>Although it is open, the amount of effort required to design and make SoCs that effective compete with ARM licensed designs could cost more than simply paying ARM licensing fees and royalties.
With ARM, only ARM can license designs to others. They'll license the ISA to you, but you'll still be unable to license your ARM implementations to others.
With RISC-V, everybody gets a free license to the ISA and anybody can license their own designs to others. As a result, there's a lively market of cores, which has been growing geometrically for the last ~3 years.
>Ultimately if a lot of experienced ARM designers defect to RISC-V firms, then the balance could tilt to RISC-V.
SiFive is, on high performance, about 2 years behind ARM's fastest designs right now. On power efficiency and area, they're already very competitive.
Tenstorrent has a team led by Jim Keller working on a high performance implementation.
Rivos, full of ex-Apple, ex-P.A.Semi engineers, is working on another.
And these are just some examples. Several other parties have huge RISC-V efforts ongoing.
RISC-V having competitive high performance implementations is something I fully expect to happen within 2 years, rather than just a possibility.
With ARM, only ARM can license designs to others. They'll license the ISA to you, but you'll still be unable to license your ARM implementations to others.
With RISC-V, everybody gets a free license to the ISA and anybody can license their own designs to others. As a result, there's a lively market of cores, which has been growing geometrically for the last ~3 years.
>Ultimately if a lot of experienced ARM designers defect to RISC-V firms, then the balance could tilt to RISC-V.
SiFive is, on high performance, about 2 years behind ARM's fastest designs right now. On power efficiency and area, they're already very competitive.
Tenstorrent has a team led by Jim Keller working on a high performance implementation.
Rivos, full of ex-Apple, ex-P.A.Semi engineers, is working on another.
And these are just some examples. Several other parties have huge RISC-V efforts ongoing.
RISC-V having competitive high performance implementations is something I fully expect to happen within 2 years, rather than just a possibility.