![Page 1: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/1.jpg)
1
Response Time Analysis for
G-EDF and G-DM Scheduling
of Sporadic DAG-Tasks
with Arbitrary Deadline
Andrea Parri, Alessandro Biondi and Mauro Marinoni
Scuola Superiore Sant’Anna – Pisa, Italy
![Page 2: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/2.jpg)
2
Introduction
Multicore revolution
New parallel programming models
for expressing parallel computational
activities
Intel TBB
![Page 3: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/3.jpg)
3
Introduction
Big Data
Novel programming models
based on the Map-Reduce
paradigm that relies on parallel
processing
![Page 4: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/4.jpg)
4
Introduction
JUNIPER EU Project – supported this work
Goal: enable application development with
performance guarantees required for real-time
exploitation of large streaming data sources and
stored data;
Case-study: applications for credit cards.
![Page 5: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/5.jpg)
5
DAG-Task
Task model for expressing parallel computations
with precedence constraints
A task is described with a Directed Acyclic Graph (DAG)
𝑒1
𝑒2
𝑒3
𝑒4
𝑒5
𝑒6
𝑒7
X
![Page 6: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/6.jpg)
6
DAG-Task
𝑒1
𝑒2
𝑒3
𝑒4
𝑒5
𝑒6
𝑒7
Vertex – sequential
computation with
WCET 𝑒𝑖
![Page 7: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/7.jpg)
7
DAG-Task
𝑒1
𝑒2
𝑒3
𝑒4
𝑒5
𝑒6
𝑒7
Edge – precedence
constraint among two
computational activities
![Page 8: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/8.jpg)
8
DAG-Task
𝑒1
𝑒2
𝑒3
𝑒4
𝑒5
𝑒6
𝑒7
Note: this model allows to
express parallelism
![Page 9: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/9.jpg)
9
DAG-Task
𝑒1
𝑒2
𝑒3
𝑒4
𝑒5
𝑒6
𝑒7
Release of a DAG-Task 𝜏𝑖
All the vertices are released simultaneously but it
can be that not all of them are enabled due to
precedence constrains
𝜏𝑖
![Page 10: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/10.jpg)
10
Sporadic DAG-Task
𝑒1
𝑒2
𝑒3
𝑒4
𝑒5
𝑒6
𝑒7
DAG-Task 𝜏𝑖
Released with a minimum inter-arrival time 𝑻𝒊
Each vertex must complete within a deadline 𝑫𝒊
𝜏𝑖
![Page 11: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/11.jpg)
11
Example
𝑒1
𝑒2
𝑒3
𝑒4
𝑒5
𝑒6
processor 1
processor 2
𝑒1
𝑒2
𝑒3
0 2 4 6 8
𝑒1 2
𝑒2 2
𝑒3 1
𝑒4 1
𝑒5 2
𝑒6 3
![Page 12: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/12.jpg)
12
Example
𝑒1
𝑒2
𝑒3
𝑒4
𝑒5
𝑒6
processor 1
processor 2
𝑒1
𝑒2
𝑒3 𝑒4
0 2 4 6 8
𝑒1 2
𝑒2 2
𝑒3 1
𝑒4 1
𝑒5 2
𝑒6 3
![Page 13: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/13.jpg)
13
Example
𝑒1
𝑒2
𝑒3
𝑒4
𝑒5
𝑒6
processor 1
processor 2
𝑒1
𝑒2
𝑒3 𝑒4 𝑒5
𝑒6
0 2 4 6 8
𝑒1 2
𝑒2 2
𝑒3 1
𝑒4 1
𝑒5 2
𝑒6 3
![Page 14: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/14.jpg)
14
Scheduling Problem
Given
a set of N sporadic DAG-Tasks;
A scheduling algorithm (G-EDF or G-DM);
A platform with m identical processors;
verify if all deadlines are guaranteed.
![Page 15: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/15.jpg)
15
State of The Art
Existing schedulability analysis can be split in 3
categories:
Based on resource augmentation (speed-up);
(Baruah et al., Bonifaci et al., Nilissen et al.,…)
Based on capacity augmentation;
(Kim et al., Li et al., Lakshmanan et al., …)
Based on Response-Time Analysis.
(Maia et al., Chwa et al., Melani et al., …)
![Page 16: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/16.jpg)
16
This Work
Response-Time Analysis of Sporadic DAG-Tasks under both G-EDF and G-DM
Contribution w.r.t. the state of the art:
• Vertices-oriented analysis;
• Tasks can have arbitrary deadlines;
• Vertices can have arbitrary utilization;
• Augmentation bounds proved for N=1.
