31
The Free Lunch Is Over

The free lunch is over

Embed Size (px)

Citation preview

Page 1: The free lunch is over

The Free Lunch Is Over

Page 2: The free lunch is over

The Free Lunch Is OverA Fundamental Turn Toward Concurrency in Software

(By Herb Sutter)

Page 3: The free lunch is over

Presenter

Muhammad Rizwan

Sr. Software Engineer

[email protected]

+92 51 111 859 859 Ext 108

Page 4: The free lunch is over

Agenda

• Moore's Law

• Amdahl's law

• Performance Factors

• Free Lunch

• Power Wall

• Free Lunch is Over

• Conclusion

• Future

• References

Page 5: The free lunch is over

Moors Law

Page 6: The free lunch is over

Moor’s Law

• Over the history of computing hardware,

the number of transistors on integrated

circuits doubles approximately every two

years Transistor

Count

If the current trend continues

to 2020, the number of

transistors would reach 32

billion.

Page 7: The free lunch is over

Amdahl's law

Page 8: The free lunch is over

Amdahl's law

The speedup of a

program using

multiple processors

in parallel

computing is limited

by the sequential

fraction of the

program. For example, if 95% of the program can

be parallelized, the theoretical maximum

speedup using parallel computing would

be 20× as shown in the diagram, no

matter how many processors are used

Page 9: The free lunch is over

Performance Factors

Page 10: The free lunch is over

Performance Factors

Clock Speed

Execution Optimization

Cache

Page 11: The free lunch is over

Free Lunch

What Free Lunch Is Actually?

Page 12: The free lunch is over

Free Lunch

Programmers haven't really had to worry much about performance or concurrency because of Moore's Law

The traditional approach to application performance was to simply wait for the next generation of processor; most software developers did not need to invest in performance tuning, and enjoyed a “free lunch” from hardware improvements.

Why we did not see 4GHz

processors in Market?

Page 13: The free lunch is over

Power Wall

Page 14: The free lunch is over

Power Wall

Page 15: The free lunch is over
Page 16: The free lunch is over

Power Wall

Page 17: The free lunch is over

Power Wall

This Free Lunch is Over

Right Now…

Page 18: The free lunch is over

Free Lunch is Over

Your free lunch will soon be over. What can you do about it?

What are you doing about it?

Page 19: The free lunch is over

The major processor manufacturers and

architectures, from Intel and AMD to Sparc and

PowerPC, have run out of room with most of their

traditional approaches to boosting CPU performance.

Instead of driving clock speeds and straight-line

instruction throughput ever higher, they are instead

turning toward hyper threading and multicore

architectures

Chip designers are under so much pressure to

deliver ever-faster CPUs that they’ll risk changing

the meaning of your program, and possibly break

it, in order to make it run faster

Page 20: The free lunch is over

Conclusion

Page 21: The free lunch is over

The free lunch is over

Performance Free Lunch

Moore’s law has changed

Futurists such as Ray Kurzweil, Bruce Sterling, and Vernor Vinge believe that the

exponential improvement described by Moore's law will ultimately lead to a technological

singularity: a period where progress in technology occurs almost instantly.

Amdahl’s law demands shift in software

concepts

Factors Hyper threading

Multicore

Cache

Page 22: The free lunch is over

Future

Page 23: The free lunch is over
Page 24: The free lunch is over
Page 25: The free lunch is over
Page 26: The free lunch is over

Now ball is in programmers

court

Page 27: The free lunch is over

-> With multicore processors, programs written in sequential mode will no longer

surf on the wave of this generation processors.

-> To surf in new wave, programs need to be well written parallel.

-> Programming language and system will increasingly be forced to deal well

with concurrency.

-> Concurrency is the next major revolution in how we write software.

-> Applications will increasingly need to be concurrent if they want to fully exploit

continuing exponential CPU throughput gains.

-> Efficiency and performance optimization will get more, not less, important.

Page 28: The free lunch is over

New Law

Only Parallel Applications

Will Survive

Page 29: The free lunch is over

References

Page 30: The free lunch is over

References

The Free Lunch Is Over By Herb Sutter

A Fundamental Turn Toward Concurrency in Software

The Free Lunch Is Over By Ricardo Hermann , Thadeo Carmo

Developing Concurrent Software

Is the free lunch really over? Scalability in Many-core Systems By Michael Wrinn

The Price of Free LunchProgramming in Multicore Era

Page 31: The free lunch is over

Any Question?