18-04-2011, 10:53 AM
[attachment=12277]
INTRODUCTION
Computers have evolved from few, huge mainframes shared by many people, and mini computers that were smaller but still shared to today’s PCs—millions in number, miniscule in size compared to the mainframes, and used by only one person at a time. The next generation could be invisible, with billions being around and each of us using more than one at a time. Welcome to the world of embedded systems, of computers that will not look like computers and won’t function like anything we’re familiar with
WHAT IS AN EMBEDDED SYSTEM?
As the name signifies, an embedded system is ‘embedded’ or built into something else, which is a non-computing device, say a car, TV, or toy. Unlike a PC, an embedded computer in a non-computing device will have a very specific function, say control a car, or display Web pages on a TV screen. So, it need not have all the functionality and hence all the components that a PC has. Similarly, the operating system and applications need not perform all the tasks that their counterparts from the PC sphere are expected to.
In short, we can define an embedded system as a computing device, built into a device that is not a computer, and meant for doing specific computing tasks. These computing tasks could range from acquiring or transferring data about the work done by the mother device to displaying information or controlling the mother device. Embedded systems could thus enable us to build intelligent machines.
Embedded systems is not a new and exotic topic that is still confined to research theses. There are many live examples of embedded systems around us. MP3 players (computing capability built into a music system), PDAs (computing in what essentially is an organizer), car-control systems, and intelligent toys are but a few examples of such systems already in place.
A typical embedded system consists of hardware (typically VLSI or very large-scale integrated circuits) specifically built for the purpose, an embedded operating system, and the specific application or applicatiospecification is considered to be an extension of the ISA bus specification. The PC /104 standard has since been extended to PC/104-plus to include the PCI bus. So, today you have PC-based embedded systems that have the ISA bus, the PCI bus, or both.
Unlike with regular PCs, in the world of the embedded PC, 386s, 486s and Pentiums are still good enough. Besides these, there are a number of CPUs meant specifically for embedded applications, like the StrongArm and the MIPS.
With embedded PCs you can even go beyond the single-function definition of an embedded system, and could build an entire PC into another machine; a PC inside a refrigerator, or a PC inside a car, for instance.
HARDWARE FOR EMBEDDED DEVICES
Universal Micro system is a general-purpose hardware that can be programmed and used to develop applications for different embedded devices
Many modern appliances like MP3 players, ‘intelligent’ refrigerators, and watches use embedded systems. However, a common obstacle for developers has been the need to develop different sets of hardware and software, for different devices. An ‘intelligent’ washing machine uses a hardware chip different from that used by an ‘intelligent’ wristwatch. In addition, the software running on the hardware chip is different. This often results in increased costs and time taken for development. (For more on embedded systems see PCQuest May2001,page38.)
The Universal Micro System (UMS) from Cradle Technologies is a solution for this problem. UMS is a general-purpose chip built around a simple instruction set. It can be used to develop applications for embedded devices because all the functionality required for a specific device can be modeled in the software.
UMS HARDWARE
Any software application expects four basic requirements from the underlying hardware: input unit, processing unit, memory unit, and output unit. Since the major functionality provided in UMS is through software, the processor and memory units must be very fast and the input-output units must be programmable and versatile.
UMS uses a large number of high speed, low power and small RISC-based processors (about 75) on a single chip. Each processor also called a PE (processing Element) coupled with two Digital Signal Processors called DSE (Digital Signal Engines) form an MSP (Multi Stream Processor), which processes voluminous chunks (stream) of data.
The UMS is structured into a number of Quads. A Quad, as shown in the diagram to the left, consists of four MSPs, program or instruction cache, data cache and a programmable DMA (Direct Memory Access) unit. There is also a high throughput (about 4 GB/sec) global bus interface, which interconnects all of them in a Quad. The use of DSEs ensures smooth digital processing, while the powerful PEs carry out the arithmetic and logical functions on the processed data. Finally, the result of the processing is transferred from the local data cache of a Quad to an external SDRAM (Synchronous Dynamic Random Access Memory) module via the DMA unit. The UMS chip does have an onboard DRAM controller to interface with external SDRAM modules. Feeding each Quad with independent chunks of data can make optimal use of the raw processing speed of the UMS chip. It is claimed that UMS has a raw speed of over 15 GFLOPS (Giga Floating Point Operations per second) while consuming just 1.5 watts of power.
The Input/Output unit of UMS is programmable. You can program it to support processing unit dependant data transfers, or do a DMA data transfer where data transfers can take place without the intervention of the processing unit. In fact, the programmable I/O is claimed to be so versatile that it can be used to model PCI, SCSI, FireWire, or DSL interfaces using software. In other words, the I/O hardware is extensively programmable through software.
SOFTWARE ON UMS
The software design has eliminated the need for customized hardware. It has been left to the developer to utilize the power of the numerous processors by using efficient software algorithms. Optimally, each Quad must be fed with independent data blocks (called data parallelism). This is the responsibility of the software developer. What Cradle has provided are some tools to speed up this development: a C compiler, an assembler and a cross assembler, linker, debuggers, and most important, a software simulator of the hardware chip. A custom C-API (Application Programming Interface), comprising of UMS specific library functions, is also provided. These include libraries for TCP/IP, OpenGL 3D, PCI, FireWire, MPEG and DV encoding and decoding.
