An Overview of Digital Signal Processor
admin
October 05, 2023
Ⅰ Introduction
A digital signal processor (DSP) is a processor composed of large-scale integrated circuit chips to complete digital signal processing tasks. Digital signal processing is the theory and technology of representing and digitally processing signals. Digital signal processing and analog signal processing are a subset of signal processing. The purpose of digital signal processing is to measure or filter continuous analog signals in the real world. Therefore, before performing digital signal processing, the signal needs to be converted from the analog domain to the digital domain, which is usually realized by an analog-to-digital converter. The output of digital signal processing often has to be converted to the analog domain, which is achieved by a digital-to-analog converter. Digital signal processors are gradually developed to meet the needs of high-speed real-time signal processing tasks. With the development of integrated circuit technology and digital signal processing algorithms, the implementation methods of DSPs are constantly changing, and the processing functions are continuously improved and expanded.
Ⅱ Features and advantages of DSP
1. Internal structure
DSP contains the following important components:
Program memory: stores the program that the DSP will use to process data;
Data storage: store the information to be processed;
Calculation engine: Perform mathematical processing, access the program in the program memory and the data in the data memory;
Input/Output: Provide a series of functions, connect with the outside.
DSP internal structure
2. Hardware features
(1) DSP belongs to Modified Harvard architecture, that is, it has two internal buses: data bus and program bus. The program and the data storage space are separated, and each has an independent address bus and data bus, and fetching and reading can be performed simultaneously. At present, it has reached 9 billion floating-point operations per second (9000 MFLOPS).
(2) Pipeline operations. The execution of each instruction is divided into several steps such as instruction fetching, decoding, fetching, and execution, which are completed by multiple functional units on-chip. Pipeline operation is equivalent to multiple instructions executed in parallel, which greatly improves the speed of operation.
(3) Independent hardware multiplier. Multiplication instructions are completed in a single cycle, optimizing a large number of repeated multiplications in algorithms such as convolution, digital filtering, FFT, correlation, and matrix operations.
(4) Circular addressing, bit-reversed, and other special instructions greatly improve the speed of addressing, sorting, and calculation in operations such as FFT and convolution. The time of 1024 points FFT is less than 1μs.
(5) Independent DMA bus and controller. There are one or more independent DMA buses, which work in parallel with the CPU program and data bus. The DMA speed has reached more than 800Mbyte/s without affecting the CPU.
(6) Multiprocessor interface. Multiple processors can easily work in parallel or serial to improve processing speed.
(7) JTAG (Joint Test Action Group) standard test interface (IEEE 1149 standard interface). It is convenient for online simulation of DSP and debugging under multi-DSP conditions.
3. Software features
(1) Immediate addressing: An operand is an immediate number, which can be obtained directly from the instruction. Example: MOV A, @0x16
(2) Direct addressing: For example, TI's TMS320 series chip divides the data memory into 512 pages with 128 words per page. Set a data page pointer DP, use 9-bit to point to a data page, and add a 7-bit page offset address to form a 16-bit data address. This helps speed up the addressing.
(3) Indirect addressing: ① 8 auxiliary registers. An auxiliary register arithmetic unit is designated by an auxiliary register pointer for a 16-bit unsigned number operation. A new address is determined, and one of the auxiliary registers is loaded. ② The contents of the eight auxiliary registers are quite flexible, and can be loaded, added, and subtracted from the immediate data; the address can be loaded from the data memory; some indexed addressing can also be done. ③Because of the reverse bit position, the bit reverse addressing can be realized.
(4) Unique multiplication instruction: Example: MAC X0, Y0, AX: (R0) +, X0 Y: (R4) + N4. YO instruction command DSP56300: multiply the numbers in registers X0 and Y0, the result is added to Acc A, load the value in the memory address pointed to by register R0 into register X0, load the value in Y memory address pointed to by register R4 into register Y0, add the value of R0 by 1, and add the value of register N4 to R4.
