Smalltalk › Frameworks & Tools › Moose

[Moose-dev] Mondrian Health Report : good news, we gained a significant speedup

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

10 messages Options

Alexandre Bergel

[Moose-dev] Mondrian Health Report : good news, we gained a significant speedup

Dear List,

You will find below the health report of today. I removed two
important bottlenecks: node translations and computing absolute
bounds. Layouting nodes is more then 100 times faster!
You can now do a "view nodes: (1 to: 20000)" in an easel without
having the time to take a coffee!

Here is the comment of Mondrian-Alexandre_Bergel.352
-=-=-=-=-=-=-=-=-=-=-=-=
Major speed improvement:
- MOGraphElement>>absoluteBounds uses now a cache. This helped
speeding up Mondrian by 35%
- Make MORectangleShape>>display:on: use absoluteBounds instead of
absoluteBoundsFor:, which speeded the UI up of 14%
- optimization when translating nodes (MOGraphElement>>translatedBy:),
we gained a significant speedup when layouting nodes
-=-=-=-=-=-=-=-=-=-=-=-=

Report produced on 2010-01-23T18:17:13+00:00
Benchmark ManyNode (simple rendering of nodes) :
100 nodes => 3 ms
200 nodes => 5 ms
300 nodes => 6 ms
400 nodes => 7 ms
500 nodes => 9 ms
600 nodes => 10 ms
700 nodes => 12 ms
800 nodes => 14 ms
900 nodes => 15 ms
1000 nodes => 17 ms
1600 nodes => 28 ms
Benchmark ManyEdges (simple rendering of edges) :
10 edges => 2 ms
20 edges => 6 ms
30 edges => 12 ms
40 edges => 28 ms
50 edges => 48 ms
60 edges => 74 ms
70 edges => 115 ms
80 edges => 169 ms
90 edges => 254 ms
100 edges => 335 ms
200 edges => 4356 ms
300 edges => 35919 ms
55.54 % of methods are covered
Progress from last time: -0.1 %

Cheers,
Alexandre
--
_,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
Alexandre Bergel http://www.bergel.eu
^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

_______________________________________________
Moose-dev mailing list
[hidden email]
https://www.iam.unibe.ch/mailman/listinfo/moose-dev

Stéphane Ducasse

[Moose-dev] Re: Mondrian Health Report : good news, we gained a significant speedup

coooool

Stef

On Jan 25, 2010, at 8:27 PM, Alexandre Bergel wrote:

> Dear List,
>
> You will find below the health report of today. I removed two important bottlenecks: node translations and computing absolute bounds. Layouting nodes is more then 100 times faster!
> You can now do a "view nodes: (1 to: 20000)" in an easel without having the time to take a coffee!
>
> Here is the comment of Mondrian-Alexandre_Bergel.352
> -=-=-=-=-=-=-=-=-=-=-=-=
> Major speed improvement:
> - MOGraphElement>>absoluteBounds uses now a cache. This helped speeding up Mondrian by 35%
> - Make MORectangleShape>>display:on: use absoluteBounds instead of absoluteBoundsFor:, which speeded the UI up of 14%
> - optimization when translating nodes (MOGraphElement>>translatedBy:), we gained a significant speedup when layouting nodes
> -=-=-=-=-=-=-=-=-=-=-=-=
>
> Report produced on 2010-01-23T18:17:13+00:00
> Benchmark ManyNode (simple rendering of nodes) :
> 100 nodes => 3 ms
> 200 nodes => 5 ms
> 300 nodes => 6 ms
> 400 nodes => 7 ms
> 500 nodes => 9 ms
> 600 nodes => 10 ms
> 700 nodes => 12 ms
> 800 nodes => 14 ms
> 900 nodes => 15 ms
> 1000 nodes => 17 ms
> 1600 nodes => 28 ms
> Benchmark ManyEdges (simple rendering of edges) :
> 10 edges => 2 ms
> 20 edges => 6 ms
> 30 edges => 12 ms
> 40 edges => 28 ms
> 50 edges => 48 ms
> 60 edges => 74 ms
> 70 edges => 115 ms
> 80 edges => 169 ms
> 90 edges => 254 ms
> 100 edges => 335 ms
> 200 edges => 4356 ms
> 300 edges => 35919 ms
> 55.54 % of methods are covered
> Progress from last time: -0.1 %
>
> Cheers,
> Alexandre
> --
> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
> Alexandre Bergel http://www.bergel.eu
> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
>
>
>
>
>
> _______________________________________________
> Moose-dev mailing list
> [hidden email]
> https://www.iam.unibe.ch/mailman/listinfo/moose-dev

_______________________________________________
Moose-dev mailing list
[hidden email]
https://www.iam.unibe.ch/mailman/listinfo/moose-dev

Tudor Girba

[Moose-dev] Re: Mondrian Health Report : good news, we gained a significant speedup

Cool indeed. It even looks like the problem of the bounds while
zooming is solved as well. That's great!

