It's certainly been exciting to see how quickly RISC-V has progressed to the point of stability, and hardware compatibility. Still would like to see some more competitive options for higher speed cores, but it's successful in areas where that speed is not essential compared to Arm and X86.
I’m still waiting for ARM (and RISC V) to start doing motherboards and replaceable CPUs. Right now upgrading would be a whole new board and that just seems a lot of e-waste.
Arguably the world's best CPU designer, Jim Keller has developed some Riscv cores Maybe he is only using them for AI workloads) So it possible we will see some good Riscv processors sometime. Also Qualcomm is definitely not going to want to be dependant on ARM so they highly likely to be developing RiISCV processors for eg phones
I got email today that my Milk-V P550 board has shipped. That should be twice as fast as the JH7110, TH1520, SpacemiT boards and in fact maybe close to the first i7s in 2009 (or at least the last Core 2 Quads). Only politics is preventing RISC-V from having boards comparable to the Pi 5 / Rock 5 / Orange Pi 5 this year, just two years behind Arm SBCs. Well, who knows .. it seems there might be some other SoCs waiting in the wings. I'm trying to get an Orion O6 which does put Arm ahead a bit more. No doubt Jeff will have early access to one o those (already?)
Well explained. I moved to Linux Mint (on a Ryzen chip) when Windows started to put ads in its search bar. 😮😮 I'm not a gamer. A "simple* processor has more than enough power. So I'm pleased to follow RISC-V's development. Thanks.
I always found it remarkable that one cannot copyright a programming language of one's own design, but that it is possible to copyright ISAs. It's not done the world any favors, and hopefully this will tip the balance.
It's a lot of work and costy to develop a computer architecture. If you can't protect it from being copied, then nobody will put that effort into it. That's why development / investment for RISC-V is rather slow. Whatever someone cooks up, someone else can copy it.
Fascinatingly complex technology topic explained from an economic standpoint. Basically, the risks of RISC-V are nil compared to the risk of continuing down the path of monopolistic processor use. It should open up new ways to use PC's and servers, instead of whatever big tech tells you. Thank you for this video and for all you do.
Chris, love these high information videos. You refreshed a lot of the economics and marketing parts of my business school background and made it relavant to why RISC V truly matters. I look forward to your continued exploration of the RISC V hardware arena. Thanks & Cheers.
RE: "The 80s" I was THERE, Sir! (I'm 58) I fancied myself the world's biggest TOS fan at the Berklee College of Music! I actually made those Atari ST and Mega 4s sequence music and master compact discs well into 2005! (TRUE STORY!) Now I'm all about Raspberry Pis and Logic Pro X on my Mac Mini M-1 - No complaints here!! 😀 Cheers! - Judson & Buddy! Stay warm out there, my fellow super nerds!!
RISC-V is still in its infancy. Until it matures and can match the performance of X86 it will remain an niche player. Even the boards state that they are for early adopters and for development.
The seriousness of this topic is no laughing matter. The thoughtful breakdown of how a trade barrier will see the U.S. shooting itself in the foot is spot on.
Another topic strengthening the case for RISC-V is trust. X86 CPUs already have spyware built into them and it’s impossible to avoid or even know how deep Alice can fall down the rabbit hole. Other proprietary cores provide cover for this premise, making all sorts of networking devices and appliances inherently untrustworthy. An open ISA establishes a context for a certification body to step in and verify that a chip is not backdoored or spying.
Interesting observation. Yes, the closed natures of the ISAs does raise that question. How would we know if there were features of the CPU that would allow for indirect observation of what the CPU is doing?
i am also in the hopes of more privacy, but im a comunist beware of china dictstorship style. Im not computer expert but ive heard their risc v implementation is kinda closed source. Understandable for inovation proteccionism but suspiciois since promoting biambu biambu biambu lol, debian and fedora are getting suport soon. Whats your view on this guys?
Another great consideration is that the development of free/libre open-source software will be significantly easier on an instruction set and architecture that is publicly documented and understood. In particular, developing things like complex drivers should be easier, at least in theory, because no reverse-engineering will be required on that specific hardware. I believe that RISC-V, once it reaches maturity, will be an excellent example of the possibility of combining free software and free hardware for everyone's benefit. Something like RISC-V being open/free will rapidly incite development of free/libre technologies in the same way we see for the Raspberry Pi currently, and I sincerely believe that large corporations like Microsoft will at least initially have trouble capitalizing on RISC-V when the free software alternatives will be developed so rapidly. This might be wishful thinking, but nonetheless seeing a free software boom on RISC-V I feel is very likely.
I wouldn't be very surprised if tablets or low-powered phones with RISC-V processors started appearing in the market in the span of a year (Q4 '25 to Q2 '26), be it China only or low-budget markets (India, SEA, South America). And I remember that 12 or so years ago I was in university and hearing about RISC-V from my professors in university, mostly as a theoretic exercise (learning basic assembly language for that ISA, comparing it to humongous instructions in x86)...
Availability of more options are always better for the consumers. Imagine if there is no Linux Operating System, what will happen to the millions of very usable computers with forced Windows11 requirements.
