Upload
tranhanh
View
226
Download
0
Embed Size (px)
Citation preview
Title 44pt sentence case
Affiliations 24pt sentence case
20pt sentence case
© ARM 2017
ARM instruction sets and CPUs for wide-ranging applications
Chris Turner
ARM Tech Forum Taipei
Director, CPU technology marketing
July 4th 2017
© ARM 2017 2
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
ARM computing is everywhere
6.6BnARM-based
embedded
chips shipped in
2016
#1shipping GPU in
the world is
Mali
> 5Bnpeople using
ARM-based
mobile phones
100BnARM-based
chips to date
© ARM 2017 3
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
Total computing from sensors to server
Efficiency
enabling
distributed
compute
Efficiency
delivering
TCO
benefits
Demands more performance with greater emphasis on efficiency and power management
Performance,
efficiency
and security
© ARM 2017 4
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
Cortex-AHighest performance
Optimized for
high-level operating
systems
Cortex-RFast response
Optimized for
high performance,
hard real-time
applications
Cortex-MSmallest/lowest power
Optimized for
discrete processing
and
microcontrollers
Optimized for
physical security
SecurCoreTamper resistant
ARM CPU architecture for total computing
© ARM 2017 5
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
Total computing in mobile
© ARM 2017 6
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
Automotive
The right combination of CPUs
Consumer
Cortex-M Low power, deterministic sensing and control
Cortex-A Rich UI and OS, open system, high performance
Cortex-R Safety, performance and real-time control
IoTMobile
© ARM 2017 7
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
Automated
exceptions
Virtual memory
Protected
memory
Programmable
exception
model
ARM architecture profiled for the application
ARMv8-A
ARMv8-R
TrustZone +
Virtualization
ARMv8-M T32
A64, A32, T32
A32, T32
NEON
SIMD
TrustZone DSP
AutomotiveConsumer IoTMobile
Virtualization
© ARM 2017 8
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
Evolving architecture that supports an ecosystem
Silicon partners
Applications
Software vendors
Development tools
ARM11ARM9
© ARM 2017 9
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
ARM ecosystem – widest choice, most innovation
© ARM 2017 10
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
▪ Cortex-A CPUs cover a wide variety of markets
▪ Scale efficiently to substantially higher performance
▪ Fit even more compute in a smaller footprint with less power
Automotive, industrial
ARM application processors are everywhere
Mobile and consumer Servers and networking IoT and embedded
© ARM 2017 11
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
ARM Cortex-A portfolio
Cortex-A57
2008 – 2013 2017201620152014
Cortex-A72
A5x
Series
A3x
Series
Cortex-A9
Well-established, mid-range processor
Cortex-A5/A7
Smallest and lowest power
ARMv7-A
Cortex-A15/A17
Infrastructure performance;
mobile efficiency
Cortex-A57
Proven infrastructure performance
Cortex-A72
For all applications
Cortex-A35
Smallest, lowest power ARMv8-A
Cortex-A53
Balanced performance and
efficiency
Cortex-A73
For mobile and consumer
Cortex-A32
Smallest, lowest power 32-bit
ARMv8-A
ARMv7-A ARMv8-AARMv8-AARMv7-A
Cortex-A55
Highest efficiency mid-range processor
64/32-bit
Cortex-A75
Ground-breaking performance for
all markets
64/32-bit
32-bit64/32-bit
64/32-bit64/32-bit64/32-bit
64/32-bit
32-bit
32-bit
32-bit
Year of IP release, volume devices in the subsequent year
A7x
Series
big.LITTLE compatible
© ARM 2017 12
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
Meeting performance needs for any application
*Performance comparison using SPECINT2000 benchmark suite
Cortex-A portfolio performance comparisons*
0%
20%
40%
60%
80%
100%
120%
140%
Rela
tive
iso
-fre
quency
perf
orm
ance
Ultra high efficiency
Cortex-A5
Cortex-A7
Cortex-A32, Cortex-A35
Cortex-A portfolio performance comparisons*
Co
rtex-A
5
Co
rtex-A
7
Co
rtex-A
32,
Co
rtex-A
35
0%
20%
40%
60%
80%
100%
120%
140%
160%
180%
200%
Rela
tive
Perf
orm
ance
at Tar
get
Fre
quency
High Performance
Cortex-A9 @ 1.5GHz Cortex-A15 @ 1.9GHz
Cortex-A17 @ 1.6GHz Cortex-A57 @ 2.1GHz
Cortex-A72 @ 2.5GHz Cortex-A73 @ 2.7GHz
Cortex-A75 @ 3GHz
Co
rtex-A
9
Co
rtex-A
15
Co
rtex-A
17
Co
rtex-A
57
Co
rtex-A
72
Co
rtex-A
73
Co
rtex-A
75
0%
20%
40%
60%
80%
100%
120%
140%
Rela
tive
Iso
-fre
quency
Perf
orm
ance
Cortex-A
High Efficiency
Cortex-A9
Cortex-A35
Cortex-A53C
ort
ex-A
9
Co
tex-A
35
Co
rtex-A
53
Co
rtex-A
55
© ARM 2017 13
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
big.LITTLE – performance and efficiency for mobile
▪ Faster, more responsive systems
▪ Increased battery life
▪ Higher, longer sustained performance
▪ Tuned for mobile, consumer and embedded
Higher
performance
Superior power
efficiency
Higher compute capacity
Cache Cache
Cache Coherent Interconnect
Interrupt controller
big clusterLITTLE
cluster
© ARM 2017 14
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
DynamIQ plus big.LITTLE
1b+4L1b+3L1b+2L1b+7L 2b+6L 4b+4L
cluster architecture
1 to 8 CPUsCPU1
Snoop
