WORLDCOMP Conference: PDPTA-2014
Las Vegas, Nevada, USA
“Impact of Thread Synchronization and Data Parallelism
on Multicore Game Programming”
July 21-24, 2014
WORLDCOMP Conference: PDPTA-2014
Las Vegas, Nevada, USA
“Impact of Thread Synchronization and Data Parallelism
on Multicore Game Programming”
Presented by:
Dr. Abu Asaduzzaman
Assistant Professor in Computer Architecture and Director of CAPPLab
Department of Electrical Engineering and Computer Science (EECS)
Wichita State University (WSU), USA
July 23, 2014
“Impact of Thread Synchronization and Data
Parallelism on Multicore Game Programming”
►
Outline
■ Introduction
 Problem Statement and Contributions
 Background and Motivation
■ Important Techniques for Game Engine
 Data Level Parallelism
 Task Level Parallelism
■ Experimentations
 Test Game Engine: Tower Defense Game
 Different Implementations
■ Experimental Results
■ Discussion
Dr. Zaman
QUESTIONS? Any time!
3
“Impact of Thread Synchronization and Data
Parallelism on Multicore Game Programming”
Authors
■ Abu Asaduzzaman, Assistant Professor
 EECS Department, Wichita State University (WSU), USA
 Director, Computer Arch & Parallel Prog Lab (CAPPLab)
■ Hin Y. Lee, MS in CS
 EECS Department, Wichita State University (WSU), USA
 Graduation: Fall 2012
■ Deepthi Gummadi, MS/CN Student
 EECS Department, Wichita State University (WSU), USA
 Graduation: Spring 2014
Dr. Zaman
4
“Impact of Thread Synchronization and Data
Parallelism on Multicore Game Programming”
Introduction
■ Xbox 360
 3 cores
 GPU card
■ PS3
 1 master core
 6 worker core
 GPU card
Dr. Zaman
5
“Impact of Thread Synchronization and Data
Parallelism on Multicore Game Programming”
Introduction
■ Game Technology
 In addition to standalone game machines, game engines
are nowadays being used for educational, engineering,
and scientific applications.
 To fulfill the high performance requirements, game
engines are adopting new hardware technologies like
multicore CPUs and software technologies like parallel
programming.
 As the number of cores in a processor increases,
multithreading can be very helpful to get as much
performance out of a system as possible to the
advancement of video game technologies.
Dr. Zaman
6
“Impact of Thread Synchronization and Data
Parallelism on Multicore Game Programming”
Introduction
■ Game Engine
 Many Components: Physics, Graphics, etc.
■ Multithreading




Concurrent and Parallel Processing
Data Level Parallelism and Task Level Parallelism
(Thread) Synchronization
Load Balancing, etc.
Dr. Zaman
7
“Impact of Thread Synchronization and Data
Parallelism on Multicore Game Programming”
Introduction
■ Contributions
 Multi-object interactive game
console in an 8-core system
 Single-Threaded Model (STM)
 Multithreaded Asynchronous Model (MAM)
 Multithreaded Synchronous Model (MSM)
 MSM with Data Parallelism (MSMDP)
 We investigate the challenges and benefits of thread
synchronization and data level parallelism.
Dr. Zaman
8
“Impact of Thread Synchronization and Data
Parallelism on Multicore Game Programming”
Introduction
■ Background and Motivation
 The game industry has surpassed the movie and music
industry in U.S. in 2005 and 2007, respectively. In 2008,
the game industry surpassed the music industry in the
U.K. and is expected to surpass DVD sales in the future.
 Multicore architecture is a recent design trend and most
vendors are adopting multicore processors to their
products. Multicore systems are very suitable for
multithreaded processing as multiple threads can be
executed on multiple cores at the same time.
Dr. Zaman
9
“Impact of Thread Synchronization and Data
Parallelism on Multicore Game Programming”
Introduction
■ Background and Motivation
 Data parallelism is where the same type of data is
parallelized in multiple threads. The use of this in a game
engine is when a component spawns multiple worker
threads to process one type of data.
 Task level parallelism is a popular method for game
engine multithreading, where components run
asynchronously in their own loop or synchronously in a
single loop with multiple forks and joins.
 To properly scale a multicore game engine, data
parallelism and task parallelism have to be employed.
Dr. Zaman
10
“Impact of Thread Synchronization and Data
Parallelism on Multicore Game Programming”
►
Outline
■ Introduction
 Problem Statement and Contributions
 Background Work and Motivation
■ Important Techniques for Game Engine
 Data Level Parallelism
 Task Level Parallelism
■ Experimentations
 Test Game Engine: Tower Defense Game
 Different Implementations
■ Experimental Results
■ Discussion
Dr. Zaman
11
“Impact of Thread Synchronization and Data
Parallelism on Multicore Game Programming”
Important Techniques for Game Engine
■ Data Level Parallelism
 Data parallelism is where the same type of data in a
component is parallelized in multiple threads.
 Example: The animation subcomponent in the graphics
component is divided into 3 batches of data for
simultaneous processing.
