O SlideShare utiliza cookies para otimizar a funcionalidade e o desempenho do site, assim como para apresentar publicidade mais relevante aos nossos usuários. Se você continuar a navegar o site, você aceita o uso de cookies. Leia nosso Contrato do Usuário e nossa Política de Privacidade.
O SlideShare utiliza cookies para otimizar a funcionalidade e o desempenho do site, assim como para apresentar publicidade mais relevante aos nossos usuários. Se você continuar a utilizar o site, você aceita o uso de cookies. Leia nossa Política de Privacidade e nosso Contrato do Usuário para obter mais detalhes.
A Seminar on
AN ARCHITECTURE FOR DARK SILICON AGE
Dept. of E&CE, GEC RAICHUR
HOD & Associate Professor
Dept. Of Electronics And Communication Engineering
27-Mar-16 Dept. of E&CE, GEC RAICHUR 3
3) DARK SILICON
4) THE UTILIZATION WALL
5) ARCHITECTURE OF GREENDROID
6) CONSERVATION CORES
27-Mar-16 Dept of E&CE, GEC RAICHUR 4
1) Mobile phone are now replaced by Smart phone that run on the
open source operating systems such as Android or IOS.
2) The Green Droid mobile application processor is a 45-nm multi
core research prototype that targets the Android mobile phone
3) It can execute general purpose mobile programs with 11 times
less energy than today’s most energy-efficient designs, at similar
or better performance levels.
4) GREENDROID will serve as a prototype for mobile application
processors in the next ﬁve to ten years.
27-Mar-16 Dept. of E&CE, GEC RAICHUR 5
5) It has a specially built structure that can analyze a current
Android phone and determine which apps, and which CPU
circuits the phone is using the most.
6) Then it can dream up a processor design that best takes
advantage of those usage habits, creating a CPU that’s
both faster and more energy efﬁcient.
27-Mar-16 Dept. of E&CE, GEC RAICHUR 6
1) A key technological problem for microprocessor architects
is the utilization wall.
2) The utilization wall says that, with each process generation, the
percentage of transistors that a chip design can switch at full
frequency drops exponentially because of power constraints.
3) A direct consequence of this is Dark Silicon.
27-Mar-16 Dept. of E&CE, GEC RAICHUR 7
1) How many transistors you can actually use simultaneously
given your power budget – the gap between area gains and
2) Dark silicon is necessary, because engineers are unable
to reduce chips operating voltages any further to offset
increases in power consumption and waste heat produced
by smaller, faster chips .
3) This dark silicon limits the utilization of the application
processors to the fullest.
THE UTILIZATION WALL
27-Mar-16 Dept. of E&CE, GEC RAICHUR 8
a) Transistor and power budgets no longer balanced
b) Exponentially increasing problem.
a) Replicated small data path
b) More ‘Dark Silicon’ than active
Observations in the wild
a) Flat frequency curve
b) Increasing cache/processor ratio
ARCHITECTURE OF GREENDROID
27-Mar-16 Dept. of E&CE, GEC RAICHUR 9
1) The GreenDroid architecture uses specialized, energy-
efficient processors, called conservation cores, or c-cores to
execute frequently used portions of the application code.
2) Collectively, the c-cores span approximately 95 percent of
the execution time of team’s test Android-based workload.
27-Mar-16 Dept. of E&CE, GEC RAICHUR 11
The system comprises
an array of 16 non
27-Mar-16 Dept. of E&CE, GEC RAICHUR 12
Each tile holds components
common to every tile
the CPU, on-chip network
(OCN) and shared Level 1
(L1) data cache—and
provides space for multiple
conservation cores (c-cores)
of various sizes.
27-Mar-16 Dept. of E&CE, GEC RAICHUR 13
The c-cores are
tightly coupled to
the host CPU via the
L1 data cache and a
27-Mar-16 Dept. of E&CE, GEC RAICHUR 14
Specialized Cores For Reducing Energy
Automatically generated from hot regions of program source
Patching support future proofs HW
Fully Automated Tool Chain
Drop-in replacements for code
Hot code implemented by C-Core, cold code runs on host CPU
HW generation/SW integration
Up to 16x for targeted hot code
27-Mar-16 Dept. of E&CE, GEC RAICHUR 16
Figure shows the projected energy savings in GreenDroid and the
origin of these savings.
The savings come from two sources
1) First, c-cores don’t require instruction fetch, instruction
decode, a conventional register file, or any of the associated
structures. Removing these reduces energy consumption by 56
2) The second source of savings (35 % of energy) comes from
the specialization of the c-cores data path.
3) The result is that average per-instruction energy drops from 91
pJ per instruction to just 8 pJ per instruction.
27-Mar-16 Dept. of E&CE, GEC RAICHUR 18
1) Less power consumption.
2) Most faster than android
3) More Efficient
4) Easy to use
5) Life time is more
6) More application is available in GreenDroid than android.
27-Mar-16 Dept. of E&CE, GEC RAICHUR 19
1) GreenDroid is a boon for dark silicon problem.
2) Can be used for desktop processors.
3) Power savings can be applied to other program, increasing
27-Mar-16 Dept. of E&CE, GEC RAICHUR 20
Over the next five to 10 years, the breakdown of conventional silicon
scaling and the resulting utilization wall will exponentially increase the
amount of dark silicon in both desktop and mobile processors.
The Green Droid prototype demonstrates that c-cores offer a new
technique to convert dark silicon into energy savings and increased
parallel execution under strict power budgets.
The estimate that the prototype will reduce processor energy consumption
by 91 percent for the code that c-cores target, and result in an overall
savings of 7.4 times.
27-Mar-16 Dept. of E&CE, GEC RAICHUR 21
1) G. Venkatesh et al., Conservation Cores: Reducing the Energy of
Mature Computations,: Proc.15th Intl Conf.Architectural Support
for Programming Languages and Operating Systems, ACM Press,
2010, pp. 205/-218. 15th Intl. Conf. Architectural Support for Prog.
Languages and Op. Sys., Mar. 2010.
2) N. Goulding et al., GreenDroid: A Mobile Application Processor
for a Future of Dark Silicon, HotChips, 2010.
3) M. Taylor et al., The Raw Processor: A Scalable 32 bit Fabric for
General Purpose and Embedded Computing, HotChips, 2001.