Cheers,
Doru

On 25 Jan 2010, at 21:12, Stéphane Ducasse wrote:

> coooool
>
> Stef
>
> On Jan 25, 2010, at 8:27 PM, Alexandre Bergel wrote:
>
>> Dear List,
>>
>> You will find below the health report of today. I removed two
>> important bottlenecks: node translations and computing absolute
>> bounds. Layouting nodes is more then 100 times faster!
>> You can now do a "view nodes: (1 to: 20000)" in an easel without
>> having the time to take a coffee!
>>
>> Here is the comment of Mondrian-Alexandre_Bergel.352
>> -=-=-=-=-=-=-=-=-=-=-=-=
>> Major speed improvement:
>> - MOGraphElement>>absoluteBounds uses now a cache. This helped
>> speeding up Mondrian by 35%
>> - Make MORectangleShape>>display:on: use absoluteBounds instead of
>> absoluteBoundsFor:, which speeded the UI up of 14%
>> - optimization when translating nodes
>> (MOGraphElement>>translatedBy:), we gained a significant speedup
>> when layouting nodes
>> -=-=-=-=-=-=-=-=-=-=-=-=
>>
>> Report produced on 2010-01-23T18:17:13+00:00
>> Benchmark ManyNode (simple rendering of nodes) :
>> 100 nodes => 3 ms
>> 200 nodes => 5 ms
>> 300 nodes => 6 ms
>> 400 nodes => 7 ms
>> 500 nodes => 9 ms
>> 600 nodes => 10 ms
>> 700 nodes => 12 ms
>> 800 nodes => 14 ms
>> 900 nodes => 15 ms
>> 1000 nodes => 17 ms
>> 1600 nodes => 28 ms
>> Benchmark ManyEdges (simple rendering of edges) :
>> 10 edges => 2 ms
>> 20 edges => 6 ms
>> 30 edges => 12 ms
>> 40 edges => 28 ms
>> 50 edges => 48 ms
>> 60 edges => 74 ms
>> 70 edges => 115 ms
>> 80 edges => 169 ms
>> 90 edges => 254 ms
>> 100 edges => 335 ms
>> 200 edges => 4356 ms
>> 300 edges => 35919 ms
>> 55.54 % of methods are covered
>> Progress from last time: -0.1 %
>>
>> Cheers,
>> Alexandre
>> --
>> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
>> Alexandre Bergel http://www.bergel.eu
>> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
>>
>>
>>
>>
>>
>> _______________________________________________
>> Moose-dev mailing list
>> [hidden email]
>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>
>
> _______________________________________________
> Moose-dev mailing list
> [hidden email]
> https://www.iam.unibe.ch/mailman/listinfo/moose-dev

--
www.tudorgirba.com

"If you interrupt the barber while he is cutting your hair,
you will end up with a messy haircut."

_______________________________________________
Moose-dev mailing list
[hidden email]
https://www.iam.unibe.ch/mailman/listinfo/moose-dev

Simon Denier-3

[Moose-dev] Re: Mondrian Health Report : good news, we gained a significant speedup

In reply to this post by Alexandre Bergel

That's cool

Jannik is checking the changes because DSM does not work completely.

I also noticed a strange thing with colors.

Open a cycletable like
MOCycleTable new cycles: (MOCycleTableTest new setUp); render

Notice how the fill color is desaturated (alpha: 0.5) in the two cells. Now if you fly over the cells with the mouse, the color is fully saturated once redrawn. Any idea where this comes from?

On 25 janv. 2010, at 20:27, Alexandre Bergel wrote:

--
Simon

_______________________________________________
Moose-dev mailing list
[hidden email]
https://www.iam.unibe.ch/mailman/listinfo/moose-dev

Screen shot 2010-01-26 at 14.30.31.png (27K) Download Attachment

Alexandre Bergel

[Moose-dev] Re: Mondrian Health Report : good news, we gained a significant speedup

Color problems are probably not related to my fixes. What is going
wrong with DSM beside the colors?
Best would be to write some tests that fails. This will give me a
starting point where to start.

Cheers,
Alexandre

On 26 Jan 2010, at 10:31, Simon Denier wrote:

>
> That's cool
>
> Jannik is checking the changes because DSM does not work completely.
>
> I also noticed a strange thing with colors.
>
> Open a cycletable like
> MOCycleTable new cycles: (MOCycleTableTest new setUp); render
>
> Notice how the fill color is desaturated (alpha: 0.5) in the two
> cells. Now if you fly over the cells with the mouse, the color is
> fully saturated once redrawn. Any idea where this comes from?
>
> <Screen shot 2010-01-26 at 14.30.31.png>
>
>
>
> On 25 janv. 2010, at 20:27, Alexandre Bergel wrote:
>
>> Dear List,
>>
>> You will find below the health report of today. I removed two
>> important bottlenecks: node translations and computing absolute
>> bounds. Layouting nodes is more then 100 times faster!
>> You can now do a "view nodes: (1 to: 20000)" in an easel without
>> having the time to take a coffee!
>>
>> Here is the comment of Mondrian-Alexandre_Bergel.352
>> -=-=-=-=-=-=-=-=-=-=-=-=
>> Major speed improvement:
>> - MOGraphElement>>absoluteBounds uses now a cache. This helped
>> speeding up Mondrian by 35%
>> - Make MORectangleShape>>display:on: use absoluteBounds instead of
>> absoluteBoundsFor:, which speeded the UI up of 14%
>> - optimization when translating nodes
>> (MOGraphElement>>translatedBy:), we gained a significant speedup
>> when layouting nodes
>> -=-=-=-=-=-=-=-=-=-=-=-=
>>
>> Report produced on 2010-01-23T18:17:13+00:00
>> Benchmark ManyNode (simple rendering of nodes) :
>> 100 nodes => 3 ms
>> 200 nodes => 5 ms
>> 300 nodes => 6 ms
>> 400 nodes => 7 ms
>> 500 nodes => 9 ms
>> 600 nodes => 10 ms
>> 700 nodes => 12 ms
>> 800 nodes => 14 ms
>> 900 nodes => 15 ms
>> 1000 nodes => 17 ms
>> 1600 nodes => 28 ms
>> Benchmark ManyEdges (simple rendering of edges) :
>> 10 edges => 2 ms
>> 20 edges => 6 ms
>> 30 edges => 12 ms
>> 40 edges => 28 ms
>> 50 edges => 48 ms
>> 60 edges => 74 ms
>> 70 edges => 115 ms
>> 80 edges => 169 ms
>> 90 edges => 254 ms
>> 100 edges => 335 ms
>> 200 edges => 4356 ms
>> 300 edges => 35919 ms
>> 55.54 % of methods are covered
>> Progress from last time: -0.1 %
>>
>> Cheers,
>> Alexandre
>> --
>> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
>> Alexandre Bergel http://www.bergel.eu
>> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
>>
>>
>>
>>
>>
>> _______________________________________________
>> Moose-dev mailing list
>> [hidden email]
>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>
> --
> Simon
>
>
>
> _______________________________________________
> Moose-dev mailing list
> [hidden email]
> https://www.iam.unibe.ch/mailman/listinfo/moose-dev

--
_,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
Alexandre Bergel http://www.bergel.eu
^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

_______________________________________________
Moose-dev mailing list
[hidden email]
https://www.iam.unibe.ch/mailman/listinfo/moose-dev

jannik laval

[Moose-dev] Re: Mondrian Health Report : good news, we gained a significant speedup

Hi Alex,

In DSM, I have a problem with edges: they do not appear since version 350 of Mondrian.
I wonder if they are not below shapes.

Cheers,
Jannik

On Jan 26, 2010, at 15:01 , Alexandre Bergel wrote:

> Color problems are probably not related to my fixes. What is going wrong with DSM beside the colors?
> Best would be to write some tests that fails. This will give me a starting point where to start.
>
> Cheers,
> Alexandre
>
>
> On 26 Jan 2010, at 10:31, Simon Denier wrote:
>
>>
>> That's cool
>>
>> Jannik is checking the changes because DSM does not work completely.
>>
>> I also noticed a strange thing with colors.
>>
>> Open a cycletable like
>> MOCycleTable new cycles: (MOCycleTableTest new setUp); render
>>
>> Notice how the fill color is desaturated (alpha: 0.5) in the two cells. Now if you fly over the cells with the mouse, the color is fully saturated once redrawn. Any idea where this comes from?
>>
>> <Screen shot 2010-01-26 at 14.30.31.png>
>>
>>
>>
>> On 25 janv. 2010, at 20:27, Alexandre Bergel wrote:
>>
>>> Dear List,
>>>
>>> You will find below the health report of today. I removed two important bottlenecks: node translations and computing absolute bounds. Layouting nodes is more then 100 times faster!
>>> You can now do a "view nodes: (1 to: 20000)" in an easel without having the time to take a coffee!
>>>
>>> Here is the comment of Mondrian-Alexandre_Bergel.352
>>> -=-=-=-=-=-=-=-=-=-=-=-=
>>> Major speed improvement:
>>> - MOGraphElement>>absoluteBounds uses now a cache. This helped speeding up Mondrian by 35%
>>> - Make MORectangleShape>>display:on: use absoluteBounds instead of absoluteBoundsFor:, which speeded the UI up of 14%
>>> - optimization when translating nodes (MOGraphElement>>translatedBy:), we gained a significant speedup when layouting nodes
>>> -=-=-=-=-=-=-=-=-=-=-=-=
>>>
>>> Report produced on 2010-01-23T18:17:13+00:00
>>> Benchmark ManyNode (simple rendering of nodes) :
>>> 100 nodes => 3 ms
>>> 200 nodes => 5 ms
>>> 300 nodes => 6 ms
>>> 400 nodes => 7 ms
>>> 500 nodes => 9 ms
>>> 600 nodes => 10 ms
>>> 700 nodes => 12 ms
>>> 800 nodes => 14 ms
>>> 900 nodes => 15 ms
>>> 1000 nodes => 17 ms
>>> 1600 nodes => 28 ms
>>> Benchmark ManyEdges (simple rendering of edges) :
>>> 10 edges => 2 ms
>>> 20 edges => 6 ms
>>> 30 edges => 12 ms
>>> 40 edges => 28 ms
>>> 50 edges => 48 ms
>>> 60 edges => 74 ms
>>> 70 edges => 115 ms
>>> 80 edges => 169 ms
>>> 90 edges => 254 ms
>>> 100 edges => 335 ms
>>> 200 edges => 4356 ms
>>> 300 edges => 35919 ms
>>> 55.54 % of methods are covered
>>> Progress from last time: -0.1 %
>>>
>>> Cheers,
>>> Alexandre
>>> --
>>> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
>>> Alexandre Bergel http://www.bergel.eu
>>> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
>>>
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> Moose-dev mailing list
>>> [hidden email]
>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>>
>> --
>> Simon
>>
>>
>>
>> _______________________________________________
>> Moose-dev mailing list
>> [hidden email]
>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>
> --
> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
> Alexandre Bergel http://www.bergel.eu
> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
>
>
>
>
>
> _______________________________________________
> Moose-dev mailing list
> [hidden email]
> https://www.iam.unibe.ch/mailman/listinfo/moose-dev

