UNIT 2
Architecture of Raspberry Pi
Single Board Computers
Single-Board Computer (SBC)
A Single-Board Computer (SBC) is a complete, functioning computer in which the microprocessor,
input/output functions, memory, and other features are all built on a singe circuit board, with RAM built
in at a pre-determined amount and with no expansion slots for peripherals.
Unlike a desktop personal computer, single board computers often do not rely on expansion slots for
peripheral functions or expansion. Single board computers have been built using a wide range
of microprocessors. Simple designs, such as those built by computer hobbyists, often use static RAM and
low-cost 8- or 16-bit processors. Other types, such as blade servers, would perform similar to
a server computer, only in a more compact format.
SBC Architecture Memory
Rock Pi X Model B x86, 64-bit 2/4 GB LPDDR3
Best Single-
Board
Computers Asus Tinker Board 2S ARM, 64-bit 2 GB LPDDR4
2022
LattePanda V1 x86, 64-bit 2/4 GB DDR3L
LattePanda Delta 432 x86, 64-bit 4 GB LPDDR4
SOC Definition
A System on Chip or an SoC is an integrated circuit that incorporates a
majority of components present on a computer. As the name suggests,
it is an entire system fabricated on a silicon chip. The beauty of an
SoC is that it integrates all the components on a single substrate. In
semiconductors, a substrate is a thin film of silicon used to fabricate
integrated circuits. In contrast to the traditional motherboard, SoC
integrates the replaceable components onto a single chip, thereby
reducing the size and increasing efficiency.
Block Diagram of System on Chip (SoC)
Processor
At the heart of the SoC is its Processor. It usually has multiple processor cores. Multiple cores
allow for different processes to run at the same time, which increases the speed of the system as
it enables your computer to perform multiple operations at the same time. Basically, the
operating system sees the multiple cores as multiple CPUs, which increases performance. As
multiple cores are fitted onto the same chip, there is less latency, which is because of faster
communication between the cores. A multiple core system has only one CPU socket with multiple
cores. For example, the following system has four cores and one socket.
Digital Signal Processor (DSP)
Digital Signal Processor (DSP) is a chip optimized for operations for digital signal processing. This
includes operations for sensors, actuators, data processing, and data analysis. It can be used for
image decoding. The use of DSP saves CPU cycles for other processing tasks, which increases
performance. Dedicated DSPs are more power-efficient, which makes them befitting for use in SoCs.
The instruction set used for DSP cores is SIMD (Single Instruction, Multiple Data) and VLIW (Very Long
Instruction Word). The use of this architecture allows for parallel processing of instructions and
superscalar execution. DSPs are used to perform operations like Fast Fourier Transform, convolution,
multiply-accumulate.
Memories on an SoC
SoCs have memories based on the application. The memories are semiconductor memory blocks
for computation purposes. Semiconductor memory usually refers to Metal Oxide Semiconductor
memory cells, which are fabricated on a single silicon chip.
On-chip communication
Traditionally, bus architecture was used to communicate between the SoC’s execution units. However,
these days, Network-On-Chip interconnect technology has emerged as a trend to overtake the bus
architecture. A popular example of bus communication is the AMBA (Advanced Microcontroller Bus
Architecture) bus protocol by ARM. The bus architecture is used to drive data between components. On-
chip bus architecture can be classified as a shared bus, hierarchical bus, and ring topologies. Different
companies have designed different architectures as per the chip design and application.
Advanced System Bus (ASB)
Advanced High-Performance Bus (AHB)
Advanced Peripheral Bus (APB)
ARM
External interfaces
SOC interfaces defer as per the intended application. The external interfaces are commonly based
on communication protocols such as WiFi, USB, Ethernet, SPI, HDMI. If required, analog interfaces
may be added for interfacing with sensors and actuators.
Other components
Other components necessary for a fully functioning SOC are timing sources like clocks, timers,
oscillators, phase lock loop systems, voltage regulators, and power management units.
Advantages & disadvantages of SoC
The main aim of an SoC is to minimize external components. Hence, it has the following advantages over a Single
Board Computer:
Size: The SoC is the size of a coin. Due to the rapidly decreasing size of MOS technology, SOCs can be made very
small while being able to perform complex tasks. The size does not impact the features of the chip.
Decreased power consumption: An SoC is optimized for low-power devices like cell-phones. Low power
consumption results in higher battery capacity in cell-phones.
Flexibility: SoCs are easily reprogrammable, which makes them flexible. They so allow the reuse of IPs.
Reliability: SoCs offer high circuit security and reduced design complexity.
Cost Efficient: Mainly due to fewer physical components and design reuse Faster circuit operation
SoCs pose some disadvantages as well:
Time Consuming: The entire process from design to fabrication can take between 6 months to 1 year. Hence, the
time to market demand is very high.
