Technical Analysis of ARM7&ARM9 Dual-core Platform

       At present, the leading products in the embedded system teaching platform market are based on ARM7 or ARM9 architecture. It is generally believed that ARM7 is a low-end product and ARM9 is a high-end product. There is also a so-called "ARM7 & ARM9 covering high-end & low-end teaching platform". It is advertised that "both sets of CPU sub-boards can be freely plugged and unplugged, and one set of experimental system can be changed into two sets. The ARM7 experimental system can realize basic ARM embedded teaching, mainly including instruction experiments, basic interface experiments, UCOS-II operating system experiments and uCLinux operating system experiments; the ARM9 experimental system can realize high-end ARM embedded teaching, mainly including extended interface experiments, Linux operating system experiments and WinCE operating system experiments." 

       This view is suspected of misleading users. Because ARM9 and ARM7 belong to ARMv41, they are low-end series of ARM microprocessors. At present, the real high-end ARM architecture processors on the market are Intel XScale compatible with the ARMV5TE system, such as PXA255 and PXA270. 

       From the perspective of the development of embedded system teaching platforms, there will be two development directions in the future. That is, on the one hand, it develops towards the high-end XScale series, which is mainly aimed at computer, software and other majors. This type of high-end platform has powerful computing power and multimedia functions. The teaching content focuses on operating systems, drivers and software applications, and cultivates embedded software talents in the fields of consumer electronics, handheld devices, wireless networks, mobile games, etc.; on the other hand, it is the mid- and low-end teaching platform of the ARM7/ARM9 series, which is mainly aimed at electronic engineering, automation, instrumentation and other majors. This type of platform has rich interfaces and functions, and the teaching content focuses on microprocessor interface design, driver development and system applications, and cultivates embedded technical talents in application fields such as industrial automation, measurement and control, and intelligent instruments.

       It is completely unnecessary to claim that two CPU daughter boards are used to realize the functions of ARM7 & ARM9, and it increases the cost and complexity of maintenance for users. The reasons are as follows: 

       1. From the teaching content of ARM architecture, the instruction set of ARM9 is fully compatible with ARM7, and there is no difference in teaching. Therefore, ARM instruction experiments and basic interface experiments are not the patents of ARM7. People who have done ARM system development know that these teaching experiments of ARM7 can also be completed with ARM9; 

       2. From the teaching content of operating system, most of the current teaching uses µCOS-II or Linux. µCOS-II has simple code and is easy to teach and learn. Linux has powerful functions, but it has high basic requirements for students. At present, most ARM7 teaching platforms on the market support µCOS-II or uCLinux, and ARM9 basically supports Linux and WinCE. But µCOS-II is not the patent of ARM7 and can run on ARM9. And uClinux is a subset of Linux. Whether from the perspective of developers or from the perspective of teaching, the Linux system can be fully compatible with uCLinux applications. 

       The main reason for claiming to support the dual cores of ARM7 and ARM9 is that their technical capabilities are weak and they cannot complete the porting and expansion of µCOS-II on ARM9, so they use ARM7 to make up for the teaching content of µCOS-II. Imagine, if there is an ARM9 platform that can run operating systems such as µCOS-II, Linux, WinCE, etc., why do we need ARM7 to make up for it? In fact, such a platform already exists. Some companies with strong technical strength have ported µCOS-II to ARM9, such as UP-NETARM2410 and UP-NETARM2410S. 

       3. From the perspective of the teaching content of hardware design, the dual core platform that supports ARM7 and ARM9 has caused a waste of hardware resources and it is difficult to give full play to the respective advantages of ARM7 and ARM9. For example: using s3c44b0 and s3c2410 processors as the cores of ARM7 and ARM9, the s3c2410 platform supports USB host and USB client, and supports true color TFT LCD. In order to be compatible with s3c44b0, the main platform has to use a 256-color STN LCD; if it wants to have a USB host or client interface, it has to use other chips for expansion. This will limit the functions of the ARM9 processor and cannot give full play to the performance of the ARM9 processor. 

       4. From the perspective of product cost service maintenance, the use of a dual-core platform that supports ARM7 and ARM9 will undoubtedly increase the cost of the product, because the most expensive chips in the teaching platform are the microprocessor and memory, and from a teaching perspective, this part of the increased cost is meaningless. From a scientific perspective, this approach wastes resources, and its cost-effectiveness is no better than a development board. In addition, the use of a dual-core platform that supports ARM7 and ARM9 reduces the reliability of the system. Students will often switch between the two cores during the experiment, and the probability of damage will greatly increase, increasing the subsequent maintenance cost. It 

       can be seen that only one ARM9 platform can meet the needs of mid- and low-end teaching at the same time. Adding an additional ARM7 to meet low-end teaching tasks can only be considered superfluous. It simply increases the hardware cost and limits the expansion of the hardware platform. 

       The purpose of embedded system teaching should be to let students learn a method of embedded platform development and design. The changes are mainly reflected in running different operating systems, rather than whether the hardware uses ARM7 or ARM9. The teaching idea should be "teaching people how to fish". What students learn is the design method. In the future, whether it is ARM7, ARM9, or XScale, or even MIPS, alpha, 68k, powerpc and other microprocessors of other architectures, they are all the same for an excellent embedded system engineer.