3 ms·

If it was my call I'd skip ARM and go straight to Risc-V. Arms days are numbered - they have reached their peak and will likely stop growing and show a gradual

If it was my call I'd skip ARM and go straight to Risc-V. Arms days are numbered - they have reached their peak and will likely stop growing and show a gradual decline as Risc-V becomes increasingly main stream.

zokula 4y ago
[dead]
booi 4y ago
the vast majority of the difficulty of building something isn't the instruction set but all the various IP required to build a complex and performant chip. For ARM there's a large ecosystem of IP for almost everything. It's great that RISC-V is an open instruction set but it will be a while before the right IP made and licensable in RISC-V.
- strangattractor 4y ago
  True - but building IP for Risc-V will likely go faster - primarily because it is open source, design tools are better and there are companies like TSMC that are foundries and will make whatever you design. I first started working with Arm when it was stilled laid out by hand. It is a great little processor but in a way the licensing thing has retarded progress in certain areas. Back in the day it was an innovative business decision to license only but the world has changed. Arm was launched in 1985. It has had an extraordinary run. Supermicro tends to stick with tried and true so maybe Arm would be the most cost effective for them. However as you said the instruction set is not the deal breaker any longer - but expensive licenses and proprietary IP can be.
- imtringued 4y ago
  The difficulty in building something is primarily the availability of hardware. If the hardware is not available nobody will buy it. Most ARM server vendors quit before they even had a product on the market.
strangattractor 4y ago
Looky here Mips is mmaking a high performance multicore CPU for the server environment. That certainly didn't take that long. https://www.mips.com/products/risc-v/ https://www.mips.com/products/risc-v/

zokula 4y ago

[dead]

booi 4y ago

the vast majority of the difficulty of building something isn't the instruction set but all the various IP required to build a complex and performant chip. For ARM there's a large ecosystem of IP for almost everything. It's great that RISC-V is an open instruction set but it will be a while before the right IP made and licensable in RISC-V.

strangattractor 4y ago
True - but building IP for Risc-V will likely go faster - primarily because it is open source, design tools are better and there are companies like TSMC that are foundries and will make whatever you design. I first started working with Arm when it was stilled laid out by hand. It is a great little processor but in a way the licensing thing has retarded progress in certain areas. Back in the day it was an innovative business decision to license only but the world has changed. Arm was launched in 1985. It has had an extraordinary run. Supermicro tends to stick with tried and true so maybe Arm would be the most cost effective for them. However as you said the instruction set is not the deal breaker any longer - but expensive licenses and proprietary IP can be.
imtringued 4y ago
The difficulty in building something is primarily the availability of hardware. If the hardware is not available nobody will buy it. Most ARM server vendors quit before they even had a product on the market.

strangattractor 4y ago

Looky here Mips is mmaking a high performance multicore CPU for the server environment. That certainly didn't take that long. https://www.mips.com/products/risc-v/ https://www.mips.com/products/risc-v/