RISC-V is a free and open source software architecture for microprocessor design. It is designed to make it easier for developers to create efficient, low-power processors. RISC-V was created by the Open Source RISC Consortium, a group of leading technology companies and universities. The consortium’s goal is to promote the use of free and open source software in the design of microprocessors. The architecture is based on the ARMv8 architecture, which is used in many high-end smartphones and tablets. RISC-V also supports other architectures, such as MIPS and PowerPC.
An Alternative to Intel and ARM Designs
Presently, two processor designs reign supreme: those created by ARM and Intel’s x86. While both companies operate at a tremendous scale, their business models are fundamentally different.
Intel designs and manufactures its own chips, while ARM licenses its designs to third-party designers, like Qualcomm and Samsung, which then add their own enhancements. While Samsung has the infrastructure to fabricate its processors in-house, Qualcomm (and other “fabless” designers) outsources this important work to third parties.
In the case of ARM, this also often requires licensors to sign non-disclosure agreements designed to keep aspects of a chip’s design private. That’s hardly surprising, considering its entire business model isn’t shaped around manufacturing, but rather, intellectual property.
Meanwhile, Intel has its own commercial design secrets under lock and key. Since both processor types are commercial, it’s hard (if not entirely impossible) for academics and open-source hackers to influence the design.
How RISC-V Is Different
RISC-V is wildly different. First, it’s not a company. It was first conceived in 2010 by academics at the University of California at Berkeley as an open-source, royalty-free alternative to the existing incumbents.
It’s similar to installing Linux instead of Windows so you don’t have to buy anything or agree to any onerous licensing agreements. RISV-V aims to do the same for semiconductor research and design.
ARM also licenses both the instruction set architecture (ISA), which refers to the commands that can natively be understood by a processor, and the microarchitecture, which shows how it can be implemented.
RISC-V merely offers the ISA, allowing researchers and manufacturers to define how they actually want to use it. This makes it scalable for devices of all stripes, from low-powered, 16-bit chips for embedded systems, to 128-bit processors for supercomputers.
As the name suggests, RISC-V uses the reduced instruction set computer (RISC) principles, the same as chips based on ARM, MIPS, SPARC, and Power designs.
What does this mean? Well, at the heart of any computer processor, there are things called instructions. In the most basic of terms, these are small programs represented in hardware that tell the processor what to do.
RISC-based chips typically have fewer instructions than chips using a complex instruction set computer (CISC) design, like those offered by Intel. Furthermore, the instructions themselves are far simpler to implement in the hardware.
Simpler instructions mean chip manufacturers can be far more efficient with their chip designs. The trade-off is these relatively complex tasks aren’t performed by the processor. Instead, they’re broken down into multiple, smaller instructions by software.
As a result, RISC has earned the nickname Relegate the Important Stuff to the Compiler. While that sounds like a bad thing, it isn’t. To understand it, though, you first have to understand what a computer processor actually is.
The processor in your phone or computer consists of billions of tiny components called transistors. In the case of CISC-based chips, many of these transistors represent the various instructions available.
Since RISC chips have fewer, simpler instructions, you don’t need many transistors. This means you have more room to do a lot of interesting things. For example, you could include more cache and memory registers, or extra functionality for AI and graphics processing.
You can also make the chip physically smaller by using fewer overall transistors. This is why RISC-based chips from MIPS and ARM are frequently found in Internet of Things (IoT) devices.
The Need for Speed
Of course, licensing isn’t the only rationale for RISC-V. David Patterson, who led the first research projects in RISC processor design, said RISC-V was designed to address the impending limits on CPU performance that can be gained from manufacturing improvements.
The more transistors you can fit on a chip, the more capable a processor ultimately becomes. As a result, chip manufacturers like TSMC and Samsung (which both manufacture processors on behalf of third parties) are working hard to shrink the size of transistors even more.
The first commercial microprocessor, the Intel 4004, had just 2,250 transistors, each measuring 10,000 nanometers (about 0.01mm). Small, certainly, but contrast that with Apple’s A14 Bionic processor, released 40 years later. That chip (which powers the new iPad Air) has 11.8 billion transistors, each measuring 5 nanometers across.
In 1965, Gordon E. Moore, the cofounder of Intel, theorized that the number of transistors that could be placed on a chip would double every two years.
“The complexity for minimum component costs has increased at a rate of roughly a factor of two per year,” Moore wrote in the 35th-anniversary issue of Electronics magazine. “Certainly, over the short-term, this rate can be expected to continue, if not to increase. Over the longer term, the rate of increase is a bit more uncertain, although there is no reason to believe it will not remain nearly constant for at least 10 years.”
Moore’s Law is expected to cease to apply this decade. There’s also considerable doubt as to whether chip manufacturers can continue this trend toward miniaturization long-term. This applies at both the basic scientific level and the economic one.
Smaller transistors are, after all, vastly more complicated and expensive to manufacture. TSMC, for example, spent over $17 billion on its factory for creating 5 nm chips. Given this brick wall, Risk-V aims to address the problem of performance by looking at ways besides shrinking the size and number of transistors.
Companies Are Already Using RISC-V
The RISC-V project started in 2010, and the first chip using the ISA was manufactured in 2011. Three years later, the project went public, and commercial interest soon followed. The technology is already being used by companies like NVIDIA, Alibaba, and Western Digital.
The irony is there’s nothing inherently groundbreaking about RISC-V. The Foundation notes on its webpage: “The RISC-V ISA is based on computer architecture ideas that date back at least 40 years.”
What, arguably, is groundbreaking, though, is the business model—or lack of one. It’s this that exposes the project to experimentation, development, and, potentially, unfettered growth. As the RISC-V Foundation also notes on its website:
“The interest is because it is a common free and open standard to which software can be ported, and which allows anyone to freely develop their own hardware to run the software.”
At this writing, RISC-V chips largely toil behind the scenes in server farms and as microcontrollers. It remains to be seen whether there’s any potential to shake up the ARM/Intel ISA duopoly in the consumer space.
However, should the incumbents stagnate, it’s within the realm of possibility that a dark horse could gallop in and change everything.