The big thing for me, besides the open architecture, is the modular nature of RISC-V. Unlike x86 and ARM, you don't need to implement instructions that you don't use. This means the savings of billions of transistors. Or if you keep the transistor count the same then you can implement more cores. There are advantages for cost, power, and heat. I've always loved RISC over CISC. CISC had its performance advantages in the early days of computing, but anymore those advantages do not exist except maybe in special use cases. RISC-V is here to stay, and I believe will eventually become the gold standard, if not the only standard.
Would you really save billions of transistors by eliminating instructions? Maybe a few million or tens of million, but I can't believe the savings would be in the billions.
One of my first jobs was converting cobol programs from ASCII to EBCDIC. A few years later I was working on software for IBM's PowerPC Risc based machines. Being in the non-profit PEG Broadcasting market now we use older computers running various linux distributions as there are plenty of software options to meet our needs. It would be fun to see Risc-V rise up, I sort of thought that IBM would accomplish this with the PowerPC but while it still "sort of lives" it never really caught on.
When speaking of the history of RISC, it's important to note that the IBM PowerPC is a RISC chip. Apple ran these RISC processors in the Mac from 1994-2005, an impressive run for the first gen of desktop RISC chips.
And based off of these are popular game consoles, like PS1-3 or Wii, so it already proved to be viable for the main driver behind most consumer developments. As always, it will only come down to compatibility, but I wouldn't mind an Emulator for the PS3 running at 40watts
The only reason Apple didn't continue with PowerPC is that neither IBM nor Motorola were interested in making PowerPC chip suitable for a new generation of MacBooks to compete with Centrino and the coming Core 2. IBM was interested in high performance unlimited energy usage workstaqtions and servers, while Motorola was interested in power efficient but slow designs (in fact very similar to the RISC-V SoCs we've had available before the Eswin P550). A benefit of RISC-V is that there will be large numbers of vendors targeting every possible market niche, much more so than either x86 or Arm has had to date.
@@paulmilligan3007 Add Alpha (from DEC) to that list. We purchased our first MIPS-based DEC Ultrix workstations in 1991, followed by two Alpha desktops in 1993. One ran DEC OSF1 and the other VMS. RISC is not new. What is new is an open RISC ISA called RISC-V and I'm super excited about it.
Really interesting video. I have hopes for RISC-V, that it will provide some disruption to the stagnant market. I don't know if it's quite fitting for this channel, but perhaps on your other channel you could make a video on the EU / European investment into RISC-V? I would personally like for Europe to have some self sufficiency when it comes to chips, to reduce the dependence on the US an China.
It has been a real treat to see RISC-V evolve over the past couple years and I remain optimistic about its future. Maybe by the end of this decade, we'll start seeing consumer RISC-V conputers hit the market! I really hope I'm right. It's good to be optimistic.
3:43 Intel/AMD are also RISC processors these days. The biggest difference for RISC-V is being an open ISA (i.e. no licensing fees to use it). The actual ISA these days has two layers: The end-user instructions that users see and the microcode that actually executes the instructions through the instruction pipeline. The microcode layer of all CPUs are basically the same. What makes Intel/AMD/NVidia such power-hungry chips is the weird race to be the fastest. The finish line ribbon was crossed about a decade ago but they decided to keep running anyway. Imagine a real race where the runners cross the finish line but decided their legs weren't on fire enough.
The biggest problem with a new ISA is that carefully hand-tuned, high-performance assembly language routines have to be rewritten from scratch. Most ordinary developers don't write a lot of assembly language these days but compiler authors and hardware driver developers do. The Windows x64 emulation layer for ARM is the only reason Windows on ARM still exists. Program compatibility is a problem between ISAs. On GitHub, there are about 25 times as many projects with C code in them than projects with assembly. So assembly language is still used and all of that code has to be ported to work on RISC-V if it is even possible to do so.
As usual that was a fantastic talk. However, as well as an alternative to X86 and ARM with RISK-5, we have an alternative to Windows, macOS and Chrome OS with the Linux distros. Although UA-cam is the best platform for distributing video content, UA-cam is also virtually a monopoly so in the spirit of free enterprise it would be great to see Explaining Computers on Rumble. Thank you again for your informative and easy to follow videos.
While the idea originally behind RISC was to have a processor with a smaller instruction set that could operate faster, it is the technology behind RISC that really matters. CISC computers, and indeed most processors originally were built using a micro sequencer processor to emulate the larger CISC instruction set that is exposed to the programmer. The code run by this micro sequencer is the MicroCode of the CPU. Smaller mini computers were hard wired, using logic gates to decode and execute the instructions, as these machines became more complex, it was easier to build the CPU using the micro arch underneath. This is what RISC machines excel at. However, thanks to Moore's Law, large portions of a CISC computer can now be totally hardwired, and they can rely less on microcode to decode and execute the CISC instruction set. The X86 CPUs today have a hugh amount of the instruction implemented with hard wired logic. Thus, the advantage that pure RISC processors once had over CISC isn't the major selling point anymore. There is quite a bit or RISC technology in today's CISC processors, and the depth of the instruction set used in many RISC processors isn't quite as small as they once were. RISC processors such as the Cortex M4 and M7 have somewhat complex DSP instructions that do quite a bit of work in a single atomic operation.