![Page 17: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/17.jpg)
17
Response-Time Analysis
For each DAG-Task 𝜏𝑖,
For each vertex 𝑣 of 𝜏𝑖,
Each job of vertex 𝑣 must complete within a deadline 𝑫𝒊
𝑒𝑣 + 𝐼𝑣 = 𝑹𝒗 ≤ 𝐷𝑖
Vertex WCET
Worst-case scheduling interference
![Page 18: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/18.jpg)
18
Response-Time Analysis
For each DAG-Task 𝜏𝑖,
For each vertex 𝑣 of 𝜏𝑖,
Each job of vertex 𝑣 must complete within a deadline 𝑫𝒊
𝑒𝑣 + 𝐼𝑣 = 𝑹𝒗 ≤ 𝐷𝑖
Not easy to compute
for multiprocessor systems!
![Page 19: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/19.jpg)
19
Response-Time Analysis
Our approach: compute an upper-bound 𝐼 𝑣 of the interference 𝐼𝑣 specific for each vertex 𝑣, so
obtaining a response-time upper-bound 𝑹𝒗
𝑒𝑣 + 𝐼𝑣 = 𝑹𝒗 ≤
𝑒𝑣 + 𝐼 𝑣 ⇒ 𝑅𝑣 ≤ 𝑹𝒗
![Page 20: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/20.jpg)
20
Response-Time Analysis
Main result of this work: we proved that
𝑅𝑣 ≤ 𝑹𝒗
𝑹𝒗 = 𝓁𝑣+ +1
𝑚
𝑣′
𝑾𝒗,𝒗′ 𝑹𝒗, 𝑌𝑣′ − 𝓁𝑣+
Critical path length:
maximum sum of WCETs in a path ending with 𝑣
![Page 21: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/21.jpg)
21
Critical Path
Critical path length: maximum sum of WCETs in a
path ending with 𝒗
1
3
4
6
2
8
4
𝒗1+4+2+4 = 11
3+6+4 = 13
𝓁𝑣+ = 13
![Page 22: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/22.jpg)
22
Response-Time Analysis
Main result: we proved that
𝑅𝑣 ≤ 𝑹𝒗
𝑹𝒗 = 𝓁𝑣+ +1
𝑚
𝑣′
𝑾𝒗,𝒗′ 𝑹𝒗, 𝑌𝑣′ − 𝓁𝑣+
Sum on all vertices 𝒗′ in the task-set
![Page 23: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/23.jpg)
23
Response-Time Analysis
Main result: we proved that
𝑅𝑣 ≤ 𝑹𝒗
𝑹𝒗 = 𝓁𝑣+ +1
𝑚
𝑣′
𝑾𝒗,𝒗′ 𝑹𝒗, 𝑌𝑣′ − 𝓁𝑣+
Upper-bound on the worst-case workloadgenerated by 𝒗′ interfering with 𝒗
![Page 24: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/24.jpg)
24
Worst-Case Workload
Upper-bound on the worst-case workload generated by 𝒗′ interfering with 𝒗
𝑾𝒗,𝒗′ 𝑹𝒗, 𝑌𝑣′
Tentative response-time
of vertex 𝑣, used in the
fixed-point iteration starting with 𝑹𝒗 = 𝒆𝒗
Response-time upper-bound
Must be always greater than
the response-time (𝑌𝑣′ = 𝐷𝑣 + 1 in the limit case)
![Page 25: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/25.jpg)
25
Worst-Case Workload
A generic vertex 𝒗′ interferes with 𝒗 released at 𝑡
𝑡 time
𝑡 − 𝑌𝑣′
𝑌𝑣′𝒗′
If shifted more on the left the job of 𝒗′
will be completed when 𝒗 is released
Release of a job of 𝒗
![Page 26: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/26.jpg)
26
Worst-Case Workload
A generic vertex 𝒗′ interferes with 𝒗 released at 𝑡
𝑡 time
𝑡 − 𝑌𝑣′𝑡 + 𝑹𝒗
𝑌𝑣′ + 𝑹𝒗
𝑌𝑣′ + 𝑹𝒗𝑇𝑣′
𝑒𝑣′Interferingworkload
In case of G-DM
Null for vertices of
lower-priority tasks
![Page 27: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/27.jpg)
27
Worst-Case Workload
A generic vertex 𝒗′ interferes with 𝒗 released at 𝑡
𝑡 time
𝑡 − 𝑌𝑣′
𝐷𝑣′𝒗′
𝑡 + 𝐷𝑣
Jobs of 𝒗′ released after𝑡 + 𝐷𝑣 − 𝐷𝑣′ will
not interfere with 𝒗
In case of G-EDF
![Page 28: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/28.jpg)
28
Worst-Case Workload
A generic vertex 𝒗′ interferes with 𝒗 released at 𝑡
𝑡 time
𝑡 − 𝑌𝑣′𝑡 + 𝐷𝑣
In case of G-EDF
𝑡 + 𝐷𝑣 − 𝐷𝑣′
𝑌𝑣′ +𝑚𝑖𝑛{𝑹𝒗, 𝐷𝑣 − 𝐷𝑣′}
𝑇𝑣′𝑒𝑣′
Interferingworkload
![Page 29: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/29.jpg)
29
Response-Time Analysis
Successors in the same job of a DAG-task cannot interfere
𝑒1
𝑒2
𝑒3
𝑒4
𝑒5
𝑒6
𝑒7
𝒗other tasks
![Page 30: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/30.jpg)
30
Response-Time Analysis
Main result: we proved that
𝑅𝑣 ≤ 𝑹𝒗
𝑹𝒗 = 𝓁𝑣+ +1
𝑚
𝑣′
𝑾𝒗,𝒗′ 𝑹𝒗, 𝑌𝑣′ − 𝓁𝑣+