INTRODUCTION
Computers have evolved from few, huge mainframes shared by many people, and mini computers that were smaller but still shared to today’s PCs—millions in number, miniscule in size compared to the mainframes, and used by only one person at a time. The next generation could be invisible, with billions being around and each of us using more than one at a time. Welcome to the world of embedded systems, of computers that will not look like computers and won’t function like anything we’re familiar with
WHAT IS AN EMBEDDED SYSTEM?
As the name signifies, an embedded system is ‘embedded’ or built into something else, which is a non-computing device, say a car, TV, or toy. Unlike a PC, an embedded computer in a non-computing device will have a very specific function, say control a car, or display Web pages on a TV screen. So, it need not have all the functionality and hence all the components that a PC has. Similarly, the operating system and applications need not perform all the tasks that their counterparts from the PC sphere are expected to.
In short, we can define an embedded system as a computing device, built into a device that is not a computer, and meant for doing specific computing tasks. These computing tasks could range from acquiring or transferring data about the work done by the mother device to displaying information or controlling the mother device. Embedded systems could thus enable us to build intelligent machines.
Embedded systems is not a new and exotic topic that is still confined to research theses. There are many live examples of embedded systems around us. MP3 players (computing capability built into a music system), PDAs (computing in what essentially is an organizer), car-control systems, and intelligent toys are but a few examples of such systems already in place.
A typical embedded system consists of hardware (typically VLSI or very large-scale integrated circuits) specifically built for the purpose, an embedded operating system, and the specific application or applicatiospecification is considered to be an extension of the ISA bus specification. The PC /104 standard has since been extended to PC/104-plus to include the PCI bus. So, today you have PC-based embedded systems that have the ISA bus, the PCI bus, or both.
Unlike with regular PCs, in the world of the embedded PC, 386s, 486s and Pentiums are still good enough. Besides these, there are a number of CPUs meant specifically for embedded applications, like the StrongArm and the MIPS.
With embedded PCs you can even go beyond the single-function definition of an embedded system, and could build an entire PC into another machine; a PC inside a refrigerator, or a PC inside a car, for instance.
HARDWARE FOR EMBEDDED DEVICES
Universal Micro system is a general-purpose hardware that can be programmed and used to develop applications for different embedded devices
Many modern appliances like MP3 players, ‘intelligent’ refrigerators, and watches use embedded systems. However, a common obstacle for developers has been the need to develop different sets of hardware and software, for different devices. An ‘intelligent’ washing machine uses a hardware chip different from that used by an ‘intelligent’ wristwatch. In addition, the software running on the hardware chip is different. This often results in increased costs and time taken for development. (For more on embedded systems see PCQuest May2001,page38.)
The Universal Micro System (UMS) from Cradle Technologies is a solution for this problem. UMS is a general-purpose chip built around a simple instruction set. It can be used to develop applications for embedded devices because all the functionality required for a specific device can be modeled in the software.
UMS HARDWARE
Any software application expects four basic requirements from the underlying hardware: input unit, processing unit, memory unit, and output unit. Since the major functionality provided in UMS is through software, the processor and memory units must be very fast and the input-output units must be programmable and versatile.
UMS uses a large number of high speed, low power and small RISC-based processors (about 75) on a single chip. Each processor also called a PE (processing Element) coupled with two Digital Signal Processors called DSE (Digital Signal Engines) form an MSP (Multi Stream Processor), which processes voluminous chunks (stream) of data.
The UMS is structured into a number of Quads. A Quad, as shown in the diagram to the left, consists of four MSPs, program or instruction cache, data cache and a programmable DMA (Direct Memory Access) unit. There is also a high throughput (about 4 GB/sec) global bus interface, which interconnects all of them in a Quad. The use of DSEs ensures smooth digital processing, while the powerful PEs carry out the arithmetic and logical functions on the processed data. Finally, the result of the processing is transferred from the local data cache of a Quad to an external SDRAM (Synchronous Dynamic Random Access Memory) module via the DMA unit. The UMS chip does have an onboard DRAM controller to interface with external SDRAM modules. Feeding each Quad with independent chunks of data can make optimal use of the raw processing speed of the UMS chip. It is claimed that UMS has a raw speed of over 15 GFLOPS (Giga Floating Point Operations per second) while consuming just 1.5 watts of power.
The Input/Output unit of UMS is programmable. You can program it to support processing unit dependant data transfers, or do a DMA data transfer where data transfers can take place without the intervention of the processing unit. In fact, the programmable I/O is claimed to be so versatile that it can be used to model PCI, SCSI, FireWire, or DSL interfaces using software. In other words, the I/O hardware is extensively programmable through software.
SOFTWARE ON UMS
The software design has eliminated the need for customized hardware. It has been left to the developer to utilize the power of the numerous processors by using efficient software algorithms. Optimally, each Quad must be fed with independent data blocks (called data parallelism). This is the responsibility of the software developer. What Cradle has provided are some tools to speed up this development: a C compiler, an assembler and a cross assembler, linker, debuggers, and most important, a software simulator of the hardware chip. A custom C-API (Application Programming Interface), comprising of UMS specific library functions, is also provided. These include libraries for TCP/IP, OpenGL 3D, PCI, FireWire, MPEG and DV encoding and decoding.