filter
Power
manager
L3
Cache
Bus
i/f
ACP and
peripheral
port i/f
CPU8
Asynchronous bridges
DynamIQ Shared Unit (DSU)
L1C & L2C L1C & L2C
© ARM 2017 15
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case▪ Enhanced multimedia user experience
▪ Intensive data processing
▪ Parallel computing
▪ AI ML CV
NEON – flexible high-performance data computing
Delivers
Performance
Wide Software
Support for TTM
Seamless
Development
© ARM 2017 16
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
ARM compute library leverages NEON for CV/ML
▪ Optimized low-level functions for CPU and GPU
▪ Most popular CV and ML functions
▪ Common functions underpinning popular ML frameworks
▪ Faster performance and development cycle
▪ Public availability as open source
Key Functions
categories
Basic arithmetic
Convolutions
Colour manipulation
Feature detection
Neural network
GEMM
Pyramids
Filters
Image reshaping
Mathematical functions
CPU version - tested on Huawei Mate 8 (single threaded)
© ARM 2017 17
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
ARM TrustZone protecting billions of devices
Device Security
Communications Security
Lifecycle Security
trusted software
Crypto
Root of Trust
non-trusted
trusted
trusted hardware
secure
system
secure
storage
Authentication
Mobile Payment
Content
Protection
Enterprise
Security
© ARM 2017 18
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
Cortex-AHighest performance
Optimized for
high-level operating
systems
Cortex-RFast response
Optimized for
high performance,
hard real-time
applications
Cortex-MSmallest/lowest power
Optimized for
discrete processing
and
microcontrollers
Optimized for
physical security
SecurCoreTamper resistant
ARM architecture for total computing
© ARM 2017 19
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
Cortex-AHighest performance
Optimized for
high-level operating
systems
Cortex-RFast response
Optimized for
high performance,
hard real-time
applications
Cortex-MSmallest/lowest power
Optimized for
discrete processing
and
microcontrollers
Optimized for
physical security
SecurCoreTamper resistant
ARM architecture for total computing
© ARM 2017 20
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
ARMv8-M – TrustZone for IoT
▪ Optimized for microcontrollers
▪ Simplified development
▪ Real-time response
▪ Efficient security
▪ Functional safety
Cort
ex-M
Cry
pto
Cell
Core
Lin
k
© ARM 2017 21
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
ARM Cortex-M portfolio
High
performance
Performance
efficiency
Lowest
power & area
Cortex-M23
TrustZone in smallest area, lowest power
Cortex-M33Flexibility,
control and
DSP with
TrustZone
ARMv8-M
Cortex-M0
Lowest cost,
low power
Cortex-M0+
Highest energy
efficiency
Cortex-M4
Mainstream
control and
DSP
Cortex-M3
Performance
efficiency
Cortex-M7Maximum
performance,
control and
DSP
ARMv6-M
ARMv7-M
25BnTotal units
shipped*
*Data as of Dec. 2016
© ARM 2017 22
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
Smallest footprint Maximum efficiency Constrained applications
Cortex-M23 – TrustZone in the smallest footprint
+50%more
efficient
than Cortex-M33
75%smaller
than Cortex-M33
Making energy harvesting IoT viable
Secure
Smart lock
Ultra efficient
Smart bandage
Safe
Medical nanorobot
Ubiquitous
Asset tracking
same ultra-high efficiency
as Cortex-M0+
© ARM 2017 23
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
Extremely compact Configurable and extensible Widely applicable
Cortex-M33 – efficiency, security and flexibility
Cortex-A5
Cortex-M33
than Cortex-A580% smaller
Cortex-A5, Cortex-M33 size based on 40nm
TrustZone
Base core
DSP
FPU
Co-p
roc
i/f
© ARM 2017 24
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
Cortex-AHighest performance
Optimized for
high-level operating
systems
Cortex-RFast response
Optimized for
high performance,
hard real-time
applications
Cortex-MSmallest/lowest power
Optimized for
discrete processing
and
microcontrollers
Optimized for
physical security
SecurCoreTamper resistant
ARM CPU architecture for total computing
© ARM 2017 25
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
Cortex-R – real-time, high performance, safety
▪ HDD and SSD storage
▪ 3G, 4G, 5G and modems
▪ Automotive – functional safety
▪ Industrial control
▪ Communications
▪ Networking
▪ SoC real-time controllers
Market-leading, real-time compute across many markets
>4.5Bnunits shipped
to date
© ARM 2017 26
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
ARM Cortex-R portfolio
Storage
and
modem
Cortex-R7
High
performance
4G modem and
storage
Cortex-R8
Highest
performance
5G modem and
storage
Functional
safety
Cortex-R4
Real-time performance
Cortex-R5
Real-time
performance
with functional
safety
Cortex-R52
Most advanced processor for
functional safety
ARMv8-RARMv7-R
Cortex-R5
Real-time
performance and
peripheral
control
© ARM 2017 27
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
Cortex-R8 – next generation mobile and storageC
ore
Mar
ks*
Scalable & efficient
Software workload can be
spread across up to 4 cores
* Total MP CoreMarks using 28nm HPM. Max multi processor config.
