BRN Discussion Ongoing

[Diogenese]

View attachment 58950

Looks like the task has breached its assigned limits, and effective and immediate communication with the work team has terminated.

 

[Diogenese]

Tenstorrent and MosChip Partner on High Performant RISC-V Design

Tenstorrent and MosChip Technologies announced that they are partnering on design for Tenstorrent’s cutting-edge RISC-V solutions.

timestech.in

“Partnering with MosChip enables us to design the strongest RISC-V solution we can to serve ourselves, our partners, and our customers alike.”
View attachment 58998 View attachment 58999

Jim Keller of Tenstorrent is not a believer in "that spikey thingy".

 

[IloveLamp]

Jim Keller of Tenstorrent is not a believer in "that spikey thingy".

But MosChip certainly is, Jim must be yet to come out of the spike closet I'd say......

Give it time, he'll come around.

Opinion only, Dyor.

 

[Diogenese]

But MosChip certainly is, Jim must be yet to come out of the spike closet I'd say......

Opinion only, Dyor.

Sounds uncomfortable.

 

[IloveLamp]

Sounds uncomfortable.

Oh it will be,..... for his .........pride

 

[Frank Zappa]

My feeling is that a company that goes to all the trouble to create their own OS will want to own their own chip. ie We should expect M-B to announce the license deal with us at the time they announce that we are the third chip partner.

As far as the $s go I think Note 2[q] of the Annual Report says something like we'll be happy to defer the payments for licensees to fit in with their revenue stream from using Akida.

If that is correct, I cannot see any reason why we should have to bend over that far in the future

 

[Frank Zappa]

This should be forbidden. It is not only very childish ... I don't play with you anymore and take all my pictures with me ... but also as stupid as one can be. Grow up and behave like an adult! You and Fact Finder have not only given information but also taken some from other people. So there is no reason to take some away. I cannot believe how someone can behave like that.

For heavens sake , do not involve FF at the same standing as that nitwit . It shows that you do not have very good judgment or respect for our messiah.

 

[Iseki]

If that is correct, I cannot see any reason why we should have to bend over that far in the future

Well not even NVIDIA make their own chips, so M-B would have to go off to a chip designer and tell them that they want something that can run Akida IP and talk directly to an NVIDIA chip, all using MB.OS, so a chip that no one else can use!, and we need to tape it out and queue up at somewhere like TSMC and test the bjesus out of it, and then take it to the M-B BoD and hope they give the go ahead! And then and only then make that fateful call to Sean, and utter these words "Sean, baby, please use the following Order Number MB-Neuromorphic-001"

 

[Diogenese]

Ok but you actually all know what I mean with “they need certificates”? Yes? Because I have the feeling maybe you don’t know what I mean with this

Some years ago PvdM said BRN would leave the qualification of Akida to the end users.

 

[cosors]

You mean something like this.

I just wonder how many days it would take me to delete all my 3123 posts especially in the small Talga group

 

[Rach2512]

New partnership to offer next-gen tech to students

The University of Western Australia has announced a cutting-edge collaboration with tech company BrainChip to equip students with the tools and resources they n

www.uwa.edu.au

Sorry if already posted.
From UWA website.

 

[7für7]

Yes. That’s why I say. Until they (the customers) do not have a certification, they can not go into mass production

 

[mkm109]

NVIDIA Drive Thor uses updated ARM Poseidon AE cores and AKIDA is compatible with Arm's product families. Just a thought bubble.


View attachment 58985



This part of an article dated September 2022 caught my eye.
View attachment 58986

Arm Announces Neoverse V2 and E2: The Next Generation of Arm Server CPU Cores

www.anandtech.com

Remember also NVIDIA is needed for driving the video images on the vast number of displays. Akida can focus on the AI side

 

[Fullmoonfever]

I don't recall this posted from mid last year?

Seems that this institute (pretty big research team and connected) have been playing with Akida Gen 1 amongst the other usual suspects of NVIDIA, Intel, Xilinx.

Now they just need Gen 2.