4. Advantages of the DSP
Digital signal processor(DSP)
(1) The interface is simple and convenient. Because of the simple electrical characteristics of digital signals, it is easy to implement on the hardware interface when different DSP systems are connected. On the data stream interface, as long as each system follows a specific standard protocol.
(2) High precision and good stability. Digital signal processing is only affected by the binarization error and the limited word length. The process does not introduce other noise, so it has a high signal-to-noise ratio. Besides, the performance of the analog system is greatly affected by the performance of the component parameters, while the digital system is basically unchanged, so the digital system is more convenient for testing, debugging, and mass production.
(3) Easy programming and easy to implement complex algorithms. In the DSP system, the DSP chip provides a high-speed computing platform, and the system function depends on the software programming. When combined with modern signal processing theory and computational mathematics, complex digital signal processing functions can be achieved.
(4) Easy integration. Modern DSP chips integrate DSP cores and peripheral circuits on a single chip. This structure is convenient for designing portable and highly integrated digital products.
Also, modern DSP chips are programmable Very-large-scale integration (VLSI) devices that implement digital signal processing functions through downloadable software and firmware. In addition to the operation and control functions of the ordinary microprocessor, the DSP chip has also made great improvements in the processor structure, instruction system, and instruction flow design for real-time digital signal processing with high data transmission rates and intensive numerical operations.
Ⅲ The history of DSP
Digital signal processor(DSP)
The foundation of informatization is digitization. One of the core technologies of digitization is digital signal processing. The task of digital signal processing largely needs to be completed by the DSP device. DSP technology has become a cutting-edge technology that people are increasingly paying attention to and have been developing rapidly. DSP can stand for Digital Signal Processor, or digital signal processing, the latter is a theoretical technology, which should be turned into an actual product through the former. The combination of the two becomes a digital signal processing solution (DSPS), which is a means of solving a practical problem and realizing a solution. This article focuses on the first interpretation of the DSP-digital signal processor. DSP is a special processor for high-speed real-time processing after the analog signal is converted into a digital signal, and its processing speed is 10 to 50 times faster than the fastest CPU. In the context of today's digital era, DSP has become a basic device in the fields of communications, computers, consumer electronics, and other fields. Industry insiders predict that DSP will be the fastest-growing electronic product in integrated circuits in the future, and will become the decisive factor in the replacement of electronic products.
Before DSP appeared, digital signal processing could only be completed by MPU (microprocessor). But the lower processing speed of MPU can not meet the high-speed real-time requirements. Therefore, in the 1970s, some people proposed the theory and algorithm of DSP. The DSP only stays in the textbooks. Even the DSP system developed is composed of discrete components, and its application field is limited to the military and aviation departments.
With the development of large-scale integrated circuit technology, the first DSP chip was born in the world in 1982. This kind of DSP device is made of micron technology NMOS. Although the power consumption and size are slightly larger, the operation speed is dozens of times faster than MPU. It has been widely used especially in speech synthesis and codec. The advent of DSP chips marks a big step for DSP application systems from large systems to miniaturization. With the progress and development of CMOS technology, the second generation of DSP chips based on CMOS technology came into being. Its storage capacity and operation speed have doubled, and it has become the basis of voice processing and image hardware processing technology. In the late 1980s, the third-generation DSP chip came out, and the operation speed was further improved, and its application range was gradually expanded to the fields of communications and computers.
In the 1990s, DSP developed the fastest, and the fourth and fifth-generation DSP devices appeared one after another. The current DSP belongs to the fifth generation product. Compared with the fourth generation, system integration is higher, and the DSP core and peripheral components are integrated on a single chip. This highly integrated DSP chip not only shows its talents in the fields of communication and computers but also gradually penetrates the daily consumption field of people.
Ⅳ Problems and challenges facing DSP
The increasingly mature DSP still has many areas that need improvement, and also faces many challenges.
