Download 28-Power - inst.eecs.berkeley.edu

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

History of electric power transmission wikipedia , lookup

Grid energy storage wikipedia , lookup

Alternating current wikipedia , lookup

Electric power system wikipedia , lookup

Power over Ethernet wikipedia , lookup

Audio power wikipedia , lookup

Switched-mode power supply wikipedia , lookup

Electrification wikipedia , lookup

Standby power wikipedia , lookup

Wireless power transfer wikipedia , lookup

Distributed generation wikipedia , lookup

Life-cycle greenhouse-gas emissions of energy sources wikipedia , lookup

Power engineering wikipedia , lookup

AC adapter wikipedia , lookup

Transcript
Power
 Motivation for design constraints of power consumption
 Power metrics
 Power consumption analysis in CMOS
 How can a logic designer control power?
CS 150 - Spring 2007 – Lec #28 – Power - 1
“X-Internet” Beyond the PC
Internet Computers
Internet Users
500
Million
Today’s Internet
1.5 Billion
Automobiles
700 Million
Telephones
4 Billion
X-Internet
Electronic Chips
60 Billion
Forrester Research, May 2001
CS 150 - Spring 2007 – Lec #28 – Power - 2 Revised 2007
“X-Internet” Beyond the PC
Millions
15000
10000
PC
Internet
5000
CS 150 - Spring 2007 – Lec #28 – Power - 3
10
20
20
08
20
07
20
06
20
05
20
04
20
03
20
02
20
20
01
0
09
X
Internet
Year
Forrester Research, May 2001
Cell Phones
 Phone w/voice command,
voice dialing, intelligent
text for short msgs
 MP3 player + headset,
digital voice recorder
 “Mobile Internet” with a
built-in WAP Browser
 Java-enabled, over the air
programmable
 Bluetooth + GPRS
Siemens SL45i
Ericsson T68
 Enhanced displays +
embedded cameras
CS 150 - Spring 2007 – Lec #28 – Power - 4
Shape of Things
 Phone + Messenger + PDA Combinations
E.g., Blackberry 5810 Wireless Phone/Handheld
Integration of PDA + Telephone
PLUS Gateway to Internet and Enterprise applications
1900 MHz GSM/GPRS (Euroversion at 900 Mhz)
SMS Messaging, Internet access
QWERTY Keyboard, 20 line display
JAVA applications capable
8 MB flash + 1 MB SRAM
CS 150 - Spring 2007 – Lec #28 – Power - 5
Shape of Things to Come
 Danger “Hiptop”
Full-featured mobile phone w/Internet Access
Email + attachments/instant messaging + PIM
Digital camera accessory
End-to-end integration of voice + data apps
Media-rich UI for graphics + sound
Large screen + QWERTY keyboard
Data nav: keyboard or push wheel
Affordable (under $200)
MIDI synthesizer for quality sound
Multi-tasking of user actions
Customizable ring tones and alerts
to personalize hiptop experience
CS 150 - Spring 2007 – Lec #28 – Power - 6
Important (Wireless)
Technology Trends
“Spectral Efficiency”:
More bits/m3
Rapidly increasing
transistor density
Rapidly declining
system cost
CS 150 - Spring 2007 – Lec #28 – Power - 7
In the Physical World: Sensor Devices
CS 150 - Spring 2007 – Lec #28 – Power - 8
Important (Wireless)
Technology Trends
Speed-Distance-Cost
Tradeoffs
Rapid Growth: Machine-toMachine Devices
(mostly sensors)
CS 150 - Spring 2007 – Lec #28 – Power - 9
Why Worry About Power?
 Portable devices:
 Handhelds, laptops, phones, MP3 players, cameras, … all need to run for
extended periods on small batteries without recharging
 Devices that need regular recharging or large heavy batteries will lose out to
those that don’t.
 Power consumption important even in “tethered” devices
 System cost tracks power consumption:
 Power supplies, distribution, heat removal
 Power conservation, environmental concerns
 In 10 years, have gone from minimal consideration of power consumption
to (designing with power consumption as a primary design constraint!
CS 150 - Spring 2007 – Lec #28 – Power - 10
Basics
 Power supply provides energy for charging and discharging wires and
transistor gates. The energy supplied is stored & then dissipated as
heat.
Power: Rate of work being done wrt time
Rate of energy being used
P  dw / dt
Units:
P  E t
Watts = Joules/seconds
 If a differential amount of charge dq is given a differential increase
in energy dw, the potential of the charge is increased by:
 By definition of current:
V  dw / dq
I  dq / dt
dw dq
dw / dt 

