Discussion RISC V Latest Developments Discussion [No Politics]

[DisEnchantment]

Some background on my experience with RISC V...
Five years ago, we were developing a CI/CD pipeline for arm64 SoC in some cloud and we add tests to execute the binaries in there as well.
We actually used some real HW instances using an ARM server chip of that era, unfortunately the vendor quickly dumped us, exited the market and leaving us with some amount of frustration.
We shifted work to Qemu which turns out to be as good as the actual chips themselves, but the emulation is buggy and slow and in the end we end up with qemu-user-static docker images which work quite well for us. We were running arm64 ubuntu cloud images of the time before moving on to docker multi arch qemu images.

Lately, we were approached by many vendors now with upcoming RISC-V chips and out of curiosity I revisited the topic above.
To my pleasant surprise, running RISC-V Qemu is smooth as butter. Emulation is fast, and images from Debian, Ubuntu, Fedora are available out of the box.
I was running ubuntu cloud images problem free. Granted it was headless but I guess with the likes of Imagination Tech offering up their IP for integration, it is only a matter of time.

What is even more interesting is that Yocto/Open Embedded already have a meta layer for RISC-V and apparently T Head already got the kernel packages and manifest for Android 10 working with RISC-V.
Very very impressive for a CPU in such a short span of time. What's more, I see active LLVM, GCC and Kernel development happening.

From latest conferences I saw this slide, I can't help but think that it looks like they are eating somebody's lunch starting from MCUs and moving to Application Processors.


And based on many developments around the world, this trend seems to be accelerating greatly.
Many high profile national and multi national (e.g. EU's EPI ) projects with RISC V are popping up left and right.
Intel is now a premium member of the consortium, with the likes of Google, Alibaba, Huawei etc..
NVDA and soon AMD seems to be doing RISC-V in their GPUs. Xilinx, Infineon, Siemens, Microchip, ST, AD, Renesas etc., already having products in the pipe or already launched.
It will be a matter of time before all these companies start replacing their proprietary Arch with something from RISC V. Tools support, compiler, debugger, OS etc., are taken care by the community.
Interesting as well is that there are lots of performant implementation of RISC V in github as well, XuanTie C910 from T Head/Alibaba, SWerV from WD, and many more.
Embedded Industry already replaced a ton of traditional MCUs with RISC V ones. AI tailored CPUs from Tenstorrent's Jim Keller also seems to be in the spotlight.

Most importantly a bunch of specs got ratified end of last year, mainly accelerated by developments around the world. Interesting times.

 

[marees]

Global Foundries to make own brand (RISC-V) hardware

GlobalFoundries to make RISC-V CPUs — fab acquires MIPS, will integrate RISC-V and AI IP into its portfolio​

News
By Anton Shilov published 7 hours ago
For an unknown sum.


GlobalFoundries announced on Tuesday that it had entered into a definitive agreement to acquire MIPS, a developer of RISC-V-based solutions and IP. The deal will enable GlobalFoundries to offer its own processors and other products based on the RISC-V instruction set architecture (ISA), which will make the company a rival to some of its foundry clients. Yet, the companies stress that MIPS will operate as a standalone business.

The takeover enhances GlobalFoundries' IP portfolio with MIPS' technologies, including the company's general-purpose CPU IP, AI inference acceleration IP, and various sensors. MIPS recently expanded its processor offerings based on the RISC-V ISA, so its Atlas product line includes a variety of cores suited for both general-purpose and real-time processing, as well as specialized cores designed for AI edge workloads. These cores are designed to enable high performance at a relatively low power consumption for compute-heavy workloads in embedded systems.


Keep in mind, though, that MIPS integration into GF will likely take years, so it is too early to talk about advantages or additional orders.