_______________________________________________
Moose-dev mailing list
[hidden email]
https://www.iam.unibe.ch/mailman/listinfo/moose-dev

Tudor Girba

[Moose-dev] Re: Mondrian Health Report : good news, we gained a significant speedup

It could very well happen. See issue:
http://code.google.com/p/moose-technology/issues/detail?id=112

It would be helpful if you could check by inspecting the graph.

Cheers,
Doru

On 26 Jan 2010, at 15:22, Laval Jannik wrote:

> Hi Alex,
>
> In DSM, I have a problem with edges: they do not appear since
> version 350 of Mondrian.
> I wonder if they are not below shapes.
>
> Cheers,
> Jannik
>
>
> On Jan 26, 2010, at 15:01 , Alexandre Bergel wrote:
>
>> Color problems are probably not related to my fixes. What is going
>> wrong with DSM beside the colors?
>> Best would be to write some tests that fails. This will give me a
>> starting point where to start.
>>
>> Cheers,
>> Alexandre
>>
>>
>> On 26 Jan 2010, at 10:31, Simon Denier wrote:
>>
>>>
>>> That's cool
>>>
>>> Jannik is checking the changes because DSM does not work completely.
>>>
>>> I also noticed a strange thing with colors.
>>>
>>> Open a cycletable like
>>> MOCycleTable new cycles: (MOCycleTableTest new setUp); render
>>>
>>> Notice how the fill color is desaturated (alpha: 0.5) in the two
>>> cells. Now if you fly over the cells with the mouse, the color is
>>> fully saturated once redrawn. Any idea where this comes from?
>>>
>>> <Screen shot 2010-01-26 at 14.30.31.png>
>>>
>>>
>>>
>>> On 25 janv. 2010, at 20:27, Alexandre Bergel wrote:
>>>
>>>> Dear List,
>>>>
>>>> You will find below the health report of today. I removed two
>>>> important bottlenecks: node translations and computing absolute
>>>> bounds. Layouting nodes is more then 100 times faster!
>>>> You can now do a "view nodes: (1 to: 20000)" in an easel without
>>>> having the time to take a coffee!
>>>>
>>>> Here is the comment of Mondrian-Alexandre_Bergel.352
>>>> -=-=-=-=-=-=-=-=-=-=-=-=
>>>> Major speed improvement:
>>>> - MOGraphElement>>absoluteBounds uses now a cache. This helped
>>>> speeding up Mondrian by 35%
>>>> - Make MORectangleShape>>display:on: use absoluteBounds instead
>>>> of absoluteBoundsFor:, which speeded the UI up of 14%
>>>> - optimization when translating nodes
>>>> (MOGraphElement>>translatedBy:), we gained a significant speedup
>>>> when layouting nodes
>>>> -=-=-=-=-=-=-=-=-=-=-=-=
>>>>
>>>> Report produced on 2010-01-23T18:17:13+00:00
>>>> Benchmark ManyNode (simple rendering of nodes) :
>>>> 100 nodes => 3 ms
>>>> 200 nodes => 5 ms
>>>> 300 nodes => 6 ms
>>>> 400 nodes => 7 ms
>>>> 500 nodes => 9 ms
>>>> 600 nodes => 10 ms
>>>> 700 nodes => 12 ms
>>>> 800 nodes => 14 ms
>>>> 900 nodes => 15 ms
>>>> 1000 nodes => 17 ms
>>>> 1600 nodes => 28 ms
>>>> Benchmark ManyEdges (simple rendering of edges) :
>>>> 10 edges => 2 ms
>>>> 20 edges => 6 ms
>>>> 30 edges => 12 ms
>>>> 40 edges => 28 ms
>>>> 50 edges => 48 ms
>>>> 60 edges => 74 ms
>>>> 70 edges => 115 ms
>>>> 80 edges => 169 ms
>>>> 90 edges => 254 ms
>>>> 100 edges => 335 ms
>>>> 200 edges => 4356 ms
>>>> 300 edges => 35919 ms
>>>> 55.54 % of methods are covered
>>>> Progress from last time: -0.1 %
>>>>
>>>> Cheers,
>>>> Alexandre
>>>> --
>>>> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
>>>> Alexandre Bergel http://www.bergel.eu
>>>> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> Moose-dev mailing list
>>>> [hidden email]
>>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>>>
>>> --
>>> Simon
>>>
>>>
>>>
>>> _______________________________________________
>>> Moose-dev mailing list
>>> [hidden email]
>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>>
>> --
>> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
>> Alexandre Bergel http://www.bergel.eu
>> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
>>
>>
>>
>>
>>
>> _______________________________________________
>> Moose-dev mailing list
>> [hidden email]
>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>
>
>
> _______________________________________________
> Moose-dev mailing list
> [hidden email]
> https://www.iam.unibe.ch/mailman/listinfo/moose-dev

--
www.tudorgirba.com

"Every thing has its own flow."

