Multicore Chips

Report
Slide 1

EPGP-EG 2013-14, IIM Indore.
Course Name : EMERGING TECHNOLOGOES

“Multicore Chips”
Group Number: 3
Group Members:
1) Mr. Arbind Kumar
2) Ms. Archana Shinde
3) Mr. Pravin Kolhe
4) Mr. Krushna Chnadra Satapathy

(EG2013-04)
(EG2013-06)
(EG2013-17)
(EG2013-18)

INTRODUCTION
• Processor:- Microchip implanted in a CPU's hard drive that processes
instructions sent to it by the computer and software programs.
• Multicore:- Multi-core processor is IC to which ≥ 2 processors are
attached for:– enhanced performance,
– reduced power consumption,
– more efficient simultaneous processing of multiple tasks

• Challenge: Much of processing power is unharnessed.
• Goal: Attract developer to join multicore movement.
• How to Use multicore:– Multithreading :– Parallel
:– Concurrent :-

Single code is executed by several processors.
Property of machine to do side by side.
Property of program to do at same time.

HOW MULTITASK WORKS?

INITIATIVES FOR MULTICORE APPLICATIONS
• Apple’s Initiative: Mac OS X Snow Leopard (Code name-Grand
Central) gives application leverage up to 8 cores of processing power.
• RapidMind: Offers software development platform that allows
developers to exploit quad-core AMD Opteron & Intel Xeon Processors.
• Nvidia Graphics Processor: Provided Parallel Programming Language
& libraries, called ‘Cuba’.
• Java Environment for Parallel Real-Time Development (Jeopard):
Java based Real-time applications on multicore chips is planned.
• Khronos Group’s Compute Working Group:- Royalty Free Standard
for building graphical multicore applications. (OpenCL)
• Microsoft: Parallel computing initiative.
• Intel’s Parallel Studio for C & C++.
• Other Vendors:- Codeplay, Polycore Software, Click Arts etc.

CHALLENGES FOR INDUSTRY
• Image, Graphics, videos & business applications will be first to benefit,
as they can be broken down in to parts for simultaneous processing.
• But, for most complex tasks, it is difficult to find single approach for
identifying sequence of parallel computations.
• Most software are written for single core chips.
• The performance of Off-the-shelf applications reduces after 4 cores.
• Lack of multicore programming tools for mainstream developers.
• Race Conditions:- Output of one calculation depends on various
elements of a task being completed in certain order.
• Programs having interdependencies of tasks.
• The tools available currently needs manual interventions for identifying
code that can be parallelized.
• Hardware issues such as Memory Latency & Slow Bus Speed due to
limited Memory Bandwidth.

REFERENCES
1)

2)

3)

4)

“Taking advantage of multicore PCs” by Tom Kaneshige, Source URL (retrieved on
2013-05-25 11:32): http://www.infoworld.com/d/developer-world/taking-advantagemulticore-pcs-711.
“Multicore chips pose next big challenge for industry” by Agam Shah, Source URL
(retrieved on 2013-05-26 12:55): http://www.infoworld.com/d/developerworld/multicore-chips-pose-next-big-challenge-industry-105
“Modern multicore and the next generation of IT” by Paul Venezia, Source URL
(retrieved on 2013-05-26 01:14): http://www.infoworld.com/d/computerhardware/modern-multicore-and-the-next-generation-it-325
“The Trouble With Multicore” by David Patterson, Source URL (retrieved on 2013-0526 15:37):
http://www.ann.ece.ufl.edu/courses/eel4713_12spr/articles/The_Trouble_With_Multicor
e-IEEE_Spectrum.pdf

8

PPT downloaded from www.pravinkolhe.com


Slide 2

EPGP-EG 2013-14, IIM Indore.
Course Name : EMERGING TECHNOLOGOES

“Multicore Chips”
Group Number: 3
Group Members:
1) Mr. Arbind Kumar
2) Ms. Archana Shinde
3) Mr. Pravin Kolhe
4) Mr. Krushna Chnadra Satapathy

(EG2013-04)
(EG2013-06)
(EG2013-17)
(EG2013-18)

INTRODUCTION
• Processor:- Microchip implanted in a CPU's hard drive that processes
instructions sent to it by the computer and software programs.
• Multicore:- Multi-core processor is IC to which ≥ 2 processors are
attached for:– enhanced performance,
– reduced power consumption,
– more efficient simultaneous processing of multiple tasks

• Challenge: Much of processing power is unharnessed.
• Goal: Attract developer to join multicore movement.
• How to Use multicore:– Multithreading :– Parallel
:– Concurrent :-

Single code is executed by several processors.
Property of machine to do side by side.
Property of program to do at same time.