Following the completion of the transaction, MIPS will continue to function as an independent business unit within GlobalFoundries. It will maintain its existing relationships with other foundries and customers, and support a wide range of technologies across multiple industries.


https://www.tomshardware.com/pc-com...p-into-its-portfolio#xenforo-comments-3882720

 

[marees]

Global Foundries to make own brand (RISC-V) hardware

GlobalFoundries to make RISC-V CPUs — fab acquires MIPS, will integrate RISC-V and AI IP into its portfolio​

News
By Anton Shilov published 7 hours ago
For an unknown sum.


GlobalFoundries announced on Tuesday that it had entered into a definitive agreement to acquire MIPS, a developer of RISC-V-based solutions and IP. The deal will enable GlobalFoundries to offer its own processors and other products based on the RISC-V instruction set architecture (ISA), which will make the company a rival to some of its foundry clients. Yet, the companies stress that MIPS will operate as a standalone business.

The takeover enhances GlobalFoundries' IP portfolio with MIPS' technologies, including the company's general-purpose CPU IP, AI inference acceleration IP, and various sensors. MIPS recently expanded its processor offerings based on the RISC-V ISA, so its Atlas product line includes a variety of cores suited for both general-purpose and real-time processing, as well as specialized cores designed for AI edge workloads. These cores are designed to enable high performance at a relatively low power consumption for compute-heavy workloads in embedded systems.


Keep in mind, though, that MIPS integration into GF will likely take years, so it is too early to talk about advantages or additional orders.

Following the completion of the transaction, MIPS will continue to function as an independent business unit within GlobalFoundries. It will maintain its existing relationships with other foundries and customers, and support a wide range of technologies across multiple industries.


https://www.tomshardware.com/pc-com...p-into-its-portfolio#xenforo-comments-3882720

MIPS has had a chequered decade. Once a pioneer of RISC processors from the work of John Hennessey and Chris Rowen at Stanford University, it competed directly with ARM for many years with the R2000 designed into the original Playstation console. It was bought by Silicon Graphics Inc (SGI) in 1992 but spun out in 1998. It was bought by UK IP developer Imagination Technologies, which also supplied graphics technologies into consoles, in 2013, before being sold to Tallwood Ventures in 2017 in a complex deal with Chinese links and then onto Wave Computing in 2018. After Wave went bankrupt in 2020 it emerged again as MIPS with a focus on the RISC-V open instruction architecture also developed by Hennessey.

https://www.eenewseurope.com/en/another-risc-v-firm-falls-as-globalfoundries-buys-mips/

 

[marees]

In its 40-year history, MIPS has come under various owners, so EE Times asked how this time would be any different. He told us that we are at an inflection point in edge AI deployment, and so this is a huge new opportunity for MIPS.

“MIPS has been acquired in the past, that’s right. And when I look at that history, what I see is that MIPS has been acquired when unique technologies or interesting markets have emerged. Graphics were early on, followed by smartphones later. What I see now is some similarity, but also a lot of new things,” Wasson commented. “If you look at the edge AI opportunity that is broader both in terms of impact and in terms of potential customer base than anything MIPS has ever seen in the past, because everything we see around us is going to get more intelligent with time. And I think there lies our biggest opportunity in edge AI. So where GF and MIPS can work together is to go after that opportunity because our technologies are suited for that.”

And then he added, “The other thing I see is that semiconductors are also broadening in terms of where in the globe the interest is coming from. Historically, it has been North America, Japan, China, and Europe, but now we’re seeing more centers of excellence and R&D development centers emerging, which will be broader and more global. The very name is global. I think that helps. I think that allows us to reach those customers.”

“And if we can continue working to reduce the adoption barrier, I think you will see customers emerge in regions that don’t exist today. And I think those regions where you have a large TAM [total addressable market], a broader TAM, not only in volume, but also in breadth, and you have more regions for MIPS. This is a great thing because now we can start leveraging GF scale to start attacking those two options,” he argued.


https://www.eetimes.com/mips-ceo-acquisition-by-globalfoundries-strengthens-edge-ai-opportunity/

 

[marees]

In its 40-year history, MIPS has come under various owners, so EE Times asked how this time would be any different. He told us that we are at an inflection point in edge AI deployment, and so this is a huge new opportunity for MIPS.