_______________________________________________
Moose-dev mailing list
[hidden email]
https://www.iam.unibe.ch/mailman/listinfo/moose-dev

jannik laval

[Moose-dev] Re: Mondrian Health Report : good news, we gained a significant speedup

Hi,

yes, the graph has edges when I inspect it.
I see the issue is opened since august 2009...

And for DSM the bug appears in the version 350.

Alex, do you have a solution ?
Cheers,

Jannik

On Jan 26, 2010, at 15:28 , Tudor Girba wrote:

> It could very well happen. See issue:
> http://code.google.com/p/moose-technology/issues/detail?id=112
>
> It would be helpful if you could check by inspecting the graph.
>
> Cheers,
> Doru
>
> On 26 Jan 2010, at 15:22, Laval Jannik wrote:
>
>> Hi Alex,
>>
>> In DSM, I have a problem with edges: they do not appear since version 350 of Mondrian.
>> I wonder if they are not below shapes.
>>
>> Cheers,
>> Jannik
>>
>>
>> On Jan 26, 2010, at 15:01 , Alexandre Bergel wrote:
>>
>>> Color problems are probably not related to my fixes. What is going wrong with DSM beside the colors?
>>> Best would be to write some tests that fails. This will give me a starting point where to start.
>>>
>>> Cheers,
>>> Alexandre
>>>
>>>
>>> On 26 Jan 2010, at 10:31, Simon Denier wrote:
>>>
>>>>
>>>> That's cool
>>>>
>>>> Jannik is checking the changes because DSM does not work completely.
>>>>
>>>> I also noticed a strange thing with colors.
>>>>
>>>> Open a cycletable like
>>>> MOCycleTable new cycles: (MOCycleTableTest new setUp); render
>>>>
>>>> Notice how the fill color is desaturated (alpha: 0.5) in the two cells. Now if you fly over the cells with the mouse, the color is fully saturated once redrawn. Any idea where this comes from?
>>>>
>>>> <Screen shot 2010-01-26 at 14.30.31.png>
>>>>
>>>>
>>>>
>>>> On 25 janv. 2010, at 20:27, Alexandre Bergel wrote:
>>>>
>>>>> Dear List,
>>>>>
>>>>> You will find below the health report of today. I removed two important bottlenecks: node translations and computing absolute bounds. Layouting nodes is more then 100 times faster!
>>>>> You can now do a "view nodes: (1 to: 20000)" in an easel without having the time to take a coffee!
>>>>>
>>>>> Here is the comment of Mondrian-Alexandre_Bergel.352
>>>>> -=-=-=-=-=-=-=-=-=-=-=-=
>>>>> Major speed improvement:
>>>>> - MOGraphElement>>absoluteBounds uses now a cache. This helped speeding up Mondrian by 35%
>>>>> - Make MORectangleShape>>display:on: use absoluteBounds instead of absoluteBoundsFor:, which speeded the UI up of 14%
>>>>> - optimization when translating nodes (MOGraphElement>>translatedBy:), we gained a significant speedup when layouting nodes
>>>>> -=-=-=-=-=-=-=-=-=-=-=-=
>>>>>
>>>>> Report produced on 2010-01-23T18:17:13+00:00
>>>>> Benchmark ManyNode (simple rendering of nodes) :
>>>>> 100 nodes => 3 ms
>>>>> 200 nodes => 5 ms
>>>>> 300 nodes => 6 ms
>>>>> 400 nodes => 7 ms
>>>>> 500 nodes => 9 ms
>>>>> 600 nodes => 10 ms
>>>>> 700 nodes => 12 ms
>>>>> 800 nodes => 14 ms
>>>>> 900 nodes => 15 ms
>>>>> 1000 nodes => 17 ms
>>>>> 1600 nodes => 28 ms
>>>>> Benchmark ManyEdges (simple rendering of edges) :
>>>>> 10 edges => 2 ms
>>>>> 20 edges => 6 ms
>>>>> 30 edges => 12 ms
>>>>> 40 edges => 28 ms
>>>>> 50 edges => 48 ms
>>>>> 60 edges => 74 ms
>>>>> 70 edges => 115 ms
>>>>> 80 edges => 169 ms
>>>>> 90 edges => 254 ms
>>>>> 100 edges => 335 ms
>>>>> 200 edges => 4356 ms
>>>>> 300 edges => 35919 ms
>>>>> 55.54 % of methods are covered
>>>>> Progress from last time: -0.1 %
>>>>>
>>>>> Cheers,
>>>>> Alexandre
>>>>> --
>>>>> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
>>>>> Alexandre Bergel http://www.bergel.eu
>>>>> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> Moose-dev mailing list
>>>>> [hidden email]
>>>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>>>>
>>>> --
>>>> Simon
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> Moose-dev mailing list
>>>> [hidden email]
>>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>>>
>>> --
>>> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
>>> Alexandre Bergel http://www.bergel.eu
>>> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
>>>
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> Moose-dev mailing list
>>> [hidden email]
>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>>
>>
>>
>> _______________________________________________
>> Moose-dev mailing list
>> [hidden email]
>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>
> --
> www.tudorgirba.com
>
> "Every thing has its own flow."
>
>
>
> _______________________________________________
> Moose-dev mailing list
> [hidden email]
> https://www.iam.unibe.ch/mailman/listinfo/moose-dev