Even if someone writes code in a higher-level language like C and their code compiles successfully, there is no guarantee that it will work as expected on a new ISA. Intel/AMD, for example, allows for unaligned memory references for both code and data while ARM and RISC-V do not. Unaligned references require a lot of extra work/overhead by the CPU to handle. Alignment is generally handled by the compiler when generating struct code but creating unaligned code/data in C is still doable.
I'm with you and viewers in thinking competition spurs innovation, rather than stagnation. We see that as the Intel behemoth is weakened right now, it is common sense to keep an open mind on an open ISA. Three main different ISAs is not too much if it means forward progress. Thank you to GNU/Linux for allowing this freedom though. 😇
What RISC-V needs is someone to make a full system on a chip (including graphics controller and onboard RAM like the Apple M series or the Qualcomm Snapdragon X Elite series) down to 5 nm circuit process. That is what will be needed to get reasonably powerful RISC-V to run on something like a touchscreen cellphone.
Great video Chris - if little green men were looking down at Earth, I am sure they would have many, many more bigger observations / concerns to consider!
RISC-ARM enabled power efficient CPUs which in turn enabled multiple cores. RISC-V may enable greater specialization of cores because one can innovate without another IP (Intellect Property) contract.
RP2350 Microcontroller: Raspberry Pi introduced the RP2350 microcontroller, which includes RISC-V cores alongside Arm Cortex-M33 cores. This chip allows users to select between RISC-V or Arm cores at startup. The RP2350 is used in the Raspberry Pi Pico 2, showcasing a blend of both architectures for different use cases. 
As you said RISC-V isn't quite there yet for general end user computing so I'm not interested in getting any yet, but I'm keeping an ear open on how they're improving through your videos, Jeff Geerling's, and from other sources. Hopefully one day RISC-V will be good enough to give me the performance of like a current gen Ryzen 7 CPU with all the hardware acceleration for encode and decode I may need to be a viable daily driver. (I value hardware decode especially for media consumption on mobile devices as software decoders aren't that power efficient.) That's assuming the software adoption gets even better to be at the point of being on par with x86_64 or ARM.
Woohoo! Another RISC-V video! Anyway, Mr Christopher, in another recent video you have mentioned your disdain for Dual Booting and as a relatively new subscriber (since mid 2024) I have no idea why do you negatively view Dual Booting a Linux-Windows PC. Can we have a reason for that?
Thanks Chris. Very interesting video. I think RISK-V will force innovation and price reductions. Trade tariffs tend to cause more problems than they resolve. The 'I'm make it more costly for you to trade' ideaset fails because the selling country will put on reciprocal trade barriers. If USA puts up barriers to China, China will do the same to the USA. In my opinion it will be China that wins this tradeoff. The panic to USA if China was to turn off that market would see a massive stock market response. BTW, have you seen the new Pimoron
Maybe one day do one the risc-v market. There are several sbc advertised with the same specs (eg Lichee pi 4a) but different prices. Is there a difference at manufacturer level?
I'm no expert on this stuff, but an open ISA sounds better than one beholden to Arm. That said, for general desktop usage, I have the same concern with RISC-V that I have with ARM. That being locked bootloaders, and Linux distributions having to make images for each specific device. x86 may not be an open ISA, but I can freely install any Linux distro on any piece of x86 hardware that I buy.
x86 used to have performance and still has the best compatibility on the market. aarch64 has amazing power efficiency and recent Apple chips show up also performance potential. I wonder where RISC-V will sit in terms of winning factor? For sure independence of any national state restrictions is an asset. This gives potential advantage for China, their chip designers and fabs. In the world of cloud computing this will spark a server war which should lead to visible advancement in all three ISAs. I hope we will benefit any of them and also this is, by analogy, where I keep fingers crossed for Intel Arc to cut NVidia GPU/AI domination
Is Risc-V faster and uses less power than X86? And also, does Arm have to pay to use Risc-V? Is their version of Risc-V more or less the same or heavily modified.
RISC and CISC are just formal distinctions these days. As far as I understand, no one actually makes pure RISC or CISC processors these days, except for some simpler ones. Perhaps the only advantage of newer CPU architectures (RISC or CISC) is that they don't have to maintain compatibility with older CPU variants of that architecture.
Well if we want to be real honest all current X64 designs are internally RISC with a x84/AMD64 CISC frontend. AMD, at one point, had a opteron design with an ARM frontend but it never turned into a commercial product and no I'm not talking about the Opteron A1100. Edit: Correction Zhaoxin also holds a x86/AMD64 licence (originating from VIA Technologies) en.wikipedia.org/wiki/Zhaoxin and there's a bonch of licence holders for embedded x86 32 bit designs lie form DM&P Electronics and ZF Micro.Not sure if the Russian 8086 clone is still around and if they ever got a licence or just did the Russian thing.
I'm still not convinced... as long as they don't find a way to run Windows on it and natively run industry-standard applications (.exe), it will remain a limited platform mostly confined to hobbyists, specialized markets, and edge devices. Without broader software compatibility, it’s unlikely to compete with established architectures like x86 or ARM in mainstream computing.