Is a presentation and some challenges via Think Project but I haven't dug into as just on ph and other stuff going on.

Snipped some pages relevant to us.

Maybe @Diogenese can pull some of the technical aspects out?

Full doc HERE

IN2P3

The National Institute of Nuclear and Particle Physics (IN2P3) performs research in the field of the 'two infinites' namely the infinitely large, with the study of cosmology and astroparticles, and the infinitely small, with nuclear physics and the physics of elementary particles.

www.in2p3.cnrs.fr

IN2P3​

The National Institute of Nuclear and Particle Physics (IN2P3) performs research in the field of the 'two infinites' namely the infinitely large, with the study of cosmology and astroparticles, and the infinitely small, with nuclear physics and the physics of elementary particles. IN2P3 is a leading international actor in these disciplines which often require very large-scale transnational facilities and its scientists are at the forefront of major discoveries, such as the Higgs boson, neutrino oscillations and gravitational waves.The Institute is also having major contributions to the development of related applied technologies, such as particle accelerators and detectors, mainly in the fields of health, energy and the environment.

 

[Diogenese]

I don't recall this posted from mid last year?

Seems that this institute (pretty big research team and connected) have been playing with Akida Gen 1 amongst the other usual suspects of NVIDIA, Intel, Xilinx.

Now they just need Gen 2.

Is a presentation and some challenges via Think Project but I haven't dug into as just on ph and other stuff going on.

Snipped some pages relevant to us.

Maybe @Diogenese can pull some of the technical aspects out?

Full doc HERE

IN2P3

The National Institute of Nuclear and Particle Physics (IN2P3) performs research in the field of the 'two infinites' namely the infinitely large, with the study of cosmology and astroparticles, and the infinitely small, with nuclear physics and the physics of elementary particles.

www.in2p3.cnrs.fr

IN2P3​

The National Institute of Nuclear and Particle Physics (IN2P3) performs research in the field of the 'two infinites' namely the infinitely large, with the study of cosmology and astroparticles, and the infinitely small, with nuclear physics and the physics of elementary particles. IN2P3 is a leading international actor in these disciplines which often require very large-scale transnational facilities and its scientists are at the forefront of major discoveries, such as the Higgs boson, neutrino oscillations and gravitational waves.The Institute is also having major contributions to the development of related applied technologies, such as particle accelerators and detectors, mainly in the fields of health, energy and the environment.



View attachment 59012 View attachment 59013 View attachment 59014 View attachment 59015 View attachment 59016

Hi Fmf,

This is the premier nuclear research organization in France - the greatest concentration of baguettes, berets and boffins on earth.

IN2P3 is the French umbrella organization for nuclear and particle physics research and works with CERN where they found Higgs.

On slide 32, Akida got excellent marks for CH1, Ch3, and Ch5, but only a B- for Ch 4 (circles) (see Slide 14). I'm guessing that these are radiation patterns from particle collision sensors - something like a DVS Geiger counter. One possible reason for this that I can think of is that the model/weights CNN2SNN conversion was dodgy, although the boffins attribute it to the 4-bit quantization, which seems wrong to me. Properly trained (model/weights) Akida could pick a circle from the largest geometric menagerie.

I wonder if there was a BRN engineer helping the boffins with the PCIe board?

In any case, they have still to savour the delights of Akida 2.

FN: The file size for challenge 4 is much larger than for the others (Slide 36). These files would be for Intel and they are not converted to SNN.

 

[IloveLamp]

https://eu1.hubs.ly/H083Nh00

Tata Elxsi: A leader in tech services | Tata Elxsi posted on the topic | LinkedIn

Tata Elxsi, led by CEO & MD Manoj Raghavan, has been delivering cutting-edge tech services to global OEMs, all the while maintaining a steady growth…

www.linkedin.com

 

[Tothemoon24]

Arm Announces New Automotive Technologies to Accelerate Development of AI-enabled Vehicles by up to Two Years

Arm today announced new automotive technologies and a roadmap of compute subsystems to enable faster time to market for AI enabled vehicles.

