docsity.com
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
Guidelines and tips
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
Multi Core Computing - Advanced Theory of Computation - Lecture Slides, Slides of Theory of Computation
This lecture is part of complete lecture series on Advanced Theory of Computation. Key points in this lecture are: Multi Core Computing, Flood of Computer Tasks, Emergence of Multi Core Processor, Dual Core Processor, Automatic Parallelization, Replicate Multiple Processor, Superscalar Operations, Multithreading, Ibm Cell Processor, Smp
Typology: Slides
2013/2014
1 / 42
This page cannot be seen from the preview
Don't miss anything!
Related documents
Partial preview of the text
Download Multi Core Computing - Advanced Theory of Computation - Lecture Slides and more Slides Theory of Computation in PDF only on Docsity!
Introduction
Motivation for Multi-Core
What is multi-core processor?
Properties of Multi-core systems
Applications benefit from multi-core
Multiprocessor memory types
Multi-core design^
Symmetric multi-core processor
Asymmetric multi-core processor
Advantages & disadvantages of multi-core
Flood of Computer Tasks(1990’s)^
Increasing of Computer user
Server management
→
Construct the database
▪
We need better performance of PC or Server. →
These demands accelerate the development of
microprocessor.
Emergence of Multi-core Processor(2000’s)^
Limits of improvement of single core
Turn over of the idea that improve the MP technology
▪
Put execution cores in one die
Exploits increased feature-size and density
Increases functional units per chip (spatialefficiency)
Limits energy consumption per operation
Constrains growth in processor complexity
dual-core processor contains two independentmicroprocessors.
A dual core set-up is somewhat comparable tohaving multiple, separate processors installedin the same computer, but because the twoprocessors are actually plugged into the samesocket, the connection between them is faster.Ideally, a dual core processor is nearly twiceas powerful as a single core processor. Inpractice, performance gains are said to beabout fifty percent: a dual core processor islikely to be about one-and-a-half times aspowerful as a single core processor.
A multi-core processor implements multiprocessing in a singlephysical package. Cores in a multi-core device may be coupledtogether tightly or loosely. For example, cores may or may notshare caches, and they may implement message passing orshared memory inter-core communication methods. Commonnetwork topologies to interconnect cores include: bus, ring, 2-dimentional mesh, and crossbar.
All cores are identical in
symmetric
multi-core systems and they
are not identical in a
symmetric
multi-core systems. Just as with
single-processor systems, cores in multi-core systems mayimplement architectures such as superscalar, vector processing,or multithreading.
with a shared on-chip cache memory,communication events can be reduced to just ahandful of processor cycles.
therefore with low latencies, communicationdelays have a much smaller impact on overallperformance.
threads can also be much smaller and still beeffective.
automatic parallelization more feasible.
Cores will be shared with a wide range of otherapplications dynamically.
Load can no longer be considered symmetric acrossthe cores.
Cores will likely not be asymmetric as acceleratorsbecome common for scientific hardware.
Source code will often be unavailable, preventingcompilation against the specific hardwareconfiguration.
Replicate multiple processor cores on a single die.
The cores fit on a single processor socket.
c o r e 1
c o r e 2
c o r e 3
c o r e 4
severalthreads
severalthreads
severalthreads
severalthreads
Most major OS support multi-core today.
OS perceives each core as a separateprocessor.
OS scheduler maps threads/processesto different cores.
Editing a photo while recording a TV show through adigital video recorder.
Downloading software while running an anti-virusprogram.
“Anything that can be threaded today will mapefficiently to multi-core”.
BUT: some applications difficult to parallelize.
Save the room of motherboard
Two single cores
In one die
We can use this room more efficiently
Simplicity ▪
We need additional systems to control the severalsingle cores.
Economical efficiency
A dual-core is much cheaper than two single cores
Shared memory: In this model, there is one (large) commonshared memory for all processors.