Schools should deviate from using monopolistic windows PC to using linux to instill adaptability mindset to more innovation. We need more young minds open to open source projects.
and alternatively it might just be forgotten in 10years. Is there any intrinsic advantage Risk V has over ARM or is it just the license fees? ARM makes only very small profit compared to other tech monopolists. It doesnt look like they are abusing their monopoly position.
@@LivingLinux I am aware of the case between ARM and Qualcomm and you are right that Q would have an interest in alternatives. I was wondering if there is an intrinsic technology advantage,
@@maxlang4440 I don't think there is an intrinsic advantage. But how can you write "It doesnt look like they are abusing their monopoly position", when you know about the court case?
@@LivingLinux I was aware of the case from reporting on arstechnica in 2024 which we can both google . It did not seem like a obvious case of abuse by ARM from their reporting, but of course I may be wrong. Is there a new judgement?. In any case I think the proposed Nvidia takeover of ARM would have been much more serious for the ecosystem.
Morning Sir! It's still FREEZING in OHIO. Anyone have a second hand space heater for sale? What's the lowest possible working temperature for a Raspberry Pi in Fahrenheit? Cheers! - Judson & Buddy !!
I could see some supercomputers equipped with powerful and energy-efficient Chinese RISC-V processors... ** Je verrais bien quelques super calculateurs équipés avec des processeurs RISC-V chinois puissants et économes en courant...
I am surprised you failed to mention Nvidia, who according to tomshardware shipped a billion RISC-V processors in 2024. Yes, they're embedded in GPUs, but as you stated, the economics of choosing RISC-V allowed Nvidia to use it without restrictions. I think that matters a WHOLE LOT to end-user computing.
RISC-V may be great for the profits of manufacturers currently tied to licencing, but it is extremely unlikely that there will be any 'trickle-down' benefit for consumers.
yes RISC-V may promote other companies to get into microprocessor development, where license fees were a deterrent, but for success, doesn't it ultimately just come down to: compatibility, MIPS, & FLOPS? It has to be better/faster, not just different.
If RISC is so good, why haven't the Americans, or at least a certain American, swooped in (On State Department instruction ?) to acquire the license Issuing rights outright. ? That would be in-line with the Incoming administration's Trade Policy i.e. putting a spoke in every other competitors' or potential competitors' wheel . . . . or is it a case of recallng the notion that you never tell your enemy that he is doing the wrong thing or something ineffective.
It's certainly been exciting to see how quickly RISC-V has progressed to the point of stability, and hardware compatibility. Still would like to see some more competitive options for higher speed cores, but it's successful in areas where that speed is not essential compared to Arm and X86.
Give it another year or two and RISC could be a decent competitor in the current ARM dominated market. But we shall see.
I’m still waiting for ARM (and RISC V) to start doing motherboards and replaceable CPUs. Right now upgrading would be a whole new board and that just seems a lot of e-waste.
Arguably the world's best CPU designer, Jim Keller has developed some Riscv cores Maybe he is only using them for AI workloads) So it possible we will see some good Riscv processors sometime.
Also
Qualcomm is definitely not going to want to be dependant on ARM so they highly likely to be developing RiISCV processors for eg phones
Hi, Jeff! I think by the end of this decade, we'll start seeing consumer RISC-V computers hitting the market. Just my hutch.
I got email today that my Milk-V P550 board has shipped. That should be twice as fast as the JH7110, TH1520, SpacemiT boards and in fact maybe close to the first i7s in 2009 (or at least the last Core 2 Quads). Only politics is preventing RISC-V from having boards comparable to the Pi 5 / Rock 5 / Orange Pi 5 this year, just two years behind Arm SBCs. Well, who knows .. it seems there might be some other SoCs waiting in the wings. I'm trying to get an Orion O6 which does put Arm ahead a bit more. No doubt Jeff will have early access to one o those (already?)
Well explained.
I moved to Linux Mint (on a Ryzen chip) when Windows started to put ads in its search bar. 😮😮
I'm not a gamer. A "simple* processor has more than enough power. So I'm pleased to follow RISC-V's development. Thanks.
*"RISC architecture is gonna change everything."* - Kate Libby, aka. 'Acid Burn', Hackers, 1995
She wasn't wrong. The iPhone was 10 years later.
Arm which is risc dominants the microcontroller markey
You didn't watch the video, it seems.
RISC-V is alive!!
Alive and well! :)
I always found it remarkable that one cannot copyright a programming language of one's own design, but that it is possible to copyright ISAs. It's not done the world any favors, and hopefully this will tip the balance.
It's a lot of work and costy to develop a computer architecture. If you can't protect it from being copied, then nobody will put that effort into it. That's why development / investment for RISC-V is rather slow. Whatever someone cooks up, someone else can copy it.
Competition is a good thing, it helps prevent 'things' becoming stagnant and/ or monopolised.
Good video Sir.
I am also an economics student, and I am also interested in RISC-V computing..
Why are you interested in RISC-V computing ?
Well, RISC and ARM is taking off mainly because of interest in lower power consumption and more efficient computing.
