Download presentation
Presentation is loading. Please wait.
Published byTiffany Tucker Modified over 9 years ago
1
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 20131 FIT5174 Distributed & Parallel Systems Lecture 5 Message Passing and MPI
2
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 20132 Acknowledgement These slides are based on slides and material by: C Evangelinos, Scott Baden, Zhiliang Xu
3
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 20133 Shared Memory Approach
4
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 20134 Message Passing Model
5
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 20135 Distributed Computing using Message Passing
6
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 20136 Single Program / Multiple Program - Multiple Data
7
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 20137 Programming with message passing
8
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 20138 Message passing concept
9
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 20139 Send and Receive concept
10
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201310 Message completion
11
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201311 Buffering concept
12
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201312 Causality concept
13
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201313 Asynchronous versus Synchronous
14
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201314 Overlapping operations
15
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201315 What is MPI?
16
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201316 MPI websites
17
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201317 MPI: Message Passing Interface
18
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201318 MPI history
19
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201319 MPI in context
20
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201320 MPI fundamentals
21
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201321 Minimal MPI subset
22
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201322 Initializing MPI
23
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201323 Getting started with MPI
24
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201324 MPI program details
25
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201325 MPI program more details
26
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201326 Simple MPI C Program
27
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201327 MPI Communicators and handles
28
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201328 MPI Communicator size
29
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201329 MPI Communicator size
30
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201330 MPI Process rank
31
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201331 MPI Process rank
32
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201332 MPI exiting
33
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201333 MPI communications
34
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201334 MPI basic datatypes for C
35
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201335 MPI C data types
36
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201336 MPI messages
37
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201337 MPI point-to-point communications
38
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201338 MPI point-to-point messages
39
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201339 MPI point-to-point messages requirements
40
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201340 Synchronous versus Asynchronous communications
41
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201341 MPI blocking standard Send
42
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201342 Send and Receive code fragment
43
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201343 Other MPI Send options
44
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201344 Considerations for MPI Send
45
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201345 Performance of MPI Send
46
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201346 MPI blocking Receive
47
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201347 MPI Receive considerations
48
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201348 MPI message passing restrictions
49
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201349 Simple MPI ‘ping-pong’
50
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201350 MPI deadlock scenario
51
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201351 MPI correctness and fairness
52
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201352 MPI bidirectional communication
53
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201353 MPI non-blocking communication
54
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201354 MPI non-blocking standard send
55
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201355 MPI non-blocking receive
56
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201356 Avoiding Deadlocks
57
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201357 MPI non-blocking sends
58
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201358 MPI non-blocking Synchronous Send
59
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201359 MPI testing instead of waiting
60
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201360 MPI with many outstanding calls
61
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201361 Other MPI communications calls
62
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201362 MPI persistent communication
63
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201363 MPI persistent communications
64
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201364 Useful MPI wildcards and constants
65
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201365 MPI collective communications
66
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201366 MPI synchronization
67
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201367 MPI broadcast
68
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201368 MPI gather
69
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201369 MPI scatter
70
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201370 MPI gather to all
71
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201371 Using vectors
72
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201372 Using vectors non-contiguously
73
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201373 Using binary trees
74
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201374 Using other kinds of trees
75
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201375 All-to-All personalized communication
76
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201376 MPI program structure summary
77
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201377 MPI program summary
78
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201378 MPI Simple C program
79
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201379 MPI Safe C program
80
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201380 MPI Deadlocking C program
81
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201381 MPI buffering dependent C program
82
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201382 Summarizing Distributed versus Shared Memory
83
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201383 Hybrid programming
84
FIT5174 Parallel & Distributed Systems Dr. Ronald Pose Lecture 5 - 201384 Why use hybrid programming?
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.