Things you have to know about FPGA

中信华

Things you have to know about FPGA [Copy link]

FPGA (Field Programmable Gate Array) is translated into Chinese as: Field Programmable Gate Array, which is an integrated circuit in which designers can program customizable digital logic on site.

1

What is an FPGA?

First of all, if you have never been exposed to FPGA (Field Programmable Gate Array), or have a little knowledge and want to continue to learn more about this industry, in this section, I will introduce you to FPGA and explain to you what problems FPGA can solve, how to solve these problems, and discuss how to optimize the design, etc.

In the four letters of FPGA, F stands for Field, which means the scene. So what is the scene? The scene in FPGA is a kind of scene in a broad sense. It not only refers to the place where the product is located, but also reflects the "immediacy" of FPGA. P stands for Programmable, which means programmable. G stands for Gate, which is the logic gate in the digital circuit. A stands for Array, which refers to the array composed of logic gates. Combining these four English words, we often call it Field Programmable Gate Array - FPGA.

Compared with other ways of building hardware platforms, FPGA does two things. First, it allows you to freely build the hardware platform you need, replacing the long development time, high development costs and design risks that you had to bear when using ASIC (application-specific integrated circuit) in the past. Another point is that compared with other hardware structures, it allows you to build the hardware you need without ASIC, and it is simpler, faster and more power-efficient than using a microprocessor core.

2

Why do we need FPGA?

①High programming freedom

FPGA seems to be omnipotent in the digital world. It can be used to build various functional modules like Lego blocks to realize various functions you want. Of course, first you must master the most basic digital logic knowledge and learn a tool language used to build various functions (here we recommend the popular Verilog). Secondly, you have to use your brain (it tests whether your logical thinking is clear). The work of an excellent architect is sketched out in his mind, not pieced together by chance with building blocks.

②Short development cycle: “Field programmable”

FPGA allows you to program products "on-site", which shortens the debugging cycle and allows products to quickly adapt to new process standards. And after the product has been installed on-site, its hardware can still be directly configured, which is what we call "field programmable". If you can reasonably configure simple logic gates or describe the hardware system you want, you can use these simple logics to do something more meaningful. When your technology is not yet mature, FPGA allows you to make multiple changes in design, and you don't have to bear high costs and some design risks that you have to face in the design process.

③High speed

The high speed of FPGA is reflected in two aspects: parallelism and PLL phase-locked loop

The first point is the parallelism of FPGA. Compared with traditional MCU, the larger the program scale, the more FPGA's advantages can be reflected. Maybe a simple "flowing light" program will not show any difference, but if it is a large system with a large amount of data flow, FPGA can call the internal FIFO core for high-speed processing, and finally can be different from MCU system by tens to hundreds of milliseconds - this is extremely fatal to the communication industry, aerospace industry, and military industry.

The second point is the PLL phase-locked loop. The MCU also contains a PLL phase-locked loop, but it is subject to great restrictions and can usually only multiply the frequency by a few gears. The PLL frequency multiplication on the FPGA has a higher degree of freedom. Taking the MAX10 produced by Altera as an example, the 50MHz clock can usually be multiplied to about 470MHz externally if the chip allows, and the internal clock can be multiplied to 1.3GHz.

Take the example of a car's reversing camera. If the camera takes a total of 250ms of delay from video capture to video display, and the national requirement is that the delay should not exceed 100ms, you can use FPGA to replace the traditional MCU to process image information to meet the government's delay requirements.

④High integration

As the semiconductor industry's products become more and more integrated, programmers are realizing that they can build entire subsystems in FPGAs, rather than being limited to using ASICs (application-specific integrated circuits). They can put more devices in the same size package - allowing for more diverse functions and more complex algorithms, with faster computing speeds and lower power consumption.

⑤ Can embed IP hard core

FPGA can have built-in IP cores, providing rich and powerful functions, so that developers do not need to waste time to repeatedly build the same functional module. At the same time, it will also reduce resource consumption and save a lot of money. At present, the storage block of FPGA can embed many IP cores, such as computing circuits, transceivers, protocol controllers, and even an MCU. However, it is important to remember that IP cores are not reprogrammable like other logic units of FPGA, but they are customized modules to achieve a certain function. In other words, for IP cores, we do not need to reconstruct them every time, which is useless. We can just call them directly.

Because of this feature of IP cores, FPGA manufacturers can sell the completed IP hard cores as commodities. It is also for this reason that most electronic systems tend to be standardized.

led2015

After watching some introductory and quick-learning videos on FPGA, I still feel a lot of pressure. Unless I really have a project to learn, I feel it is difficult to learn.