LWN: Comments on "6th RISC-V Workshop Proceedings"

6th RISC-V Workshop Proceedings

ttelford — Tue, 20 Jun 2017 00:57:47 +0000

The big thing is that RISC-V is a more modern version of the DLX processor - which is literally _the_ textbook processor design. Practically every engineer who has taken a class on CPU design implemented the DLX.

RISC-V is not much of a leap from that.

So the first problem is that the design is very similar to a design virtually everyone in the field already knows.

The second problem is That RISC-V is simple/elegant enough it doesn't take a world-class engineer to understand it. It can easily replace DLX as the textbook design.

The third problem is that fully synthesizable and open source implementations of RISC-V exist already. If you have $100 you can get an FPGA and program it to be a RISC-V.

The fourth problem is that other large companies are already planning on using RISC-V going forward. For example, NVIDIA plans on using it to replace a management component in its GPU's.

6th RISC-V Workshop Proceedings

XERC — Tue, 13 Jun 2017 12:56:58 +0000

Is there anything that stops an Oracle-versus-Google type of case from happening with a small, academic, "RISC-V Our-nice-CPU-or-software Foundation"?

If the patent trolls have more money than the small men, then they can collect taxes just like the federal government collects taxes from smaller businesses. Be it with lawyers or by just saying that we've got more money to pay to assassins than You do.

That is to say, what's really protecting the RISC-V foundation and the Red_Hat-in-its-early-days from 90-ties Microsoft'sand 2017 Oracle, other than the plain incompetence of the adversaries? Like, would the Linux Foundation and RISC-V Foundation survive, if they resided in Russia and Putin with its KGB/FSB chose to kill some of the "dissidents"? Or may be You want to develop some Signal or Telegram.org in China or Iran? What's really stopping the Russian/Chinese/Iranian case from happening with the Our-Nice-Little-Foundation in the U.S.? It's not that Microsoft didn't want to destroy Linux, back in the days of the Halloween Documents and SCO FUD, only the incompetence and lack of proper effort by the Microsoft officials saved Linux from being outlawed by the Microsoft.

What am I missing? What am I not seeing?

6th RISC-V Workshop Proceedings

gdt — Sat, 03 Jun 2017 14:10:18 +0000

Also interesting was Dave Patterson's take on past and future of CPU architecture: youtu.be/1FtEGIp3a_M.

6th RISC-V Workshop Proceedings

pabs — Wed, 31 May 2017 03:19:02 +0000

I found bunnie's talk about hardware freedom and transparency interesting:

https://riscv.org/wp-content/uploads/2017/05/Wed1100-impe...
https://youtu.be/zXwy65d_tu8