![Page 31: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/31.jpg)
31
Schedulability Test
Algorithm RTA(N) Maximum number of iterations
![Page 32: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/32.jpg)
32
Schedulability Test
1. We start with 𝑌𝑣 = 𝐷𝑣 + 1, ∀𝑣, i = 1
2. Compute the fixed-point of
𝑹𝒗 = 𝓁𝑣+ +
1
𝑚 𝑣′𝑾𝒗,𝒗′ 𝑹𝒗, 𝑌𝑣′ − 𝓁𝑣
+
3. If 𝑹𝒗 ≤ 𝑫𝒗 return SCHEDULABLE
4. If 𝑌𝑣 == 𝑹𝒗, ∀𝑣 OR i==N return NOT SCHEDULABLE
5. Else, update response-times as 𝑌𝑣 = 𝑹𝒗, ∀𝑣 and
go to step 2
Pseudo-Polynomial Complexity𝑖 + +
Algorithm RTA(N)
![Page 33: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/33.jpg)
33
Polynomial-Time Schedulability Test
If we set 𝑌𝑣 = 𝐷𝑣 + 1 and 𝑅𝑣 = 𝐷𝑣 it is possible to obtain a simple polynomial-time schedulabilitytest without involving any iteration
𝑹𝒗 = 𝓁𝑣+ +1
𝑚
𝑣′
𝑾𝒗,𝒗′ 𝐷𝑣, 𝐷𝑣′ + 1 − 𝓁𝑣+
Polynomial Complexity
![Page 34: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/34.jpg)
34
Augmentation Bound
In case of a task-set composed of a single DAG-
Task (N=1) we proved that
Our test based on response-time analysis has
Augmentation bound < 3 for G-EDF;
Augmentation bound < 5 for G-DM.
![Page 35: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/35.jpg)
35
Experimental Results
The proposed schedulability tests have been
evaluated by using synthetic workload
libdag – DAG-Tasks generator and
schedulability test. Soon publicly available online!
Comparison against the test based on
augmentation bound proposed in
V. Bonifaci, A. Marchetti-Spaccamela, S. Stiller, and A. Wiese.
“Feasibility analysis in the sporadic DAG task model”, In proc. of ECRTS 2013
To the best of our knowledge it is the onlytest dealing with arbitrary deadlines
![Page 36: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/36.jpg)
36
Experimental Results
Number N of external iterations in our algorithm
The test of Bonifaci et al.
is based on a
workload approximation
up to an 𝜖-error
with 2−𝛿
𝛿
![Page 37: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/37.jpg)
37
Experimental Results
Running times of the schedulability tests
Exponential increase
of the running time as the test precision increases
RTA test has
running time lower oftwo orders of magnitude
(Intel Xeon @ 3.5 Ghz)
![Page 38: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/38.jpg)
38
Experimental Results
Take-away messages
• RTA test outperforms
the speed-up based
test in all the tested
configurations;
• In some cases our
polynomial-time test
performs better than
the speed-up based
test that has pseudo-
polynomial complexity
𝑈 = 10
![Page 39: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/39.jpg)
39
Conclusions
We proposed a new Response-Time Analysis for
the sporadic DAG-Task model under both G-EDF
and G-DM scheduling;
The analysis handles DAG-Tasks with arbitrary
deadline and arbitrary utilization;
Two schedulability tests have been derived
(pseudo-polynomial and polynomial complexity);
Extensive set of experimental results confirmed
the effectiveness of the test.
![Page 40: Response Time Analysis for G-EDF and G-DM Scheduling of ...€¦ · Response Time Analysis for G-EDF and G-DM Scheduling of Sporadic DAG-Tasks with Arbitrary Deadline Andrea Parri,](https://reader035.vdocument.in/reader035/viewer/2022062505/5edd9d0fad6a402d6668c0b6/html5/thumbnails/40.jpg)
40
Future Work
More accurate characterization of the interfering
workload;
Support for conditional statements in the DAG-
Task;
Integration of locking protocols in the analysis;
Handle distributed computations.