0
5000
10000
15000
20000
25000
30000
Cortex-R4 Cortex-R5 Cortex-R7 Cortex-R8
Software compatible
Reduce time-to-market and
protect software investment
Market-leading performance
Best-in-class hard real-time
performance and power efficiency
Spanning performance needs
5G
© ARM 2017 28
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
Cortex-R52 – ARM’s most advanced processor for safety
ARM’s highest performance, real-time
processor for safety applications
Enabling partner choice through the standardized ARM
architecture and #1 ecosystem
Simplifying functional safety. Providing
enhanced safety features and safety support
© ARM 2017 29
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
Cortex-AHighest performance
Optimized for
high-level operating
systems
Cortex-RFast response
Optimized for
high performance,
hard real-time
applications
Cortex-MSmallest/lowest power
Optimized for
discrete processing
and
microcontrollers
Optimized for
physical security
SecurCoreTamper resistant
ARM CPU architecture for total computing
© ARM 2017 30
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
ARM SecurCore for physical security
SecurCore
SC000
Optimized area,anti-tampering
SC300
Performance,
anti-tampering
• 32-bit embedded, high performance
CPU with anti-tampering
• De facto standard for secure elements
• De facto standard for SIM and identification
• Small 32-bit embedded secure CPU for
constrained applications
ARMv6-M
Proven SolutionAnti-tampering PerformanceUltra Low Power
ARMv7-M
The easiest and most proven path to meet for physical security
© ARM 2017 31
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
ARM SecurCore portfolio
High
performance
Performance
efficiency
Lowest
power & area
Cortex-M0
Lowest cost,
low power
Cortex-M0+
Highest energy
efficiency
Cortex-M4
Mainstream
control and
DSP
Cortex-M3
Performance
efficiency
Cortex-M7Maximum
performance,
control and
DSP
ARMv6-M
ARMv7-M
Anti-tampering
SC000
Optimized area,anti-tampering
SC300
Performance,
anti-tampering
ARMv6-M
ARMv7-M
2.5BnSecurCore
shipments in
2016
© ARM 2017 32
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
Summary
ARM provides the world’s most power-efficient processors
ARM’s diverse portfolio has solutions for a wide range of applications
The ARM ecosystem is innovating for Total Computing
© ARM 2017 33
Title 40pt sentence case
Bullets 24pt sentence case
bullets 20pt sentence case
Performance and scalability for a diverse range of applications
ARMv5 ARMv6
ARM11MPCore
ARM1176JZ(F)-S
ARM1136J(F)-S
Cort
ex
-A
Cort
ex
-R
Cort
ex
-M
ARMv7-A ARMv8-A
ARMv7-R ARMv8-R
ARMv7-M ARMv8-M
Previous ARMv7 ARMv8
ARM968E-S
ARM946E-S
ARM926EJ-S
Cortex-R8
Cortex-R7
Cortex-R5
Cortex-R4
ARM7TDMI
ARM920T
Cortex-A72
Cortex-A73
Cortex-R52
Cortex-A17
Cortex-A15
ARMv4 ARMv6-M
ARMv6
Cortex-M0+
Cortex-M0
Cortex-A7
Cortex-A5
Cortex-A9
Cortex-A8Cortex-A53
Cortex-A35
Cortex-A32
Cortex-M23
High
performance
High
efficiency
Ultra high
efficiency
Real Time
Performance
efficiency
Lowest power
and area
Cortex-A57
ARM1156T2(F)-S
Cortex-M33
Cortex-M7
Cortex-M4
Cortex-M3
High
performance
Cortex-A75
Cortex-A55