Object-centric profiling:Advanced Visualizations to Tame Wild

Program Execution

Vanessa Peña, Juan Pablo Sandoval, Pablo Estefo, Alexandre Bergel

Object Profile & University of Chile

2

Execution profiling with Kai

Problem:

Traditional code profilers are driven by the method stack, discarding the notion of sending messages

Why the problem is important:

How to answer to “Is there a slow function that is called too often?”

Solution:

An intuitive visual representation of the execution that visually compare the time spent and the number of executions

3

54.8% {11501ms} MOCanvas>>drawOn: 54.8% {11501ms} MORoot(MONode)>>displayOn: 30.9% {6485ms} MONode>>displayOn: | 18.1% {3799ms} MOEdge>>displayOn: ... | 8.4% {1763ms} MOEdge>>displayOn: | | 8.0% {1679ms} MOStraightLineShape>>display:on: | | 2.6% {546ms} FormCanvas>>line:to:width:color: ... 23.4% {4911ms} MOEdge>>displayOn: ...

Execution sampling profiler

4

Structural profiling blueprint

legend for methods

(color)

#different

receiver

# executions

execution

time

5

Behavioral profiling blueprint

6

legend for methods

gray = side

effect

yellow = no side effect

# executions

execution time

m2m1

invokes m2 and m3

m1 m3

DEMO

Structural profiling blueprint

legend for methods

(color)

#different

receiver

# executions

execution

time

bounds

7

legend for methods

gray = side

effect

yellow = no side effect

# executions

execution time

m2m1

invokes m2 and m3

m1 m3

Behavioral profiling blueprint

bounds8

Code of the bounds method

MOGraphElement>>bounds "Answer the bounds of the receiver."

| basicBounds |

self shapeBoundsAt: self shape ifPresent: [ :b | ^ b ].

basicBounds := shape computeBoundsFor: self. self shapeBoundsAt: self shape put: basicBounds.

^ basicBounds

9

Memoizing

MOGraphElement>>bounds "Answer the bounds of the receiver."

| basicBounds | boundsCache ifNotNil: [ ^ boundsCache ]. self shapeBoundsAt: self shape ifPresent: [ :b | ^ b ].

basicBounds := shape computeBoundsFor: self. self shapeBoundsAt: self shape put: basicBounds.

^ boundsCache := basicBounds

10

A

B

C

Upgrading MOGraphElement>>bounds

11

A

B

C

Upgrading MOGraphElement>>bounds

43%speedup

12

B

A

Upgrading MOGraphElement>>bounds

13

AB C D

cached

absoluteBoundsmake display:on:

call absoluteBoundsinstead of absoluteBoundsFor:

A'

C'

B'

C'

14

AB C D

cached

absoluteBoundsmake display:on:

call absoluteBoundsinstead of absoluteBoundsFor:

A'

C'

B'

C'

15

AB C D

cached

absoluteBoundsmake display:on:

call absoluteBoundsinstead of absoluteBoundsFor:

A'

C'

B'

C'

16

Test coverage with Hapao

Problem:

Traditional code coverage tools have a binary view of the world

Why the problem is important:

Which method should you test first in order to increase the coverage?

Is my code well covered or not?

Solution:

An intuitive visual representation of a qualitative assessment of the coverage

17

Test blueprint

cd

# calling methods

complexity # executions

Legend for methods (inner boxes)

red = not executedblue = abstract

invocation on self

C1

C2

18DEMO

Successive improvement

Version 2.227.27%

Version 2.354.54%

Version 2.487.71%

Version 2.5100%

19

4 patterns

20

Moose-Test-Core.13 Moose-Core.313

Moose-Test-Core.48 Moose-Core.326

21.42%

56.86%

73.58%

68.25%

0%

36.78%

100%

96.66%

64.55%

100%

100%

100%

21

Reducing code complexity

Version 1.58.1Coverage: 40.57%

Version 1.58.9Coverage: 60.60%

22

Reducing code complexity

Version 2.10

Version 2.1723

Visualizing data with Roassal

Roassal is an agile and interactive visualization engine

Visualizing Incidents at the Fire Department of NY

Supported Platforms

XMLTokenizer>>nextName

XMLNestedStreamReader>>atEnd

node color

∆ number of executions

∆ numberof

messages

source code red: change black: do not change

green: # messages < than before

light red: # messages > than before and # executions <= than before

red: # messages > than before and # executions > than before

white: # messages = than before

yellow: this method did not exist before

Legend for methods

m1

m2m1 invokes m2

Multidimentional profiling

Differences between tests

Conclusion

Little innovation in the tools we commonly use

Profilers, debuggers, testing tools have not significantly evolved

Fantastic opportunities for improvement

Kai, Hapao and Roassal are just a beginning

Feel free to provide feedback on our tool

30

Spy @ Cincom StoreSpy @ SqueakSourceRoassal @ ...

ObjectProfile.com

Moose-Test-Core.13 Moose-Core.313

Moose-Test-Core.48 Moose-Core.326

21.42%

56.86%

73.58%

68.25%

0%

36.78%

100%

96.66%

64.55%

100%

100%

100%

Thanks to ESUG Chris T Cincom All of you guys!