HOW MULTITASK WORKS?

INITIATIVES FOR MULTICORE APPLICATIONS
• Apple’s Initiative: Mac OS X Snow Leopard (Code name-Grand
Central) gives application leverage up to 8 cores of processing power.
• RapidMind: Offers software development platform that allows
developers to exploit quad-core AMD Opteron & Intel Xeon Processors.
• Nvidia Graphics Processor: Provided Parallel Programming Language
& libraries, called ‘Cuba’.
• Java Environment for Parallel Real-Time Development (Jeopard):
Java based Real-time applications on multicore chips is planned.
• Khronos Group’s Compute Working Group:- Royalty Free Standard
for building graphical multicore applications. (OpenCL)
• Microsoft: Parallel computing initiative.
• Intel’s Parallel Studio for C & C++.
• Other Vendors:- Codeplay, Polycore Software, Click Arts etc.

CHALLENGES FOR INDUSTRY
• Image, Graphics, videos & business applications will be first to benefit,
as they can be broken down in to parts for simultaneous processing.
• But, for most complex tasks, it is difficult to find single approach for
identifying sequence of parallel computations.
• Most software are written for single core chips.
• The performance of Off-the-shelf applications reduces after 4 cores.
• Lack of multicore programming tools for mainstream developers.
• Race Conditions:- Output of one calculation depends on various
elements of a task being completed in certain order.
• Programs having interdependencies of tasks.
• The tools available currently needs manual interventions for identifying
code that can be parallelized.
• Hardware issues such as Memory Latency & Slow Bus Speed due to
limited Memory Bandwidth.

REFERENCES
1)

2)

3)

4)

“Taking advantage of multicore PCs” by Tom Kaneshige, Source URL (retrieved on
2013-05-25 11:32): http://www.infoworld.com/d/developer-world/taking-advantagemulticore-pcs-711.
“Multicore chips pose next big challenge for industry” by Agam Shah, Source URL
(retrieved on 2013-05-26 12:55): http://www.infoworld.com/d/developerworld/multicore-chips-pose-next-big-challenge-industry-105
“Modern multicore and the next generation of IT” by Paul Venezia, Source URL
(retrieved on 2013-05-26 01:14): http://www.infoworld.com/d/computerhardware/modern-multicore-and-the-next-generation-it-325
“The Trouble With Multicore” by David Patterson, Source URL (retrieved on 2013-0526 15:37):
http://www.ann.ece.ufl.edu/courses/eel4713_12spr/articles/The_Trouble_With_Multicor
e-IEEE_Spectrum.pdf

8

PPT downloaded from www.pravinkolhe.com


Slide 3

EPGP-EG 2013-14, IIM Indore.
Course Name : EMERGING TECHNOLOGOES

“Multicore Chips”
Group Number: 3
Group Members:
1) Mr. Arbind Kumar
2) Ms. Archana Shinde
3) Mr. Pravin Kolhe
4) Mr. Krushna Chnadra Satapathy

(EG2013-04)
(EG2013-06)
(EG2013-17)
(EG2013-18)

INTRODUCTION
• Processor:- Microchip implanted in a CPU's hard drive that processes
instructions sent to it by the computer and software programs.
• Multicore:- Multi-core processor is IC to which ≥ 2 processors are
attached for:– enhanced performance,
– reduced power consumption,
– more efficient simultaneous processing of multiple tasks

• Challenge: Much of processing power is unharnessed.
• Goal: Attract developer to join multicore movement.
• How to Use multicore:– Multithreading :– Parallel
:– Concurrent :-

Single code is executed by several processors.
Property of machine to do side by side.
Property of program to do at same time.

