Maybe doing a SpaceTally new printSpaceAnalysis and a Smalltalk garbageCollect.
before each test....and then compare the different files.. cheers mariano On Tue, Nov 9, 2010 at 12:46 PM, Max Leske <[email protected]> wrote: > Um... I should be able to get a visualisation by running somthing like this > right (I only installed it, didn't subclass anything)? > > MemoryProfiler > viewProfiling: [ FSDiskFilesystem current working / foo ]. > > I get a window but there's nothing in it... > Also, is selecting the expression in the block above, right clicking and > selecting "memory profile it" equivalent to sending #viewProfiling: ? > > Reading your paper I get the idea that this should work, when looking at > the instructions on the moose site it rather seems that I need to subclass > MemoryProfiler to actually profile anything. > > Cheers, > Max > > > > On 09.11.2010, at 15:02, Alexandre Bergel wrote: > > > Ok, but please, provide feedback. > > > > Alexandre > > > > > > On 9 Nov 2010, at 10:40, Max Leske wrote: > > > >> Sounds like what I just needed! I'll try it out. > >> > >> Cheers, > >> Max > >> > >> > >> On 09.11.2010, at 12:42, Alexandre Bergel wrote: > >> > >>> Hi Guido, > >>> > >>> We wrote a profiler to keep track of memory consumption. > >>> > >>> http://bergel.eu/download/papers/Berg10g-MemoryProfiling.pdf > >>> > >>> It will be presented at Smalltalks next week. > >>> > >>> Cheers, > >>> Alexandre > >>> > >>> On 8 Nov 2010, at 23:47, Guido Stepken wrote: > >>> > >>>> E.g.: Three different solutions for a Euler problem: > >>>> > >>>> (1 to: 999) inject: 0 into: [:sum :i | (i \\ 3 = 0) | (i \\ 5 = 0) > >>>> ifTrue: [sum + i] > >>>> ifFalse: [sum]] > >>>> > >>>> > >>>> (1 to: 999) detectSum: [:i | (i \\ 3 = 0) | (i \\ 5 = 0) ifTrue: [i] > ifFalse: [0]] > >>>> > >>>> (1 to: 999) detectSum: [:i | i * (#(0 0 1 0 1 1 0 0 1 1 0 1 0 0 1) > atWrap: i)] > >>>> > >>>> How do i measure the exact memory footprint of each solution? > >>>> > >>>> Pharo claims having reduced memory footprint at many places. Can that > be proofed, e.g. with one comparable demo app, once run in old squeak, once > in Pharo 1.1/1.2? > >>>> > >>>> And no, MessageTally spyOn: does not do right! > >>>> > >>>> Tnx in advance, Guido Stepken > >>>> > >>>> > >>> > >>> -- > >>> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: > >>> Alexandre Bergel http://www.bergel.eu > >>> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;. > >>> > >>> > >>> > >>> > >>> > >>> > >> > >> > > > > -- > > _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: > > Alexandre Bergel http://www.bergel.eu > > ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;. > > > > > > > > > > > > > > >