Clear and exhaustive, as usual ! Thanks sir for your dedication 👍
Fascinatingly complex technology topic explained from an economic standpoint. Basically, the risks of RISC-V are nil compared to the risk of continuing down the path of monopolistic processor use. It should open up new ways to use PC's and servers, instead of whatever big tech tells you. Thank you for this video and for all you do.
You nearly smiled on that last outro!
Chris, love these high information videos. You refreshed a lot of the economics and marketing parts of my business school background and made it relavant to why RISC V truly matters. I look forward to your continued exploration of the RISC V hardware arena. Thanks & Cheers.
Chris, thank you for a very timely and useful video. Best wishes.
I love RISC-V it's so exciting!!
What in RISC-V is exciting you ?
Me too! :D
I love to see your videos, they are very informative and enjoyable.
RE: "The 80s"
I was THERE, Sir! (I'm 58)
I fancied myself the world's biggest TOS fan
at the Berklee College of Music!
I actually made those Atari ST and Mega 4s
sequence music and master compact discs
well into 2005! (TRUE STORY!)
Now I'm all about Raspberry Pis and Logic Pro X on my Mac Mini M-1 - No complaints here!! 😀
Cheers! - Judson & Buddy!
Stay warm out there, my fellow super nerds!!
Would you go back to a new TOS operating system ?
Oh, wow! That sounds pretty wicked. 😄
You two stay warm yourselves!
AMIIIIIIIIGAAAAAA!!!
:-)
More competition is always good for overall progress and end users.
Thanks for video, Chris 👍!
RISC-V is still in its infancy. Until it matures and can match the performance of X86 it will remain an niche player. Even the boards state that they are for early adopters and for development.
The best thing about RISC-V is that because it is open, it cannot be bought off the market.
edit:
Thanks for keeping us up to date! Great work!
I appreciate your publishing the stance, it's once that I hadn't really embraced until now. With your background, it makes sense!
Is this the channels first BOT?
@@dlewis9760Nope, that's one of our regulars. :)
Hi, Leslie!
The seriousness of this topic is no laughing matter. The thoughtful breakdown of how a trade barrier will see the U.S. shooting itself in the foot is spot on.
In a lot of cases, it's more the threats of tariffs causing the other side to change that succeeds.
Agreed.
It is very exciting times. Things have moved very fast for the risc-v. I am cheering and hope to get my first board with RISC-V this year.
Another topic strengthening the case for RISC-V is trust. X86 CPUs already have spyware built into them and it’s impossible to avoid or even know how deep Alice can fall down the rabbit hole. Other proprietary cores provide cover for this premise, making all sorts of networking devices and appliances inherently untrustworthy. An open ISA establishes a context for a certification body to step in and verify that a chip is not backdoored or spying.
Interesting observation. Yes, the closed natures of the ISAs does raise that question. How would we know if there were features of the CPU that would allow for indirect observation of what the CPU is doing?
i am also in the hopes of more privacy, but im a comunist beware of china dictstorship style. Im not computer expert but ive heard their risc v implementation is kinda closed source. Understandable for inovation proteccionism but suspiciois since promoting biambu biambu biambu lol, debian and fedora are getting suport soon. Whats your view on this guys?
Another great consideration is that the development of free/libre open-source software will be significantly easier on an instruction set and architecture that is publicly documented and understood. In particular, developing things like complex drivers should be easier, at least in theory, because no reverse-engineering will be required on that specific hardware. I believe that RISC-V, once it reaches maturity, will be an excellent example of the possibility of combining free software and free hardware for everyone's benefit. Something like RISC-V being open/free will rapidly incite development of free/libre technologies in the same way we see for the Raspberry Pi currently, and I sincerely believe that large corporations like Microsoft will at least initially have trouble capitalizing on RISC-V when the free software alternatives will be developed so rapidly. This might be wishful thinking, but nonetheless seeing a free software boom on RISC-V I feel is very likely.
Interesting and insightful, as always. Thanks Chris.
Thank you, Sir!
Now I'm really interested in RISC-V.
I wouldn't be very surprised if tablets or low-powered phones with RISC-V processors started appearing in the market in the span of a year (Q4 '25 to Q2 '26), be it China only or low-budget markets (India, SEA, South America). And I remember that 12 or so years ago I was in university and hearing about RISC-V from my professors in university, mostly as a theoretic exercise (learning basic assembly language for that ISA, comparing it to humongous instructions in x86)...
Availability of more options are always better for the consumers. Imagine if there is no Linux Operating System, what will happen to the millions of very usable computers with forced Windows11 requirements.
It's easy to imagine because they would just use another Unix-based thing instead like NetBSD, FreeBSD, OpenBSD and so on..
Keep running 10, I think a lot of people will.
The big thing for me, besides the open architecture, is the modular nature of RISC-V. Unlike x86 and ARM, you don't need to implement instructions that you don't use. This means the savings of billions of transistors. Or if you keep the transistor count the same then you can implement more cores. There are advantages for cost, power, and heat.
I've always loved RISC over CISC. CISC had its performance advantages in the early days of computing, but anymore those advantages do not exist except maybe in special use cases. RISC-V is here to stay, and I believe will eventually become the gold standard, if not the only standard.