(1) How to arrange the data flow reasonably so that it can be executed smoothly without conflicts between DSP execution units is still an important issue facing DSP developers. Due to the complexity of the design, when the algorithm is mapped to the specific target hardware of the DSP, a high-level programming language cannot be used. Assembly language must be used, and the parallel execution mechanism of the device has a very clear understanding. And this kind of programming design limited to assembly language is the bottleneck of improving software development efficiency.
(2) There are still problems with the parallel structure. To achieve higher throughput, it is necessary to process more data bits in a specific unit of time. VLIW technology represents parallelism at the instruction level. Superscalar structures and super pipeline structures also try to get more instructions in one instruction cycle. Data-level parallelism is represented by wider data words, vectorization, and data stream structure. Because the width of the data word is larger, the instruction can process more data per instruction cycle, increasing the number of data bits that can be processed per clock cycle. Task-level or transaction-level parallelism is reflected in multitasking, multithreading, and multiprocessor designs. These structures are expected to improve data processing throughput, but the increased data and instruction width and the consequent increase in data processing throughput will have to pay a price. When the code density and data width match the application, they work. However, when the data word width is different from the processor, it can cause a lot of trouble.
(3) A large amount of available on-chip cache is becoming more and more important to the overall throughput of the system because the standard memory bus and interface can no longer support the gigabyte data transfer rate of each MAC in the system. Whether the rest of the system can be matched with high-speed processors is also becoming a big problem. The dual MAC processor with 2 ALU units may require 4 data words per clock cycle or more than 4 gigabits of data per the second word.
(4) The challenges faced by the development of DSP are also reflected in the rapid increase in CPU speed and the continued decline in prices, making DSP manufacturers face two options, one is to accelerate the development of DSP, and the other is to withdraw from the competition. Each manufacturer must diversify from diversified investment to a single investment, and establish DSPS as the main development product, that is, integrate all technologies and all products in DSP.
Ⅴ The development trend and the prospect of DSP
DSP continues to meet the increasing demands of people on its development road and is gradually developing towards the direction of personalization and low power consumption.
(1) System-level integrated DSP is the trend
Reducing the size of DSP chips has always been the direction of DSP technology. Most current DSPs are based on the RISC (Reduced Instruction Set Computing) structure, which has the advantages of small size, low power consumption, and high performance. Various DSP manufacturers have adopted new processes to improve the DSP core and integrated several DSP cores, MPU cores, special processing units, peripheral circuit units, and storage units on one chip to become DSP system-level integrated circuits.
(2) Programmable DSP is the leading product
Programmable DSP provides manufacturers with great flexibility. Manufacturers can develop various types of products on the same DSP platform to meet the needs of different users. At the same time, the programmable DSP also provides users with a good way to upgrade easily.
(3) Fixed-point DSP is the mainstream
In theory, although the dynamic range of floating-point DSP is larger than that of fixed-point DSP and more suitable for DSP applications, the cost of DSP devices for fixed-point arithmetic is lower, the requirements for memory are also lower, and the power consumption is less. Therefore, programmable DSP devices for fixed-point arithmetic are still mainstream products on the market. According to statistics, most of the DSP devices currently sold are 16-bit fixed-point programmable DSP devices, and it is expected that the proportion will gradually increase in the future.
(4) Pursue higher calculation speed
At present, the general DSP operation speed is 100MIPS, that is, 100 million instructions per second can be calculated. Due to the personalization and customization of electronic devices, DSP must pursue higher and faster computing speeds to keep up with the pace of updating electronic devices. The improvement of DSP operation speed mainly relies on new technology to improve the chip structure. At present, most DSP devices use 0.5μm～0.35μm CMOS technology. According to the development trend of CMOS, it is entirely possible to increase the DSP's operation speed by 100 times (up to 1600GIPS).
Ⅵ Conclusion
The increasing application of DSP in various fields has driven the development of DSP itself, and its application in the field of 3C (Communications, Computers, Consumer) has occupied 90% of the current DSP market share, indicating that the potential of DSP in other fields is still huge. In future development, it will appear in various fields with more excellent performance.
