Exploring Visual Comparison of Multivariate Runtime Statistics - - PowerPoint PPT Presentation

▶
exploring visual comparison of multivariate runtime
SMART_READER_LITE
LIVE PREVIEW

Exploring Visual Comparison of Multivariate Runtime Statistics - - PowerPoint PPT Presentation

Apr 09, 2023 104 likes •402 views

9TH SYMPOSIUM ON SOFTWARE PERFORMANCE 2018 Exploring Visual Comparison of Multivariate Runtime Statistics Hagen Tarner, Veit Frick, Martin Pinzger, and Fabian Beck Motivation Performance data is multivariate VIS o ff ers techniques for this

slide-1
SLIDE 1

Exploring Visual Comparison of Multivariate Runtime Statistics

Hagen Tarner, Veit Frick, Martin Pinzger, and Fabian Beck

9TH SYMPOSIUM ON SOFTWARE PERFORMANCE 2018

slide-2
SLIDE 2

Motivation

Performance data is multivariate

2

VIS offers techniques for this let’s apply them

slide-3
SLIDE 3

Use Cases of Execution Comparison

3

Execution environment Input Process Output

Three scenarios:

  • 1. Input changes
  • 2. Code changes
  • 3. Environment changes
slide-4
SLIDE 4

Use Cases of Execution Comparison

4

Execution environment Input Process Output

Three scenarios:

  • 1. Input changes
  • 2. Code changes
  • 3. Environment changes

Inputs can vary in shape, size, etc.

slide-5
SLIDE 5

Use Cases of Execution Comparison

5

Execution environment Input Process Output

Three scenarios:

  • 1. Input changes
  • 2. Code changes
  • 3. Environment changes

The code of an application changes during development.

slide-6
SLIDE 6

Use Cases of Execution Comparison

6

Execution environment Input Process Output

Three scenarios:

  • 1. Input changes
  • 2. Code changes
  • 3. Environment changes

Software runs in different setups and under different conditions.

slide-7
SLIDE 7

juxtaposition, superposition, explicit encoding Grid-based, Glyph-based,
 PCP

Visual Design Space

7

  • Use established comparison techniques
  • Apply multivariate data visualizations to performance metrics
slide-8
SLIDE 8

Visual Design Space

8

(I) Grid-based

50

METHOD #1 METRIC A

34

METRIC B

17

METRIC C

juxtaposition

slide-9
SLIDE 9

Visual Design Space

(I) Grid-based comparison

50 34 17

VERSION I VERSION II METHOD #1 METRIC A METRIC B METRIC C

juxtaposition

9

slide-10
SLIDE 10

Visual Design Space

(I) Grid-based comparison

50 34 17

VERSION I VERSION II METHOD #1 METRIC A METRIC B METRIC C METHOD #2 METHOD #3 METHOD #4 METHOD #5 METHOD #6 METHOD #7 METHOD #8

slide-11
SLIDE 11

Visual Design Space

11

(II) Glyph-based

slide-12
SLIDE 12

Visual Design Space

12

(II) Glyph-based

slide-13
SLIDE 13

Visual Design Space

13

(II) Glyph-based

superposition

slide-14
SLIDE 14

14

Visual Design Space

(II) Glyph-based

slide-15
SLIDE 15

Visual Design Space

15

(III) Parallel Coordinates Plot (PCP)

slide-16
SLIDE 16

Visual Design Space

16

(III) Parallel Coordinates Plot

slide-17
SLIDE 17

Visual Design Space

17

(III) Parallel Coordinates Plot

slide-18
SLIDE 18

Visual Design Space

18

(III) Parallel Coordinates Plot

slide-19
SLIDE 19

Application Examples

19

Apache Commons JProfiler Jupyter Notebook

COMMONS.APACHE.ORG EJ-TECHNOLOGIES.COM JUPYTER.ORG

slide-20
SLIDE 20

Application Examples

20

Apache Commons JProfiler Jupyter Notebook

  • Opensource Java library
  • Already researched by Baltes et al.
  • Contains known performance bugs

Baltes, S., Moseler, O., Beck, F., & Diehl, S. (2015, October). Navigate, understand, communicate: How developers locate performance

  • bugs. In Empirical Software Engineering and Measurement (ESEM), 2015 ACM/IEEE International Symposium on (pp. 1-10). IEEE.
slide-21
SLIDE 21

Application Examples

21

Apache Commons JProfiler Jupyter Notebook

  • State-of-the-art profiler for JVM
  • Yields method-level granularity results
slide-22
SLIDE 22

Application Examples

22

Apache Commons JProfiler Jupyter Notebook

  • Post-mortem static analysis
  • Used to generate static images
slide-23
SLIDE 23

Application Examples

  • Systematic tests of scenario-visualization combinations
  • Scenario-Visualization combinations with highest readability:
  • Input Changes + Radar Charts
  • Code Changes + PCP
  • Environment Changes + grid-based Bar Charts

23

slide-24
SLIDE 24

Application Examples

24

(I) Input Changes Scenario

slide-25
SLIDE 25

25

Application Examples

(II) Code Changes Scenario

slide-26
SLIDE 26

26

Application Examples

(III) Environment Changes Scenario

slide-27
SLIDE 27

Summary

  • First steps in using basic visualizations for comparison of multivariate

data in a software performance context.

  • Three Scenario-Visualization combinations:
  • 1. Input changes + grid-based Radar Charts
  • 2. Code changes + PCP
  • 3. Environment changes + grid-based Bar Charts

27

slide-28
SLIDE 28

Future Work

28

Visual Analytics System, that features

  • multiple coordinated views with a strong interaction concept
  • zooming/filtering + overview
slide-29
SLIDE 29

Hagen Tarner <hagen.tarner@paluno.uni-due.de> https://www.vis.wiwi.uni-due.de

THANK YOU!