HOW MULTITASK WORKS?

INITIATIVES FOR MULTICORE APPLICATIONS
• Apple’s Initiative: Mac OS X Snow Leopard (Code name-Grand
Central) gives application leverage up to 8 cores of processing power.
• RapidMind: Offers software development platform that allows
developers to exploit quad-core AMD Opteron & Intel Xeon Processors.
• Nvidia Graphics Processor: Provided Parallel Programming Language
& libraries, called ‘Cuba’.
• Java Environment for Parallel Real-Time Development (Jeopard):
Java based Real-time applications on multicore chips is planned.
• Khronos Group’s Compute Working Group:- Royalty Free Standard
for building graphical multicore applications. (OpenCL)
• Microsoft: Parallel computing initiative.
• Intel’s Parallel Studio for C & C++.
• Other Vendors:- Codeplay, Polycore Software, Click Arts etc.

CHALLENGES FOR INDUSTRY
• Image, Graphics, videos & business applications will be first to benefit,
as they can be broken down in to parts for simultaneous processing.
• But, for most complex tasks, it is difficult to find single approach for
identifying sequence of parallel computations.
• Most software are written for single core chips.
• The performance of Off-the-shelf applications reduces after 4 cores.
• Lack of multicore programming tools for mainstream developers.
• Race Conditions:- Output of one calculation depends on various
elements of a task being completed in certain order.
• Programs having interdependencies of tasks.
• The tools available currently needs manual interventions for identifying
code that can be parallelized.
• Hardware issues such as Memory Latency & Slow Bus Speed due to
limited Memory Bandwidth.

REFERENCES
1)

2)

3)

4)

“Taking advantage of multicore PCs” by Tom Kaneshige, Source URL (retrieved on
2013-05-25 11:32): http://www.infoworld.com/d/developer-world/taking-advantagemulticore-pcs-711.
“Multicore chips pose next big challenge for industry” by Agam Shah, Source URL
(retrieved on 2013-05-26 12:55): http://www.infoworld.com/d/developerworld/multicore-chips-pose-next-big-challenge-industry-105
“Modern multicore and the next generation of IT” by Paul Venezia, Source URL
(retrieved on 2013-05-26 01:14): http://www.infoworld.com/d/computerhardware/modern-multicore-and-the-next-generation-it-325
“The Trouble With Multicore” by David Patterson, Source URL (retrieved on 2013-0526 15:37):
http://www.ann.ece.ufl.edu/courses/eel4713_12spr/articles/The_Trouble_With_Multicor
e-IEEE_Spectrum.pdf

8

PPT downloaded from www.pravinkolhe.com


Slide 4

EPGP-EG 2013-14, IIM Indore.
Course Name : EMERGING TECHNOLOGOES

“Multicore Chips”
Group Number: 3
Group Members:
1) Mr. Arbind Kumar
2) Ms. Archana Shinde
3) Mr. Pravin Kolhe
4) Mr. Krushna Chnadra Satapathy

(EG2013-04)
(EG2013-06)
(EG2013-17)
(EG2013-18)

INTRODUCTION
• Processor:- Microchip implanted in a CPU's hard drive that processes
instructions sent to it by the computer and software programs.
• Multicore:- Multi-core processor is IC to which ≥ 2 processors are
attached for:– enhanced performance,
– reduced power consumption,
– more efficient simultaneous processing of multiple tasks

• Challenge: Much of processing power is unharnessed.
• Goal: Attract developer to join multicore movement.
• How to Use multicore:– Multithreading :– Parallel
:– Concurrent :-

Single code is executed by several processors.
Property of machine to do side by side.
Property of program to do at same time.

HOW MULTITASK WORKS?

INITIATIVES FOR MULTICORE APPLICATIONS
• Apple’s Initiative: Mac OS X Snow Leopard (Code name-Grand
Central) gives application leverage up to 8 cores of processing power.
• RapidMind: Offers software development platform that allows
developers to exploit quad-core AMD Opteron & Intel Xeon Processors.
• Nvidia Graphics Processor: Provided Parallel Programming Language
& libraries, called ‘Cuba’.
• Java Environment for Parallel Real-Time Development (Jeopard):
Java based Real-time applications on multicore chips is planned.
• Khronos Group’s Compute Working Group:- Royalty Free Standard
for building graphical multicore applications. (OpenCL)
• Microsoft: Parallel computing initiative.
• Intel’s Parallel Studio for C & C++.
• Other Vendors:- Codeplay, Polycore Software, Click Arts etc.