 P V I
dq dt
t
w
 Pdt

total energy
A very practical
formulation!
If we would like
to know total energy
CS 150 - Spring 2007 – Lec #28 – Power - 11
Basics
 Warning! In everyday language, the term “power” is
used incorrectly in place of “energy”
 Power is not energy
 Power is not something you can run out of
 Power can not be lost or used up
 It is not a thing, it is merely a rate
 It can not be put into a battery any more than
velocity can be put in the gas tank of a car
CS 150 - Spring 2007 – Lec #28 – Power - 12
This is how electric tea pots work ...
Heats 1 gram of water
0.24 degree C
0.24 Calories per Second
1A
1V
+
-
1 Joule of Heat
Energy per Second
1 Ohm
Resistor
20 W rating: Maximum power
the package is able to transfer
to the air. Exceed rating and
resistor burns.
CS 150 - Spring 2007 – Lec #28 – Power - 13
Cooling an iPod nano ...
Like a resistor, iPod relies
on passive transfer of heat
from case to the air
Why? Users don’t want
fans in their pocket ...
To stay “cool to the touch” via passive cooling,
power budget of 5 W
If iPod nano used 5W all the time, its battery would last
15 minutes ...
CS 150 - Spring 2007 – Lec #28 – Power - 14
Powering an iPod nano (2005 edition)
Battery has 1.2 W-hour
rating: Can supply
1.2 W of power for 1 hour
1.2 W / 5 W = 15 minutes
More W-hours require bigger battery
and thus bigger “form factor” -it wouldn’t be “nano” anymore!
Real specs for iPod nano :
14 hours for music,
4 hours for slide shows
85 mW for music
300 mW for slides
CS 150 - Spring 2007 – Lec #28 – Power - 15
0.55 ounces
12 hour
battery life
$79.00
1 GB
CS 150 - Spring 2007 – Lec #28 – Power - 16
20 hour battery life for audio,
6.5 hours for movies (80GB version)
Up from 14
24 hour
battery life hours for 2005
iPod nano
for audio
Thinner than 2005 iPod nano
5 hour
Up from 4
battery life hours for 2005
for photos
iPod nano
12 hour
battery life
CS 150 - Spring 2007 – Lec #28 – Power - 17
Notebooks ... now most of the PC market
Apple MacBook -- Weighs 5.2 lbs
8.9 in
1 in
12.8 in
Performance: Must be “close enough” to desktop
performance ... many people no longer own a desktop
Size and Weight: Ideal: paper notebook
Heat: No longer “laptops” -- top may get “warm”,
bottom “hot”. Quiet fans OK
CS 150 - Spring 2007 – Lec #28 – Power - 18
Battery: Set by size and weight limits ...
Battery rating:
55 W-hour
46x energy than iPod nano.
iPod lets you listen to music
for 14 hours!
At 2.3 GHz,
Intel Core Duo
CPU consumes 31
W running a
heavy load under 2 hours
battery life! And,
just for CPU!
Almost full 1 inch
At 1 GHz, CPU consumes
depth. Width and
13 Watts. “Energy saver”
height set by available
option uses this mode ...
space, weight.
CS 150 - Spring 2007 – Lec #28 – Power - 19
Battery Technology
 Battery technology has developed slowly
 Li-Ion and NiMh still the dominate technologies
 Batteries still contribute significantly to the weight
of mobile devices
Nokia 61xx 33%
Handspring
PDA - 10%
Toshiba Portege
3110 laptop - 20%
CS 150 - Spring 2007 – Lec #28 – Power - 20
55 W-hour battery stores
the energy of
1/2 a stick of dynamite.
If battery short-circuits,
catastrophe is possibleCS...
150 - Spring 2007 – Lec #28 – Power - 21
CPU Only Part of Power Budget
2004-era notebook running a
full workload.
“other”
GPU
LCD
Backlight
CPU
If our CPU took no power
at all to run, that would
only double battery life!
LCD
CS 150 - Spring 2007 – Lec #28 – Power - 22
Servers: Total Cost of Ownership (TCO)
Machine rooms
are expensive …
removing heat
dictates how
many servers to
put in a machine
room.
Reliability: running computers hot
makes them fail more often
CS 150 - Spring 2007 – Lec #28 – Power - 23
Electric bill adds
up! Powering the
servers +
powering the air
conditioners is a
big part of TCO
Thermal Image of Typical Cluster Rack
Rack
Switch
QuickTime™ and a
TIFF (LZW) decompressor
are needed to see this picture.
M. K. Patterson, A. Pratt, P. Kumar,
“From UPS to Silicon: an end-to-end evaluation of datacenter efficiency”, Intel Corporation
CS 150 - Spring 2007 – Lec #28 – Power - 24
How Do We Measure and Compare
Power Consumption?
 One popular metric for microprocessors is: MIPS/watt
 MIPS, millions of instructions per second
Typical modern value?
 Watt, standard unit of power consumption
Typical value for modern processor?