_______________________________________________
Moose-dev mailing list
[hidden email]
https://www.iam.unibe.ch/mailman/listinfo/moose-dev

Alexandre Bergel

[Moose-dev] Re: Mondrian Health Report : good news, we gained a significant speedup

> Alex, do you have a solution ?

I am not sure to fully understand the issue actually.

Should the two following scripts have the same rendering:
-=-=-=-=-=-=-=-=-=-=-=-=
view node: #a forIt: [ view node: 1 ].
view node: #b forIt: [
view node: 2.
view edges: {1 -> 2} from: #key to: #value ].
-=-=-=-=-=-=-=-=-=-=-=-=
view node: #a forIt: [ view node: 1 ].
view node: #b forIt: [ view node: 2 ].
view edges: {1 -> 2} from: #key to: #value
-=-=-=-=-=-=-=-=-=-=-=-=

Alexandre

>
> On Jan 26, 2010, at 15:28 , Tudor Girba wrote:
>
>> It could very well happen. See issue:
>> http://code.google.com/p/moose-technology/issues/detail?id=112
>>
>> It would be helpful if you could check by inspecting the graph.
>>
>> Cheers,
>> Doru
>>
>> On 26 Jan 2010, at 15:22, Laval Jannik wrote:
>>
>>> Hi Alex,
>>>
>>> In DSM, I have a problem with edges: they do not appear since
>>> version 350 of Mondrian.
>>> I wonder if they are not below shapes.
>>>
>>> Cheers,
>>> Jannik
>>>
>>>
>>> On Jan 26, 2010, at 15:01 , Alexandre Bergel wrote:
>>>
>>>> Color problems are probably not related to my fixes. What is
>>>> going wrong with DSM beside the colors?
>>>> Best would be to write some tests that fails. This will give me a
>>>> starting point where to start.
>>>>
>>>> Cheers,
>>>> Alexandre
>>>>
>>>>
>>>> On 26 Jan 2010, at 10:31, Simon Denier wrote:
>>>>
>>>>>
>>>>> That's cool
>>>>>
>>>>> Jannik is checking the changes because DSM does not work
>>>>> completely.
>>>>>
>>>>> I also noticed a strange thing with colors.
>>>>>
>>>>> Open a cycletable like
>>>>> MOCycleTable new cycles: (MOCycleTableTest new setUp); render
>>>>>
>>>>> Notice how the fill color is desaturated (alpha: 0.5) in the two
>>>>> cells. Now if you fly over the cells with the mouse, the color
>>>>> is fully saturated once redrawn. Any idea where this comes from?
>>>>>
>>>>> <Screen shot 2010-01-26 at 14.30.31.png>
>>>>>
>>>>>
>>>>>
>>>>> On 25 janv. 2010, at 20:27, Alexandre Bergel wrote:
>>>>>
>>>>>> Dear List,
>>>>>>
>>>>>> You will find below the health report of today. I removed two
>>>>>> important bottlenecks: node translations and computing absolute
>>>>>> bounds. Layouting nodes is more then 100 times faster!
>>>>>> You can now do a "view nodes: (1 to: 20000)" in an easel
>>>>>> without having the time to take a coffee!
>>>>>>
>>>>>> Here is the comment of Mondrian-Alexandre_Bergel.352
>>>>>> -=-=-=-=-=-=-=-=-=-=-=-=
>>>>>> Major speed improvement:
>>>>>> - MOGraphElement>>absoluteBounds uses now a cache. This helped
>>>>>> speeding up Mondrian by 35%
>>>>>> - Make MORectangleShape>>display:on: use absoluteBounds instead
>>>>>> of absoluteBoundsFor:, which speeded the UI up of 14%
>>>>>> - optimization when translating nodes
>>>>>> (MOGraphElement>>translatedBy:), we gained a significant
>>>>>> speedup when layouting nodes
>>>>>> -=-=-=-=-=-=-=-=-=-=-=-=
>>>>>>
>>>>>> Report produced on 2010-01-23T18:17:13+00:00
>>>>>> Benchmark ManyNode (simple rendering of nodes) :
>>>>>> 100 nodes => 3 ms
>>>>>> 200 nodes => 5 ms
>>>>>> 300 nodes => 6 ms
>>>>>> 400 nodes => 7 ms
>>>>>> 500 nodes => 9 ms
>>>>>> 600 nodes => 10 ms
>>>>>> 700 nodes => 12 ms
>>>>>> 800 nodes => 14 ms
>>>>>> 900 nodes => 15 ms
>>>>>> 1000 nodes => 17 ms
>>>>>> 1600 nodes => 28 ms
>>>>>> Benchmark ManyEdges (simple rendering of edges) :
>>>>>> 10 edges => 2 ms
>>>>>> 20 edges => 6 ms
>>>>>> 30 edges => 12 ms
>>>>>> 40 edges => 28 ms
>>>>>> 50 edges => 48 ms
>>>>>> 60 edges => 74 ms
>>>>>> 70 edges => 115 ms
>>>>>> 80 edges => 169 ms
>>>>>> 90 edges => 254 ms
>>>>>> 100 edges => 335 ms
>>>>>> 200 edges => 4356 ms
>>>>>> 300 edges => 35919 ms
>>>>>> 55.54 % of methods are covered
>>>>>> Progress from last time: -0.1 %
>>>>>>
>>>>>> Cheers,
>>>>>> Alexandre
>>>>>> --
>>>>>> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
>>>>>> Alexandre Bergel http://www.bergel.eu
>>>>>> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> Moose-dev mailing list
>>>>>> [hidden email]
>>>>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>>>>>
>>>>> --
>>>>> Simon
>>>>>
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> Moose-dev mailing list
>>>>> [hidden email]
>>>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>>>>
>>>> --
>>>> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
>>>> Alexandre Bergel http://www.bergel.eu
>>>> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> Moose-dev mailing list
>>>> [hidden email]
>>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>>>
>>>
>>>
>>> _______________________________________________
>>> Moose-dev mailing list
>>> [hidden email]
>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>>
>> --
>> www.tudorgirba.com
>>
>> "Every thing has its own flow."
>>
>>
>>
>> _______________________________________________
>> Moose-dev mailing list
>> [hidden email]
>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>
>
>
> _______________________________________________
> Moose-dev mailing list
> [hidden email]
> https://www.iam.unibe.ch/mailman/listinfo/moose-dev