CHALLENGES FOR INDUSTRY
• Image, Graphics, videos & business applications will be first to benefit,
as they can be broken down in to parts for simultaneous processing.
• But, for most complex tasks, it is difficult to find single approach for
identifying sequence of parallel computations.
• Most software are written for single core chips.
• The performance of Off-the-shelf applications reduces after 4 cores.
• Lack of multicore programming tools for mainstream developers.
• Race Conditions:- Output of one calculation depends on various
elements of a task being completed in certain order.
• Programs having interdependencies of tasks.
• The tools available currently needs manual interventions for identifying
code that can be parallelized.
• Hardware issues such as Memory Latency & Slow Bus Speed due to
limited Memory Bandwidth.

REFERENCES
1)

2)

3)

4)

“Taking advantage of multicore PCs” by Tom Kaneshige, Source URL (retrieved on
2013-05-25 11:32): http://www.infoworld.com/d/developer-world/taking-advantagemulticore-pcs-711.
“Multicore chips pose next big challenge for industry” by Agam Shah, Source URL
(retrieved on 2013-05-26 12:55): http://www.infoworld.com/d/developerworld/multicore-chips-pose-next-big-challenge-industry-105
“Modern multicore and the next generation of IT” by Paul Venezia, Source URL
(retrieved on 2013-05-26 01:14): http://www.infoworld.com/d/computerhardware/modern-multicore-and-the-next-generation-it-325
“The Trouble With Multicore” by David Patterson, Source URL (retrieved on 2013-0526 15:37):
http://www.ann.ece.ufl.edu/courses/eel4713_12spr/articles/The_Trouble_With_Multicor
e-IEEE_Spectrum.pdf

8

PPT downloaded from www.pravinkolhe.com


Slide 5

EPGP-EG 2013-14, IIM Indore.
Course Name : EMERGING TECHNOLOGOES

“Multicore Chips”
Group Number: 3
Group Members:
1) Mr. Arbind Kumar
2) Ms. Archana Shinde
3) Mr. Pravin Kolhe
4) Mr. Krushna Chnadra Satapathy

(EG2013-04)
(EG2013-06)
(EG2013-17)
(EG2013-18)

INTRODUCTION
• Processor:- Microchip implanted in a CPU's hard drive that processes
instructions sent to it by the computer and software programs.
• Multicore:- Multi-core processor is IC to which ≥ 2 processors are
attached for:– enhanced performance,
– reduced power consumption,
– more efficient simultaneous processing of multiple tasks

• Challenge: Much of processing power is unharnessed.
• Goal: Attract developer to join multicore movement.
• How to Use multicore:– Multithreading :– Parallel
:– Concurrent :-

Single code is executed by several processors.
Property of machine to do side by side.
Property of program to do at same time.

HOW MULTITASK WORKS?

INITIATIVES FOR MULTICORE APPLICATIONS
• Apple’s Initiative: Mac OS X Snow Leopard (Code name-Grand
Central) gives application leverage up to 8 cores of processing power.
• RapidMind: Offers software development platform that allows
developers to exploit quad-core AMD Opteron & Intel Xeon Processors.
• Nvidia Graphics Processor: Provided Parallel Programming Language
& libraries, called ‘Cuba’.
• Java Environment for Parallel Real-Time Development (Jeopard):
Java based Real-time applications on multicore chips is planned.
• Khronos Group’s Compute Working Group:- Royalty Free Standard
for building graphical multicore applications. (OpenCL)
• Microsoft: Parallel computing initiative.
• Intel’s Parallel Studio for C & C++.
• Other Vendors:- Codeplay, Polycore Software, Click Arts etc.

