際際滷

際際滷Share a Scribd company logo
Intellectual Property (IP) Cores
By Jannin Joy A. Ramirez
Semiconductor industry: Then and Now
THEN
SC companies provided
 product definition
 design
 manufacturing
 assembly
 customer support
Goal: system integration
NOW
chipless companies
only provide
 design information
e.g. ARM 
microprocessors,
peripherals, and chips
Amphion  digital
video and broadband
wireless applications
IP Core: What is it exactly?
 A design function with well-defined interfaces.
 a design block for a specific chip that handles a
well-defined piece of functionality
 A block of logic or data that can be used in
making application-specific integrated circuits
(ASICs) and field programmable gate arrays
(FPGAs)
e.g. CPUs, Ethernet controllers, UARTs
Ethernet  LAN technology, IOBASE-T
Universal Asynchronous Receiver/Transmitter 
microchip w/ programming that controls PCs
interface to attached serial devices
IP core Design flow
 息Whitney and Neville-Neil
How are IP cores created?
 Register Transfer Language (RTL)  a
hardware programming language
 Synthesis programs  read RTL and
translate them directly to circuits that are
implemented on the silicon
 Simulation programs  read RTL and let
designers exercise functionality and check
its correctness
EDA (electronic design automation) Tools
DESIGN STAGE TOOLS SUPPORTED
RTL Simulation
And Gate Level Simulation
Synopsys Chronologic VCS
v5.0.1A  Unix
Cadence Verilog XL
v2.6.37 Unix
ModelTech
ModelSim v5.3d WinNT
Logic Synthesis Synopsys Design Compiler
v99.10-4
Test Synthesis Synopsys Test Compiler
v99.10-4
IP core categories
 SOFT IP  core is not mapped onto silicon
(+) portable across technology generations
(-) raw RTL code must be tailored to a target
technology through logic design process
 HARD IP  physical manifestations of the IP
design
(+) best for plug-and-play applications
(-) less portable and flexible
 FIRM IP  combination of best attributes of both
soft and hard
(+) portability and silicon optimization
Comparison of Attributes of IP formats
SOFT IP FIRM IP HARD IP
Delivery
format
RTL code Technology
specific netlist
GDSII data
Portability/
Reusability
HIGH HIGH LOW
Silicon
Optimization
LOW HIGH HIGHEST
Integration
effort required
HIGH
Logic
design
process
required
LOW
Place and rout
tools
LOW
Drop into
SoC design
ARM CPU cores
 ARM CPU cores cover a wide range of
performance and features enabling system
designers to create solutions that meet their precise
requirements.
 Three system categories:
Embedded real-time  storage, automotive,
industrial and networking applications
Open platforms  devices running platform OS
like Linus, PalmOS, Windows CE
Secure applications  SIM cards and payment
terminals
Cache Size
(Inst/Data)
Tightly
Coupled
Memory
Memory
management
AHB Bus
Interface
Clock
MHz
EMBEDDED CORES
ARM7TDMI No No No Yes 133
ARM7TDMI-S No No No Yes 100-133
ARM1026EJ-S Variable Yes MMU+MPU Dual AHB 266-325
PLATFORM CORES
ARM720T 8K unified No MMU Yes 100
ARM1136J(F)-S Variable Yes MMU Quad 64-
bit AHB
333-400
Latest version of ARM CPU cores
 ARM v6 Architecture KEY AREAS:
Memory Management
 average fetch and data latency is reduced
 more efficient bus usage; less bus activity
yields significant power savings as a result of
reduced memory access
 Multiprocessing
Multiprocessor systems share data
efficiently by sharing memory
ARM v6 data sharing and synchronization
capabilities make it easier to implement
multiprocessor systems
 Multimedia Support
 Single Instruction Multiple Data (SIMD)
capabilities enable more efficient software
implementation of media applications such
as audio and video encoders
 Data Handling
Endianism refers to the way data is
referenced and stored in memory
SoC integration: little endian OS
environment and interfaces (USB,PCI) with
big endian data (TCP/IP packet, MPEG
stream)
ARM v6 handles data in mixed-endian
systems
Latest version of ARM CPU core
Benefits of ARM v6 Architecture
30 percent increase in system performance
* level-one memory system, which includes features
such as a tightly-coupled Direct Memory Access
(DMA) controller and re-architected cache,
Up to 8x performance increase for media
applications
* SIMD capabilities boost the performance of media
applications such as audio and video
encoder/decoders by up to four times,
* also includes enhanced instruction set support for
motion estimation.
ARM11 Family: ARM1136J-S
and ARM 1136J(F)-S
 Low power consumption
o < 0.4mW/MHz (0.13袖m, 1V) including cache controllers
o Energy saving power-down modes address static leakage currents in
advanced processes
 High performance integer processor
o 8-stage integer pipeline delivers high clock frequency
o Separate load-store and arithmetic pipelines
Branch Prediction and Return Stack
 High performance memory system design
o Supports 4-64k cache sizes
o Optional tightly coupled memories with DMA for multi-media
applications
o High-performance 64-bit memory system speeds data access for media
processing and networking applications
o ARMv6 memory system architecture accelerates OS context-switch