newsroom.arm.com

Arm Announces New Automotive Technologies to Accelerate Development of AI-enabled Vehicles by up to Two Years​

By Dipti Vachani, SVP and GM of the Automotive Line of Business, Arm
Artificial Intelligence (AI)AutomotiveSoftware
Share

News Highlights​

• New safety-enabled Arm Automotive Enhanced (AE) processors will bring leading-edge Armv9 and server-class performance to AI-driven use cases
• Future Arm Compute Subsystems (CSS) for Automotive will further reduce development time and cost, providing maximum flexibility for high-performance automotive systems
• For the first time, the Arm ecosystem can develop software on virtual prototyping solutions for the automotive market ahead of physical silicon being available, accelerating development cycles by up to two years

The automotive market is going through an unprecedented transformation with demand for more autonomy, increasingly advanced user experiences and the trend toward electrification creating an explosion of software and AI. Vehicle electronics are becoming more complex and to deliver at pace, a fundamental rethink of product development is needed. Today, in a series of industry firsts, Arm and our ecosystem are unveiling the latest Arm Automotive Enhanced (AE) processors alongside new virtual platforms, all made available to the industry from day one to accelerate automotive development cycles by up to two years.
Next-generation AE processors bring Armv9 and server-class performance to automotive for the first time
For the first time, we are bringing Armv9-based technologies to automotive, enabling the industry to take advantage of the AI, security and virtualization capabilities that this latest generation of the Arm architecture delivers. To meet the growing performance demands of today’s vehicles, we are leveraging our leadership in the infrastructure market by bringing server-class Neoverse technology to automotive, along with new Armv9-based Cortex-A products for scalability. The full line-up of products we’re announcing today includes:

• Arm Neoverse V3AE: Bringing Neoverse technology to the automotive sector for the first time, delivering server-class performance for AI-accelerated autonomous and ADAS workloads
• Arm’s first v9-based Cortex-A processors purpose-built for automotive:
  • Arm Cortex-A720AE: Delivering industry-leading sustained performance with SoC design flexibility for a broad range of software-defined vehicle (SDV) applications
  • Arm Cortex-A520AE: Providing leading power efficiency with functional safety features to scale across automotive use cases
• Arm Cortex-R82AE: The highest-ever performing real-time processor for functional safety which delivers 64-bit computing to real-time processing for the very first time
• Arm Mali-C720AE: A configurable ISP optimized for the most demanding computer and human vision use cases
• A range of configurable system IP to enable the Arm silicon ecosystem to deliver scalable, high-performance automotive SoCs

Read more about this new technology, which is already being adopted by leading players, including Marvell, MediaTek, NVIDIA, NXP, Renesas, Telechips, Texas Instruments, and others, in this blog post.
The road ahead: Arm Compute Subsystems (CSS) for Automotive
We are not stopping there. Looking ahead, we see that systems are only going to get more complex, the need for safety will become even greater, and we’ll need to bring all of these compute components together for our partners, standardizing for consistency where it matters. Arm CSS for Automotive will deliver pre-integrated and validated configurations of Arm AE IP optimized for performance, power, and area using a leading-edge foundry process. Our first CSS for automotive is expected to be delivered in 2025.
Virtual prototypes will enable ecosystem to leverage Arm AE IP for software development ahead of silicon
Traditional automotive development cycles are linear: First processor IP is delivered, then silicon development begins, and only when hardware becomes available around two years later, can software developers start. Today we’re changing that with a new approach: virtual prototyping on industry-leading, latest-generation Arm AE IP, which allows software developers to start before the physical silicon is available. We’re partnering with industry leaders including Amazon Web Services (AWS), Cadence, Corellium, Siemens and others to deliver these virtual platforms and cloud solutions, which allow for earlier, more seamless development across the full software stack in collaboration with partners including Autoware Foundation, BlackBerry QNX, Elektrobit, Kernkonzept, LeddarTech, Mapbox, Sensory, Tata Technologies, TIER IV, Vector and more. Read more about the new virtual platforms here, and more about our collaborations at the foundational and application software layer here.
Today’s announcement marks a critical step in enabling the automotive ecosystem by unlocking new opportunities for developers and accelerating time to market for leading OEMs, automotive Tier 1 suppliers, silicon partners, and software providers that are all building the future of software-defined vehicles, on Arm.