Design Verification requirements are very high and consume 70% of the total time. DV is tedious due to the
increasing complexity of SoC design.
Availability and compatibility of IPs play a very significant role, which can add to the time to market.
Exponentially increasing fabrication costs.
For low volume products, SoC may not be the best option.
Applications of SoCs
The most common application of SOCs today is in mobile applications, including
smartphones, smart watches, tablets. Other applications include signal speech
processing, PC interfaces, data communication. SoCs are being applied to
personal computers as well due to the integration of communication modules like
LTE and wireless networks onto the chip.
The most popular SoCs in the market today are manufactured by Qualcomm
Technologies for smartphones, smart watches, and the upcoming 5G network
compatibility. Other manufacturers include Intel Technology, Samsung Inc, Apple
Inc., among many others.
Architecture of Raspberry Pi 3 B+
The architecture of the Raspberry Pi 3 B+ is based heterogeneous platform
The basic principle of a heterogeneous platform is that they contain multiple processing elements
with different kinds of computational units. We can describe the architecture of an heterogeneous
platform with three components: Processing element (PE), Medium and Link. A PE models the
processing element type, it is used to describe the main properties of a computational unit available
on a given platform. A PE may refer to any element in the IoT architecture that is capable of being
part of the dataflow network, i.e. can produce or consume data items (tokens in dataflow parlance).
A PE therefore can be a sensor that can produce tokens, an embedded computer or IoT application
gateway that can run part of the dataflow network or a server instance. A Medium models a
communication channel or a memory, it is used to describe the kind of communication that a PE
communicates with another PE or an external communication channel. A Link describes the
properties of the interface between a PE and a Medium. A PE might be represented as nested PE in
order to represent a multi-core or many-cores computing device. As an example the architecture of
the Raspberry Pi B+ is depicted in Figure below. A Raspberry Pi B+ contains a Broadcom A-53 ARM
processor (BCM2837B0) with four processing units and variety of media that can be interconnected
with it. In addition, the architecture on Figure 4 represents the device as a standalone platform
without any other platform connected to it. Furthermore, this architecture is sufficient for describing a
Raspberry Pi that runs an operating system that controls all the interfaces and it provides four
processing cores. A dataflow application can map actors to each ARM core.
Example of heterogeneous architecture
Figure shows An heterogeneous architecture
thats contains a server and an embedded
device that receives the temperature data.
Figure 5 depicts an heterogeneous platform
composed of a server that contains a Xeon
processor with six PEs, an embedded
Raspberry Pi that communicates with the
server via WiFi and a RISC IOT device that
captures the ambient temperature and sends
its data to the Raspberry Pi through
Bluetooth. The media present in Figure but
not used by this dataflow application is
omitted in Figure.
Using this architecture representation the
dataflow application will map its actors to
each processing element and the actor FIFOs
are mapped to links which are connected to
the defined media.
BCM2837
This is the Broadcom chip used in the Raspberry Pi 3, and in later models of the
Raspberry Pi 2. The underlying architecture of the BCM2837 is identical to the
BCM2836. The only significant difference is the replacement of the ARMv7 quad
core cluster with a quad-core ARM Cortex A53 (ARMv8) cluster.
The ARM cores run at 1.2GHz, making the device about 50% faster than the
Raspberry Pi 2. The VideoCore IV runs at 400MHz.
Please refer to the following BCM2836 document for details on the ARM
peripherals specification, which also applies to the BCM2837.
Architectural features
TCM, Tightly-Coupled Memory
A graphics processing unit (GPU)
A graphics processing unit (GPU) is a specialized electronic
circuit designed to rapidly manipulate and alter memory to accelerate the
creation of images in a frame buffer intended for output to a display
device. GPUs are used in embedded systems, mobile phones, personal
computers, workstations, and game consoles. Modern GPUs are very
efficient at manipulating computer graphics and image processing. Their
highly parallel structure makes them more efficient than general-
purpose central processing units (CPUs) for algorithms that process large
blocks of data in parallel. In a personal computer, a GPU can be present on
a video card or embedded on the motherboard. In certain CPUs, they are
embedded on the CPU die
The Videocore IV quad processor (QPU) is a SIMD machine with the following key
features:
• A highly uniform 64-bit instruction encoding
• Four-way physical parallelism
• 16-way virtual parallelism (4-way multiplexed over four successive clock cycles)
• A dual-issue floating-point ALU (one add and multiply per cycle)
• Two large single-ported register files
• Five accumulators
• I/O mapped into the register space
• Support for two hardware threads
• Instruction and register level coupling to the 3D hardware
Thank You
