Loongson has overcome 3 major problems and gradually achieved the goal, leaving the last 1 problem to be solved
I have to say that the typical representative of domestic CPU is in the growth history of Loongson, just like the history of fighting monsters and upgrading online games, constantly working hard, growing, solving the shortcomings one by one, and finally becoming the king.
In the past, Loongson actually had three major shortcomings, and these shortcomings seemed very difficult to solve one by one, but after more than 20 years of efforts, now these three shortcomings have been solved by two, and only the last one, Loongson is also working hard to solve it together.
Let's start with the first major disadvantage, which is the dependence on foreign instruction sets.
In the past, Loongson used foreign MIPS, although Loongson later developed its own Loongisa instruction set on the basis of MIPS, but it is still inseparable from MIPS instruction set.
Prior to the Loongson 3A5000, Loongson used the MIPS instruction set, and each time the CPU was sold, a licensing fee was paid to the owner of the MIPS instruction set.
But since 3A5000, the situation is different, Loongson has developed its own LoongArch instruction set, which is 100% self-developed and has nothing to do with MIPS, fundamentally getting rid of the dependence on MIPS, and truly realizing autonomy and controllability.
The second major issue is performance.
For any CPU, performance is a very important indicator, if the performance is not good, no one will use it, and when Loongson was first launched, it lagged behind Intel and AMD and even differed by more than 15 years.
However, Loongson was not discouraged, and continued to work hard to develop and improve the architecture, and by the time of 3A6000, Loongson's performance was almost at the level of Intel's 10th generation Core i3, and the IPC value even reached the level of 13th generation Core i3.
Although on the whole, Loongson lags behind Intel and AMD by 3-5 years, this gap is not a world of difference, and in some fields, it can be completely replaced.
The third big issue is ecology.
For CPU, performance is an important factor, but without ecology, no matter how good the performance is, no one will use it, because the home computer is used to solve problems, and without an ecological computer, it can only create problems, not solve them.
At present, Loongson can only use the Linux ecosystem, which is indeed far from Windows and cannot be solved in a short time.
But now the two major operating systems of Kirin home software and UOS are fully supported by Loongson, and support the latest Loongarch instruction set, these two family systems, the number of compatible software and hardware exceeds 3 million, which can largely meet everyone's daily needs.
It can be seen that although the third problem has not been solved, it can be regarded as partially solved, and with the development of operating systems in various countries, it is estimated that this gap will be solved soon.
When Loongson can fully replace Intel and AMD and become the real domestic CPU of the Chinese people, then we will be able to gradually get rid of the dependence on Intel and AMD in the CPU and embark on an independent, autonomous and controllable road.