Tudor Girba

[Moose-dev] Re: Mondrian Health Report : good news, we gained a significant speedup

Hi Alex,

Yes, they should.

This was indeed a feature that is implemented in the VW version and it
implies a smart ordering of edges. The forces are:
- an edge has to be drawn behind the origin and destination nodes
- the edge must be drawn above both of the origin container and of the
destination container

IO digged a bit in the VW implementation, and this I found the zOrder
function:

Edge>>zOrder
^(self fromFigure zOrder max: self toFigure zOrder) - 1

Node>>zOrder
^self parent zOrder + 2

This function was used to order the elements to draw:

AbstractFigure>>elements
elements isNil ifTrue: [
| sorted |
sorted := SortedCollection sortBlock: [:a :b | a zOrder < b zOrder].
self deepNodeFigures do: [:e | sorted add: e].
self deepEdgeFigures do: [:e | sorted add: e].
elements := sorted asOrderedCollection.
].
^elements

Cheers,
Doru

On 26 Jan 2010, at 16:08, Alexandre Bergel wrote:

>> Alex, do you have a solution ?
>
> I am not sure to fully understand the issue actually.
>
> Should the two following scripts have the same rendering:
> -=-=-=-=-=-=-=-=-=-=-=-=
> view node: #a forIt: [ view node: 1 ].
> view node: #b forIt: [
> view node: 2.
> view edges: {1 -> 2} from: #key to: #value ].
> -=-=-=-=-=-=-=-=-=-=-=-=
> view node: #a forIt: [ view node: 1 ].
> view node: #b forIt: [ view node: 2 ].
> view edges: {1 -> 2} from: #key to: #value
> -=-=-=-=-=-=-=-=-=-=-=-=
>
> Alexandre
>
>
>>
>> On Jan 26, 2010, at 15:28 , Tudor Girba wrote:
>>
>>> It could very well happen. See issue:
>>> http://code.google.com/p/moose-technology/issues/detail?id=112
>>>
>>> It would be helpful if you could check by inspecting the graph.
>>>
>>> Cheers,
>>> Doru
>>>
>>> On 26 Jan 2010, at 15:22, Laval Jannik wrote:
>>>
>>>> Hi Alex,
>>>>
>>>> In DSM, I have a problem with edges: they do not appear since
>>>> version 350 of Mondrian.
>>>> I wonder if they are not below shapes.
>>>>
>>>> Cheers,
>>>> Jannik
>>>>
>>>>
>>>> On Jan 26, 2010, at 15:01 , Alexandre Bergel wrote:
>>>>
>>>>> Color problems are probably not related to my fixes. What is
>>>>> going wrong with DSM beside the colors?
>>>>> Best would be to write some tests that fails. This will give me
>>>>> a starting point where to start.
>>>>>
>>>>> Cheers,
>>>>> Alexandre
>>>>>
>>>>>
>>>>> On 26 Jan 2010, at 10:31, Simon Denier wrote:
>>>>>
>>>>>>
>>>>>> That's cool
>>>>>>
>>>>>> Jannik is checking the changes because DSM does not work
>>>>>> completely.
>>>>>>
>>>>>> I also noticed a strange thing with colors.
>>>>>>
>>>>>> Open a cycletable like
>>>>>> MOCycleTable new cycles: (MOCycleTableTest new setUp); render
>>>>>>
>>>>>> Notice how the fill color is desaturated (alpha: 0.5) in the
>>>>>> two cells. Now if you fly over the cells with the mouse, the
>>>>>> color is fully saturated once redrawn. Any idea where this
>>>>>> comes from?
>>>>>>
>>>>>> <Screen shot 2010-01-26 at 14.30.31.png>
>>>>>>
>>>>>>
>>>>>>
>>>>>> On 25 janv. 2010, at 20:27, Alexandre Bergel wrote:
>>>>>>
>>>>>>> Dear List,
>>>>>>>
>>>>>>> You will find below the health report of today. I removed two
>>>>>>> important bottlenecks: node translations and computing
>>>>>>> absolute bounds. Layouting nodes is more then 100 times faster!
>>>>>>> You can now do a "view nodes: (1 to: 20000)" in an easel
