions of Matlab on different machines similar to bench() functions? -

so matlab has set of tools check if processor , setting of computer enough , how ranks popular settings. range of matrix based tests(lu,fft,ode,sparse,2d,3d)

however, according matlab here: http://www.mathworks.com/help/matlab/ref/bench.html note: "a benchmark intended compare performance of 1 particular version of matlab on different machines. not offer direct comparisons between different versions of matlab because tasks , problem sizes change version version."

so there set of problems or similar package can compare between machines use different versions of matlab? or best way compare different versions of matlab on different machine? there universal benchmark can provide bigger picture?

the goals focused on vector based numerical operation no graphic , no graphical animation. university cloud based matlab version different our desktop version or personal version on laptop.

you may rename module bench , run in different matlab versions.this undo changes mathworks apparently applies bench.m.

we have done matlab versions our customers use:

procedure:

-1- copy "c:\program files\matlab\r2013a\toolbox\matlab\general\bench.m" "mybench.m"

-2- run "mybench.m" subsequentlty in 2012a (32 bit), 2013a (64 bit), 2016a (64 bit). runs 5 repetitions of benchmark.

date:

august 15, 2016

results:

matlab 2012a (32 bit)

lu fft ode sparse 2-d 3-d

0.0525    0.1022    0.0524    0.1203    0.2312    0.6745  0.0602    0.1075    0.0520    0.1276    0.2285    0.6913  0.0646    0.1032    0.0523    0.1119    0.2280    0.6711  0.0546    0.1056    0.0518    0.1120    0.2288    0.6992  0.0539    0.1025    0.0520    0.1165    0.2374    0.7007 

matlab 2013a (64 bit)

lu fft ode sparse 2-d 3-d

0.0473    0.0729    0.0545    0.0948    0.2308    0.6913  0.0494    0.0778    0.0560    0.1083    0.2502    0.6734  0.0585    0.0767    0.0567    0.0944    0.2184    0.6804  0.0539    0.0754    0.0552    0.0951    0.2188    0.6858  0.0520    0.0761    0.0557    0.0948    0.2257    0.6798 

matlab 2014b (64 bit)

lu fft ode sparse 2-d 3-d

0.0389    0.0670    0.0729    0.0886    0.7948    0.5959  0.0472    0.0691    0.0739    0.0927    0.7582    0.6114  0.0427    0.0675    0.0737    0.0996    0.7288    0.6271  0.0391    0.0682    0.0740    0.0978    0.7316    0.6303  0.0388    0.0717    0.0751    0.0955    0.7096    0.6469 

matlab 2016a (64 bit)

lu fft ode sparse 2-d 3-d

0.0712    0.1004    0.0542    0.0888    0.7549    0.6586  0.0823    0.1003    0.0527    0.0892    0.7273    0.6345  0.0816    0.0977    0.0523    0.0885    0.7547    0.6280  0.0842    0.1008    0.0542    0.0954    0.6634    0.6355  0.0837    0.1000    0.0514    0.0919    0.6641    0.6440