In recent years, with the gradual maturity of the process and IP, 32-bit MCU is growing rapidly, the strength of the wind and even 16-bit MCU has basically been skipped. Nowadays, when it comes to embedded MCU, it is either 8-bit or 32-bit, and there are only a few 16-bit MCU product models.
So how about the case of 8-bit MCU, many embedded engineers have some misconceptions, the following is a brief analysis.
Misconception 1
8-bit MCU is being eliminated
This is the most common misconception. First of all, the facts, according to the latest Gartner market report, 8-bit market revenue and growth compared to 32-bit are just a few percentage points worse.
Considering the fact that 8-bit individual chips are much cheaper than 32-bit chips, 8-bit shipments are actually much higher than 32-bit ones.
To use an intuitive analogy, now we have a high-speed rail, is not all the traditional general express, express trains are immediately eliminated?
Obviously this is not the case, and there are too many reasons for this. The reality is that the application areas where 8-bit MCUs once existed cannot be immediately replaced directly with 32-bit MCUs.
Misconception 2
8-bit MCUs lack innovation
Many people would think that since the market is now dominated by 32-bit MCUs, manufacturers are not investing R&D resources in 8-bit products. People who think so may think of 8-bit MCU, the image of 40DIP “classic 8051” will come to mind.
In fact, chip makers have not stopped innovating. For example, the CIP-51 core uses a clock cycle equivalent to an instruction cycle design, instantly increasing the performance of the 8051 by 12 times at the same frequency. Some domestic semiconductor manufacturers also have innovations based on the 8051 or other 8-bit cores.
Misconception 3
8-bit MCUs are difficult to program with C/C++
If you understand the design principles of Arduino, this misconception will be self-evident.
Of course, frankly speaking, 8-bit MCUs are more difficult to program in high-level languages than 32-bit MCUs, and the main obstacle is the inconsistency of memory addresses.
For example, the memory address of the 8051 core is divided into CODE, data, sfr, idata and xdata, and it is even more complicated when it comes to banking. 8-bit PICs also have a more “non-mainstream” design such as hardware Stack, but these obstacles can be alleviated by tool optimization. The 8-bit PIC also has more “non-mainstream” designs like hardware Stack, but these obstacles can be mitigated by tool optimizations.
Misconception 4
8-bit MCUs are made for simple applications
This view is true, but embedded applications are mostly simple applications. The characteristics of embedded system applications themselves determine that there are still many uses for 8-bit. The evolution of peripherals and compilers will slowly expand the scope of 8-bit processor applications.
Misconception 5
8-bit MCUs are not capable of handling the needs of IoT applications
IoT applications are not a single application, but a composite application. Smart watches, smart speakers, host controllers, and gateways of course require complex processors to implement.
But IoT applications also contain a large number of sensor nodes, execution nodes and conversion nodes. Such nodes are more suitable to be implemented with low-power 8-bit processors.
Misconception 6
8-bit MCU response is slow
This is a complete misunderstanding.
In a typical embedded application, the response speed is mainly related to interrupt response and wake-up latency. 8-bit processors have the natural advantage (less address translation effort, fewer gates implemented in IP cells) to be at least as good as 32-bit processors.
Myth 7
8-bit MCUs are less energy efficient than 32-bit MCUs
I once read a book written by an authoritative engineer from ARM. The book’s point of view is that the energy efficiency ratio of 32-bit processors is higher than that of 8-bit MCUs, on the grounds that 32-bit processors can quickly finish processing tasks and have a greater proportion of hibernation time, but this conclusion contains an assumption that the task has a certain degree of complexity.
If the task itself is very simple and the power consumption of the wake-up process is large, then this assumption is not valid. For different application scenarios, you cannot simply say which has a higher energy efficiency ratio, 8-bit or 32-bit. At least in very simple applications, the energy efficiency ratio of 8-bit is higher. If we add some tasks that are individually responsive and do not require CPU intervention, the energy efficiency ratio of 8-bit can even be much higher.
Myth 8
32-bit MCUs are far more powerful than 8-bit MCUs for the same price
There is a degree of credibility to this one, but don’t forget that there are a large number of applications where an 8-bit MCU is sufficient, in which case the cost of having to buy a 32-bit MCU at a higher average price will go up. For many embedded products that are largely standardized, 8-bit MCUs still have some cost advantages.
Myth 9
8-bit MCU design applications can not adapt to future changes
This is a matter of thinking perspective, as an embedded programmer, it is more important to consider the task at hand. No matter what type of MCU, if the product form has changed or the demand itself has changed, it should be redesigned. The future is not visible to anyone.
Misconception 10
8-bit MCUs are more onerous to develop and have no upgrade path
32-bit processors are more software-centric and can do more code reuse. 8-bit processors, on the other hand, make more use of hardware peripherals to get the job done. On balance, there is no absolute difference.
As long as it is an embedded processor, the upgrade path is not very clear. If you use a manufacturer that has both 8-bit and 32-bit products, you will find that many of the peripherals are very similar. Considering the current trend of graphically configuring peripherals, the upgrade path gradually becomes less important, as they are all graphical or scripted to generate the base driver code anyway.
