Explaining RISC-V: An x86 & ARM Alternative

RISC-V is an alternative microprocessor technology to x86 and ARM, with its instruction set architecture (ISA) being open rather than closed. This video explains what RISC-V is all about, including its origins, key market players, hardware, applications, intellectual property (IP), and the likely role of global politics and international trade barriers in determining RISC-V’s success.

My previous review of the VisionFive RISC-V SBC that can run a Linux OS is here:

And my review of the Nezha RISC-V SBC that can also run a Linux OS is here:

REFERENCES
Specific sources included in the video are as follows:

The first RISC-V Instruction Set Manual (from 2011):

Alibaba (T-Head) open sources Xuantie RISC-V cores:

Chapters:
00:00 Introduction
01:26 Open & Closed ISAs
03:55 RISC-V Origins
05:15 Market Players
08:43 Entering the Mainstream
10:34 The Third Platform

#RISC-V #x86 #ARM #ExplainingComputers

[derekla]

12:30 The biggest reason China, India, and Russia want to move away from x86 is because the built in Management Engine on Intel and AMD's Secure Technology can be used as backdoors by government intelligence agencies.
Those subprocessors have been included on CPUs for over a decade now. On normal units, trying to disable the IME will render the processor unbootable. Except for special high security models made exclusively for the US Govt which have it hardware disabled from the factory.

[scottduede]

This is easily the best open-source/closed-source hardware explanation I have been given. Thank you, ExplainingComputers! I really needed this.

[alliejr]

It comes down to compiler and compiler-related technologies that drive virtualization as Apple has demonstrated with it's seamless move from x86 to (Apple) ARM.

[carecavoador]

Somehow Chris seems to read my mind. As he almost every week publishes a video about a subject I've been thinking about. As always, a very interesting video, very informative and top quality production. Thank you so much!

[PeterJasper]

Open architecture and interoperability appear to be the hallmarks of advancement in technology as it fosters continued development and competition. I am sure that RISC-V will do well in an ‘uncertain’ world. Thanks for another great video.

[vdivanov]

Great presentation on a pertinent topic, as always! Many thanks!

[rorytruman]

When I started out as an engineer the principle CPU I encountered was the Motorola 68K, CISC soon became dominant not necessarily because it was better but more due I believe to commercial partnerships. As a big fan of RISC over CISC during the late 80s and early 90s the debate as to which one was better fizzled out much to my disappointment so it is good to see a resurgence in interest. It just proves that good ideas never go away and the next few years promise to be quite exciting.

Thanks for covering this subject Chris, I agree that geopolitical concerns will likely trigger an end to the dominance of the small number of players driving the current hardware/software market to the benefit of consumers who will have more choice in how they do their computing. Similarly crucial is open and common messaging protocols to support open computing

[RoboNuggie]

A great look at RISC-V, and something I think will be a force majeure in IT terms.... exciting!
Thank you Christopher!

[MichelMorinMontreal]

I don't think there are many YT channels where processor architecture, global economics, geo-political considerations and "Open-source" issues are discussed in a single and clear presentation! I think the correct expression is "Master Class"! And, indeed, Professor Barnatt, it seems clear (to me) that "This is going to happend"! Congratulations and thanks!

[lawrenceallwright]

"Unlike some in the industry, I do believe that this is going to happen."
And if Chris says it's going to happen, it's going to happen.
Another eloquently explained video, thank you for all the research that you put in to bring us these insights, Mr B.

[johnmckown]

One ISA which was not mentioned is from IBM in its proprietary "enterprise" mainframe, the "series z" which is the great^n grandchild of the S/360 mainframe from the 1960s. Granted, this is not something the average person will come across. But it is a very interesting, to me, because I've worked on them for 40+ years. It is a very CISC architecture.

One reason for its popularity in very large enterprise systems is its extreme reliability. I have had a CPU fail & the hardware did a fall over to a backup CPU in the machine with NO outage to the code running on the failed CPU. That code was restarted by the hardware automatically and the OS was simply "informed" of the fact for reporting purposes. In addition, the machine automatically "called" IBM service, who came out and "hot swapped" a new CPU replacement. Again, with no outage to anything running on the machine.

I love the very insightful analysis at the end of the video. This is the first material about RISC-V I have seen on Youtube which takes into account not only technical aspects, but also economical and political ones. Like the Shakti initiative, which is largely ignored, but will probably soon have a huge impact on the whole chip market.

[stolz]

Great Video! As russian I may confirm that inside my country there are some attempts to build own CPU. But there is a problem: it's relative easy to design CPU, but it's very hard to establish a production and even harder to have enough of software. RISC-V Instruction Set is very variable. So here is the problem - CPU-designers should have specific compilers for specific CPU...

[charlesburnaford]

The first RISC computer that I worked on was the DEC PDP 8. which had only 13 instructions. The assembly language source code for the BASIC interpreter was about 6000 lines of code in PAL. It ran in 4 to 8K of ferrite core memory.

[locusf]

Cost is also going to be a factor, since the licensing fees to produce ARM/x86 ISA chips is probably quite eye-watering.

EDIT:
Lot of really good replies have created more information on this. Top-of-the chain costs are minimal but it can be so that during chip development the costs can be expected to rise.

Also one could consider the fact that chinese chip makers were embargoed from erratas from ARM due to international issues. I believe the embargo was not that long (some months) and has been lifted since.

[encyclpedia-]

I think it's always a good thing for hardware and software not to belong to monopolies or oligopolies. That NVidia Effect you mentioned is a prime reason why having so few control so much is dangerous and nerve-wracking. ISPs here in the United States have oligopolies and geographical monopolies in many regions and charge way more than pretty much everywhere else in the world for internet access.

[blademan]

The last 4 min is the most conversational you’ve ever sounded. Please have more conversational commentary!

[IAmPattycakes]

I'm just hoping there are massively parallel RISC-V server chips at some point soon. Web hosting doesn't need crazy fast single core, and is basically implicitly parallel. Getting even half the performance for that workload with a quarter the silicon being used is very reasonable when you see the amount of bloat that's crammed into x86 now that doesn't ever get used in many programs. Most of the software I write would definitely benefit from that, but I'm only a one server kinda guy right now so I don't really drive the market.

[billkaroly]

Wow. Great presentation. This made me think back to the MOS 6502 for some unrelated reason.

[v]

I first learned about RISC-V when I was looking up the PS2 and how I could make homebrew games for it. The PS2 was made with a RISC processor and it was made BRILLIANTLY by Sony.