>>>>>>> without having the time to take a coffee!
>>>>>>>
>>>>>>> Here is the comment of Mondrian-Alexandre_Bergel.352
>>>>>>> -=-=-=-=-=-=-=-=-=-=-=-=
>>>>>>> Major speed improvement:
>>>>>>> - MOGraphElement>>absoluteBounds uses now a cache. This
>>>>>>> helped speeding up Mondrian by 35%
>>>>>>> - Make MORectangleShape>>display:on: use absoluteBounds
>>>>>>> instead of absoluteBoundsFor:, which speeded the UI up of 14%
>>>>>>> - optimization when translating nodes
>>>>>>> (MOGraphElement>>translatedBy:), we gained a significant
>>>>>>> speedup when layouting nodes
>>>>>>> -=-=-=-=-=-=-=-=-=-=-=-=
>>>>>>>
>>>>>>> Report produced on 2010-01-23T18:17:13+00:00
>>>>>>> Benchmark ManyNode (simple rendering of nodes) :
>>>>>>> 100 nodes => 3 ms
>>>>>>> 200 nodes => 5 ms
>>>>>>> 300 nodes => 6 ms
>>>>>>> 400 nodes => 7 ms
>>>>>>> 500 nodes => 9 ms
>>>>>>> 600 nodes => 10 ms
>>>>>>> 700 nodes => 12 ms
>>>>>>> 800 nodes => 14 ms
>>>>>>> 900 nodes => 15 ms
>>>>>>> 1000 nodes => 17 ms
>>>>>>> 1600 nodes => 28 ms
>>>>>>> Benchmark ManyEdges (simple rendering of edges) :
>>>>>>> 10 edges => 2 ms
>>>>>>> 20 edges => 6 ms
>>>>>>> 30 edges => 12 ms
>>>>>>> 40 edges => 28 ms
>>>>>>> 50 edges => 48 ms
>>>>>>> 60 edges => 74 ms
>>>>>>> 70 edges => 115 ms
>>>>>>> 80 edges => 169 ms
>>>>>>> 90 edges => 254 ms
>>>>>>> 100 edges => 335 ms
>>>>>>> 200 edges => 4356 ms
>>>>>>> 300 edges => 35919 ms
>>>>>>> 55.54 % of methods are covered
>>>>>>> Progress from last time: -0.1 %
>>>>>>>
>>>>>>> Cheers,
>>>>>>> Alexandre
>>>>>>> --
>>>>>>> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
>>>>>>> Alexandre Bergel http://www.bergel.eu
>>>>>>> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> Moose-dev mailing list
>>>>>>> [hidden email]
>>>>>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>>>>>>
>>>>>> --
>>>>>> Simon
>>>>>>
>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> Moose-dev mailing list
>>>>>> [hidden email]
>>>>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>>>>>
>>>>> --
>>>>> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
>>>>> Alexandre Bergel http://www.bergel.eu
>>>>> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> Moose-dev mailing list
>>>>> [hidden email]
>>>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> Moose-dev mailing list
>>>> [hidden email]
>>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>>>
>>> --
>>> www.tudorgirba.com
>>>
>>> "Every thing has its own flow."
>>>
>>>
>>>
>>> _______________________________________________
>>> Moose-dev mailing list
>>> [hidden email]
>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>>
>>
>>
>> _______________________________________________
>> Moose-dev mailing list
>> [hidden email]
>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev
>
> --
> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
> Alexandre Bergel http://www.bergel.eu
> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
>
>
>
>
>
> _______________________________________________
> Moose-dev mailing list
> [hidden email]
> https://www.iam.unibe.ch/mailman/listinfo/moose-dev

--
www.tudorgirba.com

"It's not what we do that matters most, it's how we do it."

_______________________________________________
Moose-dev mailing list
[hidden email]
https://www.iam.unibe.ch/mailman/listinfo/moose-dev