 MIPS/watt reflects tradeoff between performance and power
 Increasing performance requires increasing power
 Problem with “MIPS/watt”
MIPS/watt values are typically not independent of MIPS
• Techniques exist to achieve very high MIPS/watt values, but at
very low absolute MIPS (used in watches)
Metric only relevant for comparing processors with a similar
performance
 One solution, MIPS2/watt. Puts more weight on performance
CS 150 - Spring 2007 – Lec #28 – Power - 25
Metrics
 How does MIPS/watt relate to energy?
 Average power consumption = energy / time
 MIPS/watt = instructions/sec / joules/sec = instructions/joule
 Equivalent metric (reciprocal) is energy per operation (E/op)
 E/op is more general - applies to more that processors
 also, usually more relevant, as batteries life is limited by total
energy draw.
 This metric gives us a measure to use to compare two alternative
implementations of a particular function.
CS 150 - Spring 2007 – Lec #28 – Power - 26
Power in CMOS
Vdd
Switching Energy:
Vdd
pullup
network
energy used to
switch a node
i(t)
v(t)
0 1
Calculate energy
dissipated in pullup:
pulldown C
network
v(t)
t0
t1
GND
t1
t1
t1
t0
t0
t0
Esw   P(t )dt   (Vdd  v)  i(t )dt   (Vdd  v)  c (dv dt ) dt 
t1
t1
t0
t0
 cVdd  dv  c  v  dv  cVdd  1 2cVdd  1 2 cVdd
Energy supplied
2
Energy stored
2
2
Energy dissipated
An equal amount of energy is dissipated on pulldown
CS 150 - Spring 2007 – Lec #28 – Power - 27
Switching Power
 Gate power consumption:
 Assume a gate output is switching its output at a rate of:
activity factor
 f
clock rate
(probability of switching on
any particular clock period)
1/f
Pavg  E t  switching rate  Esw
Therefore:
Pavg    f 1 2 cVdd
 Chip/circuit power consumption:
2 Pavg
Pavg  n  avg  f 1 2 cavgVdd
2
number of nodes (or gates)
CS 150 - Spring 2007 – Lec #28 – Power - 28
clock f
Other Sources of Energy Consumption
 “Short Circuit” Current:
 Junction Diode Leakage :
Vout
I
Vin
Vin
Vout
Transistor drain regions
“leak” charge to substrate.
I
I
Vin
10-20% of total chip power
Diode
Characteristic
V
~1nWatt/gate
few mWatts/chip
CS 150 - Spring 2007 – Lec #28 – Power - 29
Other Sources of Energy Consumption
 Consumption caused by “DC leakage current” (Ids leakage):
Ids
Vin=0
Vout=Vdd
Ioff
Transistor s/d conductance
never turns off all the way
Vgs
Vth
Low voltage processes much worse
 This source of power consumption is becoming increasing significant
as process technology scales down
 For 90nm chips around 10-20% of total power consumption
Estimates put it at up to 50% for 65nm
CS 150 - Spring 2007 – Lec #28 – Power - 30
Controlling Energy Consumption: What
Control Do You Have as a Designer?
 Largest contributing component to CMOS power consumption is
switching power:
Pavg  n  avg  f 1 2 cavgVdd
2
 Factors influencing power consumption:
 n: total number of nodes in circuit
 : activity factor (probability of each node switching)
 f: clock frequency (does this effect energy
consumption?)
 Vdd: power supply voltage
 What control do you have over each factor?
 How does each effect the total Energy?
Our design projects do not optimize for power consumption
CS 150 - Spring 2007 – Lec #28 – Power - 31
Scaling Switching Energy per Gate
Moore’s Law
at work …
Due to reduced
V and C (length
and width of Cs
decrease, but
plate distance
gets smaller)
Recent slope
reduced
because V is
scaled less
aggressively
From: “Facing the Hot Chips Challenge Again”, Bill Holt, Intel, presented at Hot Chips 17, 2005.
CS 150 - Spring 2007 – Lec #28 – Power - 32
Device Engineers Trade Speed and Power
We can reduce CV2 (Pactive)
by lowering Vdd
We can increase speed
by raising Vdd and
lowering Vt
We can reduce leakage
(Pstandby) by raising Vt
From: Silicon Device Scaling to the Sub-10-nm Regime
Meikei Ieong,1* Bruce Doris,2 Jakub Kedzierski,1 Ken Rim,1 Min Yang1
CS 150 - Spring 2007 – Lec #28 – Power - 33
Customize processes for product types ...
From: “Facing the Hot Chips Challenge Again”, Bill Holt, Intel, presented at Hot Chips 17, 2005.
CS 150 - Spring 2007 – Lec #28 – Power - 34
Intel: Comparing 2 CPU Generations ...
Find enough
tricks, and you
can afford to
raise Vdd a
little so that
you can raise
the clock
speed!
Clock speed
unchanged ...
Lower Vdd, lower C,
but more leakage
Design tricks:
architecture & circuits
CS 150 - Spring 2007 – Lec #28 – Power - 35