“MIPS has been acquired in the past, that’s right. And when I look at that history, what I see is that MIPS has been acquired when unique technologies or interesting markets have emerged. Graphics were early on, followed by smartphones later. What I see now is some similarity, but also a lot of new things,” Wasson commented. “If you look at the edge AI opportunity that is broader both in terms of impact and in terms of potential customer base than anything MIPS has ever seen in the past, because everything we see around us is going to get more intelligent with time. And I think there lies our biggest opportunity in edge AI. So where GF and MIPS can work together is to go after that opportunity because our technologies are suited for that.”

And then he added, “The other thing I see is that semiconductors are also broadening in terms of where in the globe the interest is coming from. Historically, it has been North America, Japan, China, and Europe, but now we’re seeing more centers of excellence and R&D development centers emerging, which will be broader and more global. The very name is global. I think that helps. I think that allows us to reach those customers.”

“And if we can continue working to reduce the adoption barrier, I think you will see customers emerge in regions that don’t exist today. And I think those regions where you have a large TAM [total addressable market], a broader TAM, not only in volume, but also in breadth, and you have more regions for MIPS. This is a great thing because now we can start leveraging GF scale to start attacking those two options,” he argued.


https://www.eetimes.com/mips-ceo-acquisition-by-globalfoundries-strengthens-edge-ai-opportunity/

“Take motor control as an example. Most of our initial product focuses on motor control and how we bring robotics into that, apply it to robotics, apply it to automotive, that’s been where a lot of our energy has gone. I think that scales up nicely.”

“That goes very well with GlobalFoundries’ manufacturing footprint on both the mature nodes at 40 nanometers, but also 22 nanometers, and then eventually at 12 nanometers as well,” Wasson added. “So, there is a significant synergy between what GF assets are and how we can overlay that with what we’ve been working on in the markets.”

 

[DrMrLordX]

GF bought MIPS? That's really odd. Even the PRC basically ditched MIPS IP after uh, creatively borrowing it:

Loongson’s LSX and LASX Vector Extensions

Loongson used to make CPUs based off the MIPS ISA, but the company recently switched to a homegrown ISA called Loongarch.

chipsandcheese.com

It sort of makes sense that MIPS is doing RISC-V now since apparently nobody wants to bother with actual MIPS64 anymore.

 

[Io Magnesso]

It's interesting to provide IP while manufacturing

 

[marees]

It's interesting to provide IP while manufacturing

Global Foundries wants to become the intel of RISC-V I guess

 

[Io Magnesso]

Global Foundries wants to become the intel of RISC-V I guess

But it's interesting, so I want to support

 

[DrMrLordX]

Global Foundries wants to become the intel of RISC-V I guess

They still need better process tech.

 

[Io Magnesso]

They still need better process tech.

But the mature process is also important.

 

[Panino Manino]

They still need better process tech.

Actually, "to be Intel" ... maybe no?

News is nice and all, but for me feels like a double lovers suicide. At least they won't die alone, poor souls.

 

[igor_kavinski]

They still need better process tech.

A little birdie told me that Intel's israel fab is going to be abandoned soon.

GoFlo go get it!

 

[Io Magnesso]

But because GF is a manufacturer that has disembarked from the EUV competition...
Even GF can't talk about people...

 

[DZero]

I was crazy, but trying to revive the MISHAP, I mean MIPS right now is a very bad idea...

 

[NostaSeronx]

M8500 is an improvement on I8500.
I8600 is an improvement on P8700/I8500.
S8200 is unique as it isn't shown here but is the RVV + MIPS Matrix core supposedly.
Still waiting if we get a RVA23 P8800 or something from S8200?


