One of the terms that appear the most lately is that of RISC-V, especially since the days when it was said that NVIDIA was going to buy ARM as an alternative to the most used ISA of all. However, despite its benefits, it is important to put the dots on the i’s and make things clear about the huge amount of misinformation circulating on the networks.
In recent years you will have read news about how RISC-V is the future of processors and will soon replace ARM. All this is said without taking into account the difficulty of switching from one set of registers and instructions to another for all the existing programs on the market. Something you can only do if you are a certain Cupertino company that controls its entire ecosystem with a steady hand. That is, it is almost impossible and we have already seen how, for example, the PC is still based on x86, despite the fact that at the time there was talk of switching to ARM. Why should it be any different with RISC-V? That’s why we thought it best to bust some myths.
RISC-V is not what you usually think
Although there are processors that have the RISC-V ISA and there are more and more of them, we must start from the fact that unlike what happens with ARM where the set of registers and instructions is fixed and untouchable, in RISC-V it is modular and, therefore, it is enough that an instruction or a set of instructions does not exist for a program to stop working. And although new instructions are added in all sorts of ISAs to solve new types of problems, in reality what is not done is to touch a base that remains stable in order to maintain the compatibility of different applications.
RISC-V’s approach is different, being modular allows chipmakers to make special-purpose processors and choose the instruction set and therefore units that the chip will carry in its final form. This allows them to focus more on speeding up common tasks and freeing up CPU cores. Although this is very different from an accelerator or a fixed function unit which is only one task.
It’s all still the same as taking a streetcar and removing weight to make it into a competition car. The car is the same, but it is not designed to transport people, but rather to reach the finish line first. This is the key to RISC-V and what will make it popular, not for general use, but for solving small problems. Therefore, its use will become more and more popular, they will even surpass ARM in terms of the volume of use, since we will see several processors under said ISA both in mobile phones and in PCs.
What’s interesting about said ISA?
Well, to the fact that it allows us to create specialized or even general purpose processors without having to pay for an ISA, this is especially important in embedded systems where we must have a processor of a certain power to do make things work, but we have to conform to the immutable specs and conditions of an x86 or ARM processor or we’ll end up paying more for features that will never be used in said application.
Additionally, the goal of every CPU is to make programs run as fast as possible, which is why it’s often ideal to have support units that take the strain off common but repetitive tasks. Although no, anyone can make their own chips, you can design one and apply it via FPGA. This is already starting to happen in the supercomputer market. In the case of the home PC, it will be different with units inside the central processors, graphics chips and even for certain functions of the chipset.