More Related Content

A block of logic or data that can be used in making application-specific integrated circuits (ASICs) and field programmable gate arrays (FPGAs)

  • 1. Intellectual Property (IP) Cores By Jannin Joy A. Ramirez
  • 2. Semiconductor industry: Then and Now THEN SC companies provided product definition design manufacturing assembly customer support Goal: system integration NOW chipless companies only provide design information e.g. ARM microprocessors, peripherals, and chips Amphion digital video and broadband wireless applications
  • 3. IP Core: What is it exactly? A design function with well-defined interfaces. a design block for a specific chip that handles a well-defined piece of functionality A block of logic or data that can be used in making application-specific integrated circuits (ASICs) and field programmable gate arrays (FPGAs) e.g. CPUs, Ethernet controllers, UARTs Ethernet LAN technology, IOBASE-T Universal Asynchronous Receiver/Transmitter microchip w/ programming that controls PCs interface to attached serial devices
  • 4. IP core Design flow 息Whitney and Neville-Neil
  • 5. How are IP cores created? Register Transfer Language (RTL) a hardware programming language Synthesis programs read RTL and translate them directly to circuits that are implemented on the silicon Simulation programs read RTL and let designers exercise functionality and check its correctness
  • 6. EDA (electronic design automation) Tools DESIGN STAGE TOOLS SUPPORTED RTL Simulation And Gate Level Simulation Synopsys Chronologic VCS v5.0.1A Unix Cadence Verilog XL v2.6.37 Unix ModelTech ModelSim v5.3d WinNT Logic Synthesis Synopsys Design Compiler v99.10-4 Test Synthesis Synopsys Test Compiler v99.10-4
  • 7. IP core categories SOFT IP core is not mapped onto silicon (+) portable across technology generations (-) raw RTL code must be tailored to a target technology through logic design process HARD IP physical manifestations of the IP design (+) best for plug-and-play applications (-) less portable and flexible FIRM IP combination of best attributes of both soft and hard (+) portability and silicon optimization
  • 8. Comparison of Attributes of IP formats SOFT IP FIRM IP HARD IP Delivery format RTL code Technology specific netlist GDSII data Portability/ Reusability HIGH HIGH LOW Silicon Optimization LOW HIGH HIGHEST Integration effort required HIGH Logic design process required LOW Place and rout tools LOW Drop into SoC design
  • 9. ARM CPU cores ARM CPU cores cover a wide range of performance and features enabling system designers to create solutions that meet their precise requirements. Three system categories: Embedded real-time storage, automotive, industrial and networking applications Open platforms devices running platform OS like Linus, PalmOS, Windows CE Secure applications SIM cards and payment terminals
  • 10. Cache Size (Inst/Data) Tightly Coupled Memory Memory management AHB Bus Interface Clock MHz EMBEDDED CORES ARM7TDMI No No No Yes 133 ARM7TDMI-S No No No Yes 100-133 ARM1026EJ-S Variable Yes MMU+MPU Dual AHB 266-325 PLATFORM CORES ARM720T 8K unified No MMU Yes 100 ARM1136J(F)-S Variable Yes MMU Quad 64- bit AHB 333-400
  • 11. Latest version of ARM CPU cores ARM v6 Architecture KEY AREAS: Memory Management average fetch and data latency is reduced more efficient bus usage; less bus activity yields significant power savings as a result of reduced memory access
  • 12. Multiprocessing Multiprocessor systems share data efficiently by sharing memory ARM v6 data sharing and synchronization capabilities make it easier to implement multiprocessor systems
  • 13. Multimedia Support Single Instruction Multiple Data (SIMD) capabilities enable more efficient software implementation of media applications such as audio and video encoders
  • 14. Data Handling Endianism refers to the way data is referenced and stored in memory SoC integration: little endian OS environment and interfaces (USB,PCI) with big endian data (TCP/IP packet, MPEG stream) ARM v6 handles data in mixed-endian systems
  • 15. Latest version of ARM CPU core Benefits of ARM v6 Architecture 30 percent increase in system performance * level-one memory system, which includes features such as a tightly-coupled Direct Memory Access (DMA) controller and re-architected cache, Up to 8x performance increase for media applications * SIMD capabilities boost the performance of media applications such as audio and video encoder/decoders by up to four times, * also includes enhanced instruction set support for motion estimation.
  • 16. ARM11 Family: ARM1136J-S and ARM 1136J(F)-S Low power consumption o < 0.4mW/MHz (0.13袖m, 1V) including cache controllers o Energy saving power-down modes address static leakage currents in advanced processes High performance integer processor o 8-stage integer pipeline delivers high clock frequency o Separate load-store and arithmetic pipelines Branch Prediction and Return Stack High performance memory system design o Supports 4-64k cache sizes o Optional tightly coupled memories with DMA for multi-media applications o High-performance 64-bit memory system speeds data access for media processing and networking applications o ARMv6 memory system architecture accelerates OS context-switch