Would you really save billions of transistors by eliminating instructions? Maybe a few million or tens of million, but I can't believe the savings would be in the billions.
One of my first jobs was converting cobol programs from ASCII to EBCDIC. A few years later I was working on software for IBM's PowerPC Risc based machines. Being in the non-profit PEG Broadcasting market now we use older computers running various linux distributions as there are plenty of software options to meet our needs. It would be fun to see Risc-V rise up, I sort of thought that IBM would accomplish this with the PowerPC but while it still "sort of lives" it never really caught on.
When speaking of the history of RISC, it's important to note that the IBM PowerPC is a RISC chip. Apple ran these RISC processors in the Mac from 1994-2005, an impressive run for the first gen of desktop RISC chips.
Not to forget SPARC or MIPS, both of which are RISC
Excellent point!
And based off of these are popular game consoles, like PS1-3 or Wii, so it already proved to be viable for the main driver behind most consumer developments. As always, it will only come down to compatibility, but I wouldn't mind an Emulator for the PS3 running at 40watts
The only reason Apple didn't continue with PowerPC is that neither IBM nor Motorola were interested in making PowerPC chip suitable for a new generation of MacBooks to compete with Centrino and the coming Core 2. IBM was interested in high performance unlimited energy usage workstaqtions and servers, while Motorola was interested in power efficient but slow designs (in fact very similar to the RISC-V SoCs we've had available before the Eswin P550). A benefit of RISC-V is that there will be large numbers of vendors targeting every possible market niche, much more so than either x86 or Arm has had to date.
@@paulmilligan3007 Add Alpha (from DEC) to that list. We purchased our first MIPS-based DEC Ultrix workstations in 1991, followed by two Alpha desktops in 1993. One ran DEC OSF1 and the other VMS. RISC is not new. What is new is an open RISC ISA called RISC-V and I'm super excited about it.
Really interesting video. I have hopes for RISC-V, that it will provide some disruption to the stagnant market.
I don't know if it's quite fitting for this channel, but perhaps on your other channel you could make a video on the EU / European investment into RISC-V? I would personally like for Europe to have some self sufficiency when it comes to chips, to reduce the dependence on the US an China.
It has been a real treat to see RISC-V evolve over the past couple years and I remain optimistic about its future. Maybe by the end of this decade, we'll start seeing consumer RISC-V conputers hit the market! I really hope I'm right. It's good to be optimistic.
3:40 No mention of Acorn RISC Machine?
Important topics, thank you 👍
P.S. What's most important is process node mfg capabilities and capacities as well as governance thereof.
Very interesting, and useful, thanks ❤
If you want to jump to a new CPU architecture you have to be willing to take a RISC……..😇
RISCY business
- Tom Cruise
Economics and politics, politics and economics. That's how things work.
3:43 Intel/AMD are also RISC processors these days. The biggest difference for RISC-V is being an open ISA (i.e. no licensing fees to use it). The actual ISA these days has two layers: The end-user instructions that users see and the microcode that actually executes the instructions through the instruction pipeline. The microcode layer of all CPUs are basically the same. What makes Intel/AMD/NVidia such power-hungry chips is the weird race to be the fastest. The finish line ribbon was crossed about a decade ago but they decided to keep running anyway. Imagine a real race where the runners cross the finish line but decided their legs weren't on fire enough.
The biggest problem with a new ISA is that carefully hand-tuned, high-performance assembly language routines have to be rewritten from scratch. Most ordinary developers don't write a lot of assembly language these days but compiler authors and hardware driver developers do. The Windows x64 emulation layer for ARM is the only reason Windows on ARM still exists. Program compatibility is a problem between ISAs. On GitHub, there are about 25 times as many projects with C code in them than projects with assembly. So assembly language is still used and all of that code has to be ported to work on RISC-V if it is even possible to do so.
As usual that was a fantastic talk.
However, as well as an alternative to X86 and ARM with RISK-5, we have an alternative to Windows, macOS and Chrome OS with the Linux distros.
Although UA-cam is the best platform for distributing video content, UA-cam is also virtually a monopoly so in the spirit of free enterprise it would be great to see Explaining Computers on Rumble.
Thank you again for your informative and easy to follow videos.
While the idea originally behind RISC was to have a processor with a smaller instruction set that could operate faster, it is the technology behind RISC that really matters. CISC computers, and indeed most processors originally were built using a micro sequencer processor to emulate the larger CISC instruction set that is exposed to the programmer. The code run by this micro sequencer is the MicroCode of the CPU. Smaller mini computers were hard wired, using logic gates to decode and execute the instructions, as these machines became more complex, it was easier to build the CPU using the micro arch underneath. This is what RISC machines excel at. However, thanks to Moore's Law, large portions of a CISC computer can now be totally hardwired, and they can rely less on microcode to decode and execute the CISC instruction set. The X86 CPUs today have a hugh amount of the instruction implemented with hard wired logic. Thus, the advantage that pure RISC processors once had over CISC isn't the major selling point anymore. There is quite a bit or RISC technology in today's CISC processors, and the depth of the instruction set used in many RISC processors isn't quite as small as they once were. RISC processors such as the Cortex M4 and M7 have somewhat complex DSP instructions that do quite a bit of work in a single atomic operation.