Same market, just not same targets.
M-series = Akeana 100 series
I-series = Non-vector Akeana 1000 series
S-series = Vector Akeana 1000 series
P-series = Akeana 5000 series

12FDX/10FD/7FD ~ 2026 is when eSoC 3.0 wafer ships
7FDX/7FD+ ~ 2028 is when Advanced SOI wafer ships
3FD/2FD ~ 2030 is when More Advanced SOI wafers ship to customers
Canon NIL tool should be shippable early. However, it will only be the six/eight cluster tool that does full production. Minimum speed for 28-nm metal pitch is officially 20 wafers per hour per cluster. True 14A, with 16-nm metal pitch is 2028+ target for Canon.

NIL goes has inverse speed from minimum pitch:
28nm MP = 20 WPH/c --- 2028+ --> 16nm MP = 20 WPH/c
40nm MP = 40 WPH/c --- 2028+ --> 28nm MP = 40 WPH/c
56nm MP = 60 WPH/c --- 2028+ --> 40nm MP = 80 WPH/c [Which 80WPH is the speed cap for a single cluster]

 

[DrMrLordX]

Actually, "to be Intel" ... maybe no?

News is nice and all, but for me feels like a double lovers suicide. At least they won't die alone, poor souls.

Nah Glofo doesn't even have foundry tech at Intel's level. 12FDX and 12nm+ are pretty much it, and I'm not even sure if they ship 12FDX yet? It's hard to tell.

A little birdie told me that Intel's israel fab is going to be abandoned soon.

GoFlo go get it!

Intel isn't going to leave anything of value (IP, etc) there for another competitor to snap up. Would be nice for GF if they did!

I was crazy, but trying to revive the MISHAP, I mean MIPS right now is a very bad idea...

It's kind of weird, but remember that MIPS isn't even pursuing their own custom uarch anymore anyway. They're just another RISC-V shop at this point.

 

[igor_kavinski]

Intel isn't going to leave anything of value (IP, etc) there for another competitor to snap up. Would be nice for GF if they did!

Intel badly needs the money. GloFlo could probably get the Saudis and Abu Dhabi Govt to pay. Intel shouldn't have a problem licensing their old tech to GloFlo. They are focused on 14A now.

 

[DZero]

Intel badly needs the money. GloFlo could probably get the Saudis and Abu Dhabi Govt to pay. Intel shouldn't have a problem licensing their old tech to GloFlo. They are focused on 14A now.

If Intel gives the 7nm to GloFo, it will be enough to put them back on track. With multiple patterning, would get to 5nm

 

[DrMrLordX]

If Intel gives the 7nm to GloFo, it will be enough to put them back on track. With multiple patterning, would get to 5nm

That would certainly be interesting. GF might become the next SMIC, only they'd be refining Intel DUV nodes rather than ones stolen from Samsung "working on their own innovations".

 

[DZero]

That would certainly be interesting. GF might become the next SMIC, only they'd be refining Intel DUV nodes rather than ones stolen from Samsung "working on their own innovations".

Talking about Samsung... makes me wonder if Samsung decides to sell the 10nm tech to SMIC in order to see if the chinese improves that. Also, their 4nm became mature and their 3nm is interestingly decent.

 

[DrMrLordX]

Talking about Samsung... makes me wonder if Samsung decides to sell the 10nm tech to SMIC in order to see if the chinese improves that. Also, their 4nm became mature and their 3nm is interestingly decent.

Why buy when they can borrow?

Also is Samsung finally out of the woods yet with SF3 nodes? I wasn't aware. It certainly looks like their SF4X node has some problems with it, or else AMD wouldn't have dumped it for TSMC N4C.

 

[LightningZ71]

From what I've read, Samsung doesn't have a commercially viable node post the "5nm" generation.

 

[igor_kavinski]

From what I've read, Samsung doesn't have a commercially viable node post the "5nm" generation.

What??? They can't even poach Intel foundry engineers???

Really must not want to invest I guess.

 

[Io Magnesso]

What??? They can't even poach Intel foundry engineers???

Really must not want to invest I guess.

Well, to be honest, the situation at Samsung Fab may be worse than Intel…

 

[LightningZ71]

There are plenty of unsubstantiated rumors out there.