Dr. Zaman
12
“Impact of Thread Synchronization and Data
Parallelism on Multicore Game Programming”
Important Techniques for Game Engine
■ Task Level Parallelism
 The use of task parallelism in game engine is by running
each component task in its own thread.
 Example: In synchronized model (left) all tasks must
finish in a single clock cycle. In asynchronous model
(right) the tasks can run and finish at their own time.
Dr. Zaman
13
“Impact of Thread Synchronization and Data
Parallelism on Multicore Game Programming”
Important Techniques for Game Engine
■ Thread Pool
 Thread pool is a system where tasks are queued in a list
and assigned to any available threads that are idle.
 Max. amount of threads should run at the same time.
Dr. Zaman
14
“Impact of Thread Synchronization and Data
Parallelism on Multicore Game Programming”
►
Outline
■ Introduction
 Problem Statement and Contributions
 Background Work and Motivation
■ Important Techniques for Game Engine
 Data Level Parallelism
 Task Level Parallelism
■ Experimentations
 Test Game Engine: Tower Defense Game
 Different Implementations
■ Experimental Results
■ Discussion
Dr. Zaman
15
“Impact of Thread Synchronization and Data
Parallelism on Multicore Game Programming”
Test Game Engine
■ CAPPLab Tower Defense Game v1
 Defend the Base Structure; players are to build Defensive
Structures and try to survive as many waves as possible. For every
enemy destroyed, the player gains credits. Maximum 20 objects in
the screen at one time.




The game is over if
(i) Player survives 3 minutes,
(ii) Base Struct. is destroyed, or
(iii) Abnormal termination
Dr. Zaman
16
“Impact of Thread Synchronization and Data
Parallelism on Multicore Game Programming”
Test Game Engine
■ Single Threaded Model (STM)
 The order of operations in the single threaded
implementation is:
1) Capture input
2) Update input operation (graphics)
3) Update game logic
4) Update AI
5) Update physics
6) Process navigational mesh updates
7) Simulate physics
8) Render graphics
Dr. Zaman
17
“Impact of Thread Synchronization and Data
Parallelism on Multicore Game Programming”
Test Game Engine
■ Multithreaded Asynchronous Model (MAM)
 The order of operations is:
Thread 1:
Thread 2:
1) Capture input
1) Update graphics
2) Update game logic 2) Render graphics
3) Update AI
4) Update physics
5) Process navigational mesh updates
6) Simulate physics
Dr. Zaman
18
“Impact of Thread Synchronization and Data
Parallelism on Multicore Game Programming”
Test Game Engine
■ Multithreaded Synchronous Model (MSM)
 The order of operations is:
Serial Stage:
Parallel stage: (One possible order)
1) Capture input 1) Process navigational mesh updates
2) Update logic 2) Simulate physics
3) Update AI
3) Render graphics
4) Update physics
5) Update graphics
Dr. Zaman
19
“Impact of Thread Synchronization and Data
Parallelism on Multicore Game Programming”
Test Game Engine
■ MSM with Data Parallelism (MSMDP)
 The order of operations is:
Serial Stage:
Parallel stage: (One possible order)
1) Capture input 1) Update navigational mesh
2) Update logic 2) Simulate physics
3) Update AI
3) Perform collision detection on object batch 1
4) Update physics
4) Perform collision detection on object batch 2
5) Update graphics
5) Render graphics
Dr. Zaman
20
“Impact of Thread Synchronization and Data
Parallelism on Multicore Game Programming”
►
Outline
■ Introduction
 Problem Statement and Contributions
 Background Work and Motivation
■ Important Techniques for Game Engine
 Data Level Parallelism
 Task Level Parallelism
■ Experimentations
 Test Game Engine: Tower Defense Game
 Different Implementations
■ Experimental Results
■ Discussion
Dr. Zaman
21
“Impact of Thread Synchronization and Data
Parallelism on Multicore Game Programming”
Experimental Results
■ Generation and processing of Frames
 MSMDP generates more frames than other implementations.
 SMT takes more time to process a frame than multithreaded ones.
Dr. Zaman
22
“Impact of Thread Synchronization and Data
Parallelism on Multicore Game Programming”
Experimental Results
■ Processing Time for Components and Speedup
 Physics simulation takes the maximum amount of time.
 About 14x speedup is achieved due to MSMDP for 512 threads.
Dr. Zaman
23
“Impact of Thread Synchronization and Data
Parallelism on Multicore Game Programming”
Conclusions and Future Extension
■ We explore the challenges due to thread
synchronization and data parallelism by developing a
multicore video game console.
■ Experimental results show that the MSMDP generates
more frames and takes less amount of time to process
the frames.
■ On an 8-core system, the speedup due to MSMDP is
about 14 with respect to the STM implementation.
■ Higher speedup needed for future game engines can
be achieved by GPU computing.
Dr. Zaman
24
WORLDCOMP Conference: PDPTA; Las Vegas, USA; 2014
“Impact of Thread Synchronization and Data Parallelism on
Multicore Game Programming”
QUESTIONS?
Contact: Abu Asaduzzaman
E-mail: abuasaduzzaman@ieee.org
Phone: +1-316-978-5261
CAPPLab: http://www.cs.wichita.edu/~capplab/
Thank You!