>A: You're dreaming, who would be stupid enough to do that?
B: If someone did do this, would it be crazy or awesome?

ARM is essentially closed source and closed from the ecological and instruction set. The same intellectual property (instruction set or IP) is authorized to be used by countless enterprises around the world. The more people use it, the lower the marginal cost.

Although RISC-V instruction set is open source, from the perspective of processor IP research and development, the whole industry belongs to wasteful competition, which also seriously restricts the development process of RISC-V high-performance processor from 0 to 1.

So, is this open source commercial grade RISC-V Processor Core crazy or sillyd or bullish in terms of enterprise strategy?

1. At least ARM doesn't want to see this happen, as it would be a huge boost to the overall RISC-V ecosystem!

2. The Processor Core is not the whole chip, the SoC is a complex systemic design that integrates various IPs and extends to the software ecosystem and tool chain. The entire RISC-V ecosystem is a grasshopper on a string, and if the ecosystem does not work, no one will benefit, and ultimately all will miss this window of opportunity.

3. Whether or not you open source your commercial grade RISC-V Processor Core depends on whether you can focus on long-term strategic benefits or whether you have to take into account short-term benefits (which are not much)?

4. As I have said many times, so many SoC design companies can fatten up one ARM, but definitely not two or three companies like ARM. If RISC-V follows the same business model as ARM and does the same as ARM, it is unlikely to succeed, let alone surpass it. So the only way to succeed is to be more ruthless than ARM, but of course cheapness alone is definitely not the main reason for success. Although, sometimes we say that RISC-V needs to bend the rules, which is escapism! You don't even have the speed, so what's the point of talking about bending the rules?

5. Open source is often a way for large companies to take the ecological initiative and a convention for smaller companies to ride the free ride.

I'm looking forward to the day when, hats off first, RISC-V will enter a whole new phase of development. Maybe tomorrow, maybe the day after, hopefully not for long.

2021/10/19 Ali Yunqi Conference

![](././../../images_dir/1638149936/1.png)

</div>

Lastly, we are working on developing a RISC-V based computing software ecosystem, and we are launching the RVCL open source project (RISC-V Computing Library), which broadly includes high performance computing libraries such as BLAS, MATH/LIBM, VML, FFT, IPP, etc. We want to go beyond the ARM Compute Library from the point of view of performance and meeting new computing needs. The simple idea is that 1. this is what we are good at and 2. if every company builds a computing library optimisation team, it is wastefully competitive. So look forward to working with great companies, and open source will bring the whole ecosystem to life.

---
![](././../../images_dir/1632828056/1.png)