Even if someone writes code in a higher-level language like C and their code compiles successfully, there is no guarantee that it will work as expected on a new ISA. Intel/AMD, for example, allows for unaligned memory references for both code and data while ARM and RISC-V do not. Unaligned references require a lot of extra work/overhead by the CPU to handle. Alignment is generally handled by the compiler when generating struct code but creating unaligned code/data in C is still doable.
I think since Pentium 3 x86 has RISC coe processing and the x86 instructions are implemented as microcode.
I'm with you and viewers in thinking competition spurs innovation, rather than stagnation. We see that as the Intel behemoth is weakened right now, it is common sense to keep an open mind on an open ISA. Three main different ISAs is not too much if it means forward progress. Thank you to GNU/Linux for allowing this freedom though. 😇
What RISC-V needs is someone to make a full system on a chip (including graphics controller and onboard RAM like the Apple M series or the Qualcomm Snapdragon X Elite series) down to 5 nm circuit process. That is what will be needed to get reasonably powerful RISC-V to run on something like a touchscreen cellphone.
Thanks Chris.
risc v❤️❤️🚀
Great video Chris - if little green men were looking down at Earth, I am sure they would have many, many more bigger observations / concerns to consider!
thanks a lot for your explain !
It will be awesome More choices
RISC-ARM enabled power efficient CPUs which in turn enabled multiple cores.
RISC-V may enable greater specialization of cores because one can innovate without another IP (Intellect Property) contract.
I hope that the Raspberry Pi foundation will release a RISC-V board. That would leader other manufacturers to do the same.
RP2350 Microcontroller: Raspberry Pi introduced the RP2350 microcontroller, which includes RISC-V cores alongside Arm Cortex-M33 cores. This chip allows users to select between RISC-V or Arm cores at startup. The RP2350 is used in the Raspberry Pi Pico 2, showcasing a blend of both architectures for different use cases.

IBM came up with RISV architecture 50 years ago. They still use it today.
As you said RISC-V isn't quite there yet for general end user computing so I'm not interested in getting any yet, but I'm keeping an ear open on how they're improving through your videos, Jeff Geerling's, and from other sources.
Hopefully one day RISC-V will be good enough to give me the performance of like a current gen Ryzen 7 CPU with all the hardware acceleration for encode and decode I may need to be a viable daily driver. (I value hardware decode especially for media consumption on mobile devices as software decoders aren't that power efficient.) That's assuming the software adoption gets even better to be at the point of being on par with x86_64 or ARM.
Great video
most RISC-y click of the day 😉
Very RISC-y business!
@@Praxibetel-Ix very disARMing! 🙄🤣
Interesting. Thanks.
I'm gonna skip RISCV and wait for cold-fusion powered quantum computers
Woohoo! Another RISC-V video!
Anyway, Mr Christopher, in another recent video you have mentioned your disdain for Dual Booting and as a relatively new subscriber (since mid 2024) I have no idea why do you negatively view Dual Booting a Linux-Windows PC. Can we have a reason for that?
Where is the chip RP2350 produced ?
Didn't ARM originally stood for Archimedes RISC Machine?
Wikipedia says originally Acorn RISC Machine, then Advanced RISK Machine :)
We need RISC-V processors manufacturers outside of China.
Thanks Chris. Very interesting video. I think RISK-V will force innovation and price reductions. Trade tariffs tend to cause more problems than they resolve. The 'I'm make it more costly for you to trade' ideaset fails because the selling country will put on reciprocal trade barriers. If USA puts up barriers to China, China will do the same to the USA. In my opinion it will be China that wins this tradeoff. The panic to USA if China was to turn off that market would see a massive stock market response.
BTW, have you seen the new Pimoron
Pimornman 5 RPi case? £50+
Because our freedom matters.
Maybe one day do one the risc-v market. There are several sbc advertised with the same specs (eg Lichee pi 4a) but different prices. Is there a difference at manufacturer level?
Zhaoxin still has an x86 license?
I'm no expert on this stuff, but an open ISA sounds better than one beholden to Arm. That said, for general desktop usage, I have the same concern with RISC-V that I have with ARM. That being locked bootloaders, and Linux distributions having to make images for each specific device. x86 may not be an open ISA, but I can freely install any Linux distro on any piece of x86 hardware that I buy.
x86 used to have performance and still has the best compatibility on the market. aarch64 has amazing power efficiency and recent Apple chips show up also performance potential. I wonder where RISC-V will sit in terms of winning factor? For sure independence of any national state restrictions is an asset. This gives potential advantage for China, their chip designers and fabs.
In the world of cloud computing this will spark a server war which should lead to visible advancement in all three ISAs. I hope we will benefit any of them and also this is, by analogy, where I keep fingers crossed for Intel Arc to cut NVidia GPU/AI domination
Is Risc-V faster and uses less power than X86? And also, does Arm have to pay to use Risc-V? Is their version of Risc-V more or less the same or heavily modified.
RISC and CISC are just formal distinctions these days. As far as I understand, no one actually makes pure RISC or CISC processors these days, except for some simpler ones. Perhaps the only advantage of newer CPU architectures (RISC or CISC) is that they don't have to maintain compatibility with older CPU variants of that architecture.