CHALLENGES FOR INDUSTRY
• Image, Graphics, videos & business applications will be first to benefit,
as they can be broken down in to parts for simultaneous processing.
• But, for most complex tasks, it is difficult to find single approach for
identifying sequence of parallel computations.
• Most software are written for single core chips.
• The performance of Off-the-shelf applications reduces after 4 cores.
• Lack of multicore programming tools for mainstream developers.
• Race Conditions:- Output of one calculation depends on various
elements of a task being completed in certain order.
• Programs having interdependencies of tasks.
• The tools available currently needs manual interventions for identifying
code that can be parallelized.
• Hardware issues such as Memory Latency & Slow Bus Speed due to
limited Memory Bandwidth.

REFERENCES
1)

2)

3)

4)

“Taking advantage of multicore PCs” by Tom Kaneshige, Source URL (retrieved on
2013-05-25 11:32): http://www.infoworld.com/d/developer-world/taking-advantagemulticore-pcs-711.
“Multicore chips pose next big challenge for industry” by Agam Shah, Source URL
(retrieved on 2013-05-26 12:55): http://www.infoworld.com/d/developerworld/multicore-chips-pose-next-big-challenge-industry-105
“Modern multicore and the next generation of IT” by Paul Venezia, Source URL
(retrieved on 2013-05-26 01:14): http://www.infoworld.com/d/computerhardware/modern-multicore-and-the-next-generation-it-325
“The Trouble With Multicore” by David Patterson, Source URL (retrieved on 2013-0526 15:37):
http://www.ann.ece.ufl.edu/courses/eel4713_12spr/articles/The_Trouble_With_Multicor
e-IEEE_Spectrum.pdf

8

PPT downloaded from www.pravinkolhe.com


Slide 6

EPGP-EG 2013-14, IIM Indore.
Course Name : EMERGING TECHNOLOGOES

“Multicore Chips”
Group Number: 3
Group Members:
1) Mr. Arbind Kumar
2) Ms. Archana Shinde
3) Mr. Pravin Kolhe
4) Mr. Krushna Chnadra Satapathy

(EG2013-04)
(EG2013-06)
(EG2013-17)
(EG2013-18)

INTRODUCTION
• Processor:- Microchip implanted in a CPU's hard drive that processes
instructions sent to it by the computer and software programs.
• Multicore:- Multi-core processor is IC to which ≥ 2 processors are
attached for:– enhanced performance,
– reduced power consumption,
– more efficient simultaneous processing of multiple tasks

• Challenge: Much of processing power is unharnessed.
• Goal: Attract developer to join multicore movement.
• How to Use multicore:– Multithreading :– Parallel
:– Concurrent :-

Single code is executed by several processors.
Property of machine to do side by side.
Property of program to do at same time.

HOW MULTITASK WORKS?

INITIATIVES FOR MULTICORE APPLICATIONS
• Apple’s Initiative: Mac OS X Snow Leopard (Code name-Grand
Central) gives application leverage up to 8 cores of processing power.
• RapidMind: Offers software development platform that allows
developers to exploit quad-core AMD Opteron & Intel Xeon Processors.
• Nvidia Graphics Processor: Provided Parallel Programming Language
& libraries, called ‘Cuba’.
• Java Environment for Parallel Real-Time Development (Jeopard):
Java based Real-time applications on multicore chips is planned.
• Khronos Group’s Compute Working Group:- Royalty Free Standard
for building graphical multicore applications. (OpenCL)
• Microsoft: Parallel computing initiative.
• Intel’s Parallel Studio for C & C++.
• Other Vendors:- Codeplay, Polycore Software, Click Arts etc.

CHALLENGES FOR INDUSTRY
• Image, Graphics, videos & business applications will be first to benefit,
as they can be broken down in to parts for simultaneous processing.
• But, for most complex tasks, it is difficult to find single approach for
identifying sequence of parallel computations.
• Most software are written for single core chips.
• The performance of Off-the-shelf applications reduces after 4 cores.
• Lack of multicore programming tools for mainstream developers.
• Race Conditions:- Output of one calculation depends on various
elements of a task being completed in certain order.
• Programs having interdependencies of tasks.
• The tools available currently needs manual interventions for identifying
code that can be parallelized.
• Hardware issues such as Memory Latency & Slow Bus Speed due to
limited Memory Bandwidth.

