[Pharo4] bench for browseAllSendersOf:
Interesting... Pharo2: [SystemNavigation new browseAllSendersOf: #+] bench '0.688 per second.â Pharo3: [SystemNavigation new browseAllSendersOf: #+] bench '0.638 per second.' Pharo4: [SystemNavigation new browseAllSendersOf: #+] bench '1.37 per second.' And for Nautilus: Pharo2: [Smalltalk tools browser open] bench '4.46 per second.' Pharo3: [Smalltalk tools browser open] bench '6.77 per second.' Pharo4: [Smalltalk tools browser open] bench '7.17 per second.â (all with the same VM) Marcus
On 29 Jul 2014, at 09:28, Marcus Denker <marcus.denker@inria.fr> wrote:
Interesting...
Pharo2: [SystemNavigation new browseAllSendersOf: #+] bench '0.688 per second.â
Pharo3: [SystemNavigation new browseAllSendersOf: #+] bench '0.638 per second.â
Here the system got larger, so even though it is slower, it already is faster. Reason: - Simplifications in SystemNavigation and SystemDictionary - Spec improvements - General Cleanups
Pharo4: [SystemNavigation new browseAllSendersOf: #+] bench '1.37 per second.â
Reasons: -> system smaller by around 1-2% -> cleanups in MessageBrowser API, SystemNavigation and Slider/MorphModel -> General Cleanups There is more to be cleaned⦠SystemNavigation is still very arcane, especially related to browsing methods. The idea is to move all the real UI related code to the Tool classes and clean up more. (this can be done step-by-step as time permits). Later we should re-visit all the concepts of âEnviromentâ that we have⦠and unify. -> Smalltalk globals -> SystemNavigation as the mediator for different âSmalltalk globalâ like environments -> RBBrowserEnvironment -> Ring Containers (maybe not related) Marcus
Nice work ! On 29 Jul 2014, at 09:44, Marcus Denker <marcus.denker@inria.fr> wrote:
On 29 Jul 2014, at 09:28, Marcus Denker <marcus.denker@inria.fr> wrote:
Interesting...
Pharo2: [SystemNavigation new browseAllSendersOf: #+] bench '0.688 per second.â
Pharo3: [SystemNavigation new browseAllSendersOf: #+] bench '0.638 per second.â
Here the system got larger, so even though it is slower, it already is faster. Reason: - Simplifications in SystemNavigation and SystemDictionary - Spec improvements - General Cleanups
Pharo4: [SystemNavigation new browseAllSendersOf: #+] bench '1.37 per second.â
Reasons: -> system smaller by around 1-2% -> cleanups in MessageBrowser API, SystemNavigation and Slider/MorphModel -> General Cleanups
There is more to be cleaned⦠SystemNavigation is still very arcane, especially related to browsing methods.
The idea is to move all the real UI related code to the Tool classes and clean up more. (this can be done step-by-step as time permits).
Later we should re-visit all the concepts of âEnviromentâ that we have⦠and unify. -> Smalltalk globals -> SystemNavigation as the mediator for different âSmalltalk globalâ like environments -> RBBrowserEnvironment -> Ring Containers (maybe not related)
Marcus
Yes and it will only improve. Once we unify RBEnv and SystemDictionary and SystemNavigation. I like the discussion we got.
Nice work !
On 29 Jul 2014, at 09:44, Marcus Denker <marcus.denker@inria.fr> wrote:
On 29 Jul 2014, at 09:28, Marcus Denker <marcus.denker@inria.fr> wrote:
Interesting...
Pharo2: [SystemNavigation new browseAllSendersOf: #+] bench '0.688 per second.â
Pharo3: [SystemNavigation new browseAllSendersOf: #+] bench '0.638 per second.â
Here the system got larger, so even though it is slower, it already is faster. Reason: - Simplifications in SystemNavigation and SystemDictionary - Spec improvements - General Cleanups
Pharo4: [SystemNavigation new browseAllSendersOf: #+] bench '1.37 per second.â
Reasons: -> system smaller by around 1-2% -> cleanups in MessageBrowser API, SystemNavigation and Slider/MorphModel -> General Cleanups
There is more to be cleaned⦠SystemNavigation is still very arcane, especially related to browsing methods.
The idea is to move all the real UI related code to the Tool classes and clean up more. (this can be done step-by-step as time permits).
Later we should re-visit all the concepts of âEnviromentâ that we have⦠and unify. -> Smalltalk globals -> SystemNavigation as the mediator for different âSmalltalk globalâ like environments -> RBBrowserEnvironment -> Ring Containers (maybe not related)
Marcus
http://pharoweekly.wordpress.com/2014/07/31/pharo-tools-are-getting-faster/ On 29/7/14 09:28, Marcus Denker wrote:
Interesting...
Pharo2: [SystemNavigation new browseAllSendersOf: #+] bench '0.688 per second.â
Pharo3: [SystemNavigation new browseAllSendersOf: #+] bench '0.638 per second.'
Pharo4: [SystemNavigation new browseAllSendersOf: #+] bench '1.37 per second.'
And for Nautilus:
Pharo2: [Smalltalk tools browser open] bench '4.46 per second.'
Pharo3: [Smalltalk tools browser open] bench '6.77 per second.'
Pharo4: [Smalltalk tools browser open] bench '7.17 per second.â
(all with the same VM)
Marcus
participants (3)
-
Marcus Denker -
stepharo -
Sven Van Caekenberghe