Well if we want to be real honest all current X64 designs are internally RISC with a x84/AMD64 CISC frontend. AMD, at one point, had a opteron design with an ARM frontend but it never turned into a commercial product and no I'm not talking about the Opteron A1100.
Edit: Correction Zhaoxin also holds a x86/AMD64 licence (originating from VIA Technologies) en.wikipedia.org/wiki/Zhaoxin
and there's a bonch of licence holders for embedded x86 32 bit designs lie form DM&P Electronics and ZF Micro.Not sure if the Russian 8086 clone is still around and if they ever got a licence or just did the Russian thing.
risc v is quite good as a low power web server although it is a bit expensive
"RISC will never take off!" Probably posted by someone using a smartphone running on a RISC CPU. Hilarious.
I'm still not convinced... as long as they don't find a way to run Windows on it and natively run industry-standard applications (.exe), it will remain a limited platform mostly confined to hobbyists, specialized markets, and edge devices. Without broader software compatibility, it’s unlikely to compete with established architectures like x86 or ARM in mainstream computing.
Mobile phones and tablets !
Schools should deviate from using monopolistic windows PC to using linux to instill adaptability mindset to more innovation. We need more young minds open to open source projects.
risc-v can easily splinter, adding a specific feature only known to that country, banning imports of foreign chips
RISC isn't rising for the first time now, it actually rose and fell before x86 did in the form of PowerPC.
You forgot MIPS and Hitachi SH.
My soldering Iron runs on risc-v :D
It's always risky to use a soldering iron. So that makes perfect sense 😂
😂😂😂
and alternatively it might just be forgotten in 10years. Is there any intrinsic advantage Risk V has over ARM or is it just the license fees? ARM makes only very small profit compared to other tech monopolists. It doesnt look like they are abusing their monopoly position.
You missed the court case between ARM and Qualcomm, and ARM demanded all Snapdragon X Elite chips to be destroyed? First round was won by Qualcomm.
@@LivingLinux I am aware of the case between ARM and Qualcomm and you are right that Q would have an interest in alternatives. I was wondering if there is an intrinsic technology advantage,
@@maxlang4440 I don't think there is an intrinsic advantage. But how can you write "It doesnt look like they are abusing their monopoly position", when you know about the court case?
@@LivingLinux I was aware of the case from reporting on arstechnica in 2024 which we can both google . It did not seem like a obvious case of abuse by ARM from their reporting, but of course I may be wrong. Is there a new judgement?. In any case I think the proposed Nvidia takeover of ARM would have been much more serious for the ecosystem.
Good Morning my best friend.........
Hey, hey, hey! :)
Morning Sir! It's still FREEZING in OHIO.
Anyone have a second hand space heater for sale?
What's the lowest possible working temperature
for a Raspberry Pi in Fahrenheit?
Cheers! - Judson & Buddy !!
Brrrrrr, stay warm over there!
I don't see RISC going mainstream as long as Microsoft exists.
I could see some supercomputers equipped with powerful and energy-efficient Chinese RISC-V processors...
**
Je verrais bien quelques super calculateurs équipés avec des processeurs RISC-V chinois puissants et économes en courant...
I am surprised you failed to mention Nvidia, who according to tomshardware shipped a billion RISC-V processors in 2024. Yes, they're embedded in GPUs, but as you stated, the economics of choosing RISC-V allowed Nvidia to use it without restrictions. I think that matters a WHOLE LOT to end-user computing.
I wonder who will blink first and make some riscv kit...intel or amd.
Congrats on the one million subs. I guess that means you get BOTs posting now. I see at least 7 right now.
RISC-V may be great for the profits of manufacturers currently tied to licencing, but it is extremely unlikely that there will be any 'trickle-down' benefit for consumers.
I want to risk to program on RISC.
Nice word play 😂
😂
Please keep in mind it's a RISCy business ...
yes RISC-V may promote other companies to get into microprocessor development, where license fees were a deterrent, but for success, doesn't it ultimately just come down to: compatibility, MIPS, & FLOPS? It has to be better/faster, not just different.
Arguments could be made that ARM processors are no longer RISC. They have added instructions that look CISC.
👍
A Raspberry Pi 5 is a RISC machine
Seems to me that in the next 10 to 15 years, RISC-V computing might look a little like Linux distros.A lot of competitors, with a few leading the way.
hopefully these RISC-V cores on which enthusiasts from all the world are grinding will not fall on their heads inside chinese missile
I dont remember microsoft investing that much in their own network. They invested heavily in WWW in the 1990s. The heck are u on about??
If RISC is so good, why haven't the Americans, or at least a certain American, swooped in (On State Department instruction ?) to acquire the license Issuing rights outright. ?
That would be in-line with the Incoming administration's Trade Policy i.e. putting a spoke in every other competitors' or potential competitors' wheel . . . . or is it a case of recallng the notion that you never tell your enemy that he is doing the wrong thing or something ineffective.
But monopolies do innovate. They invent ways to lock out competition all the time.
Even if you're using an x86 processor you're still actually using RISC.
Morning!
Good morning! A silver medal for you. 🥈