REFERENCES
1)

2)

3)

4)

“Taking advantage of multicore PCs” by Tom Kaneshige, Source URL (retrieved on
2013-05-25 11:32): http://www.infoworld.com/d/developer-world/taking-advantagemulticore-pcs-711.
“Multicore chips pose next big challenge for industry” by Agam Shah, Source URL
(retrieved on 2013-05-26 12:55): http://www.infoworld.com/d/developerworld/multicore-chips-pose-next-big-challenge-industry-105
“Modern multicore and the next generation of IT” by Paul Venezia, Source URL
(retrieved on 2013-05-26 01:14): http://www.infoworld.com/d/computerhardware/modern-multicore-and-the-next-generation-it-325
“The Trouble With Multicore” by David Patterson, Source URL (retrieved on 2013-0526 15:37):
http://www.ann.ece.ufl.edu/courses/eel4713_12spr/articles/The_Trouble_With_Multicor
e-IEEE_Spectrum.pdf

8

PPT downloaded from www.pravinkolhe.com


Slide 7

EPGP-EG 2013-14, IIM Indore.
Course Name : EMERGING TECHNOLOGOES

“Multicore Chips”
Group Number: 3
Group Members:
1) Mr. Arbind Kumar
2) Ms. Archana Shinde
3) Mr. Pravin Kolhe
4) Mr. Krushna Chnadra Satapathy

(EG2013-04)
(EG2013-06)
(EG2013-17)
(EG2013-18)

INTRODUCTION
• Processor:- Microchip implanted in a CPU's hard drive that processes
instructions sent to it by the computer and software programs.
• Multicore:- Multi-core processor is IC to which ≥ 2 processors are
attached for:– enhanced performance,
– reduced power consumption,
– more efficient simultaneous processing of multiple tasks

• Challenge: Much of processing power is unharnessed.
• Goal: Attract developer to join multicore movement.
• How to Use multicore:– Multithreading :– Parallel
:– Concurrent :-

Single code is executed by several processors.
Property of machine to do side by side.
Property of program to do at same time.

HOW MULTITASK WORKS?

INITIATIVES FOR MULTICORE APPLICATIONS
• Apple’s Initiative: Mac OS X Snow Leopard (Code name-Grand
Central) gives application leverage up to 8 cores of processing power.
• RapidMind: Offers software development platform that allows
developers to exploit quad-core AMD Opteron & Intel Xeon Processors.
• Nvidia Graphics Processor: Provided Parallel Programming Language
& libraries, called ‘Cuba’.
• Java Environment for Parallel Real-Time Development (Jeopard):
Java based Real-time applications on multicore chips is planned.
• Khronos Group’s Compute Working Group:- Royalty Free Standard
for building graphical multicore applications. (OpenCL)
• Microsoft: Parallel computing initiative.
• Intel’s Parallel Studio for C & C++.
• Other Vendors:- Codeplay, Polycore Software, Click Arts etc.

CHALLENGES FOR INDUSTRY
• Image, Graphics, videos & business applications will be first to benefit,
as they can be broken down in to parts for simultaneous processing.
• But, for most complex tasks, it is difficult to find single approach for
identifying sequence of parallel computations.
• Most software are written for single core chips.
• The performance of Off-the-shelf applications reduces after 4 cores.
• Lack of multicore programming tools for mainstream developers.
• Race Conditions:- Output of one calculation depends on various
elements of a task being completed in certain order.
• Programs having interdependencies of tasks.
• The tools available currently needs manual interventions for identifying
code that can be parallelized.
• Hardware issues such as Memory Latency & Slow Bus Speed due to
limited Memory Bandwidth.

REFERENCES
1)

2)

3)

4)

“Taking advantage of multicore PCs” by Tom Kaneshige, Source URL (retrieved on
2013-05-25 11:32): http://www.infoworld.com/d/developer-world/taking-advantagemulticore-pcs-711.
“Multicore chips pose next big challenge for industry” by Agam Shah, Source URL
(retrieved on 2013-05-26 12:55): http://www.infoworld.com/d/developerworld/multicore-chips-pose-next-big-challenge-industry-105
“Modern multicore and the next generation of IT” by Paul Venezia, Source URL
(retrieved on 2013-05-26 01:14): http://www.infoworld.com/d/computerhardware/modern-multicore-and-the-next-generation-it-325
“The Trouble With Multicore” by David Patterson, Source URL (retrieved on 2013-0526 15:37):
http://www.ann.ece.ufl.edu/courses/eel4713_12spr/articles/The_Trouble_With_Multicor
e-IEEE_Spectrum.pdf

8

PPT downloaded from www.pravinkolhe.com


Slide 8

EPGP-EG 2013-14, IIM Indore.
Course Name : EMERGING TECHNOLOGOES

“Multicore Chips”
Group Number: 3
Group Members:
1) Mr. Arbind Kumar
2) Ms. Archana Shinde
3) Mr. Pravin Kolhe
4) Mr. Krushna Chnadra Satapathy

(EG2013-04)
(EG2013-06)
(EG2013-17)
(EG2013-18)

INTRODUCTION
• Processor:- Microchip implanted in a CPU's hard drive that processes
instructions sent to it by the computer and software programs.
• Multicore:- Multi-core processor is IC to which ≥ 2 processors are
attached for:– enhanced performance,
– reduced power consumption,
– more efficient simultaneous processing of multiple tasks

• Challenge: Much of processing power is unharnessed.
• Goal: Attract developer to join multicore movement.
• How to Use multicore:– Multithreading :– Parallel
:– Concurrent :-

Single code is executed by several processors.
Property of machine to do side by side.
Property of program to do at same time.

HOW MULTITASK WORKS?

INITIATIVES FOR MULTICORE APPLICATIONS
• Apple’s Initiative: Mac OS X Snow Leopard (Code name-Grand
Central) gives application leverage up to 8 cores of processing power.
• RapidMind: Offers software development platform that allows
developers to exploit quad-core AMD Opteron & Intel Xeon Processors.
• Nvidia Graphics Processor: Provided Parallel Programming Language
& libraries, called ‘Cuba’.
• Java Environment for Parallel Real-Time Development (Jeopard):
Java based Real-time applications on multicore chips is planned.
• Khronos Group’s Compute Working Group:- Royalty Free Standard
for building graphical multicore applications. (OpenCL)
• Microsoft: Parallel computing initiative.
• Intel’s Parallel Studio for C & C++.
• Other Vendors:- Codeplay, Polycore Software, Click Arts etc.

CHALLENGES FOR INDUSTRY
• Image, Graphics, videos & business applications will be first to benefit,
as they can be broken down in to parts for simultaneous processing.
• But, for most complex tasks, it is difficult to find single approach for
identifying sequence of parallel computations.
• Most software are written for single core chips.
• The performance of Off-the-shelf applications reduces after 4 cores.
• Lack of multicore programming tools for mainstream developers.
• Race Conditions:- Output of one calculation depends on various
elements of a task being completed in certain order.
• Programs having interdependencies of tasks.
• The tools available currently needs manual interventions for identifying
code that can be parallelized.
• Hardware issues such as Memory Latency & Slow Bus Speed due to
limited Memory Bandwidth.

REFERENCES
1)

2)

3)

4)

“Taking advantage of multicore PCs” by Tom Kaneshige, Source URL (retrieved on
2013-05-25 11:32): http://www.infoworld.com/d/developer-world/taking-advantagemulticore-pcs-711.
“Multicore chips pose next big challenge for industry” by Agam Shah, Source URL
(retrieved on 2013-05-26 12:55): http://www.infoworld.com/d/developerworld/multicore-chips-pose-next-big-challenge-industry-105
“Modern multicore and the next generation of IT” by Paul Venezia, Source URL
(retrieved on 2013-05-26 01:14): http://www.infoworld.com/d/computerhardware/modern-multicore-and-the-next-generation-it-325
“The Trouble With Multicore” by David Patterson, Source URL (retrieved on 2013-0526 15:37):
http://www.ann.ece.ufl.edu/courses/eel4713_12spr/articles/The_Trouble_With_Multicor
e-IEEE_Spectrum.pdf

8

PPT downloaded from www.pravinkolhe.com


similar documents