Additional Resources:​

• Supporting Quotes from the Arm Automotive Ecosystem
• Arm’s Broadest Ever Automotive Enhanced IP Portfolio Designed for the Future of Computing in Vehicles
• Virtual Platforms from Arm and Partners Available Now to Accelerate and Transform Automotive Development
• Ecosystem Collaborations Bring Full Stack Software Solutions to Develop Leading-edge Automotive Applications From Day One
• Functional Safety Compute for the Software-defined Vehicle

Arm introduces new safety-enabled AE processors | Arm posted on the topic | LinkedIn

🆕 We're bringing Armv9 and Neoverse-based technologies to automotive for the very first time. With our new safety-enabled AE processors, industry leaders can… | 28 comments on LinkedIn

www.linkedin.com

 

[Diogenese]

Arm Announces New Automotive Technologies to Accelerate Development of AI-enabled Vehicles by up to Two Years

Arm today announced new automotive technologies and a roadmap of compute subsystems to enable faster time to market for AI enabled vehicles.

newsroom.arm.com

Arm Announces New Automotive Technologies to Accelerate Development of AI-enabled Vehicles by up to Two Years​

By Dipti Vachani, SVP and GM of the Automotive Line of Business, Arm
Artificial Intelligence (AI)AutomotiveSoftware
Share

News Highlights​

• New safety-enabled Arm Automotive Enhanced (AE) processors will bring leading-edge Armv9 and server-class performance to AI-driven use cases
• Future Arm Compute Subsystems (CSS) for Automotive will further reduce development time and cost, providing maximum flexibility for high-performance automotive systems
• For the first time, the Arm ecosystem can develop software on virtual prototyping solutions for the automotive market ahead of physical silicon being available, accelerating development cycles by up to two years

The automotive market is going through an unprecedented transformation with demand for more autonomy, increasingly advanced user experiences and the trend toward electrification creating an explosion of software and AI. Vehicle electronics are becoming more complex and to deliver at pace, a fundamental rethink of product development is needed. Today, in a series of industry firsts, Arm and our ecosystem are unveiling the latest Arm Automotive Enhanced (AE) processors alongside new virtual platforms, all made available to the industry from day one to accelerate automotive development cycles by up to two years.
Next-generation AE processors bring Armv9 and server-class performance to automotive for the first time
For the first time, we are bringing Armv9-based technologies to automotive, enabling the industry to take advantage of the AI, security and virtualization capabilities that this latest generation of the Arm architecture delivers. To meet the growing performance demands of today’s vehicles, we are leveraging our leadership in the infrastructure market by bringing server-class Neoverse technology to automotive, along with new Armv9-based Cortex-A products for scalability. The full line-up of products we’re announcing today includes:

• Arm Neoverse V3AE: Bringing Neoverse technology to the automotive sector for the first time, delivering server-class performance for AI-accelerated autonomous and ADAS workloads
• Arm’s first v9-based Cortex-A processors purpose-built for automotive:
  • Arm Cortex-A720AE: Delivering industry-leading sustained performance with SoC design flexibility for a broad range of software-defined vehicle (SDV) applications
  • Arm Cortex-A520AE: Providing leading power efficiency with functional safety features to scale across automotive use cases
• Arm Cortex-R82AE: The highest-ever performing real-time processor for functional safety which delivers 64-bit computing to real-time processing for the very first time
• Arm Mali-C720AE: A configurable ISP optimized for the most demanding computer and human vision use cases
• A range of configurable system IP to enable the Arm silicon ecosystem to deliver scalable, high-performance automotive SoCs

Read more about this new technology, which is already being adopted by leading players, including Marvell, MediaTek, NVIDIA, NXP, Renesas, Telechips, Texas Instruments, and others, in this blog post.
The road ahead: Arm Compute Subsystems (CSS) for Automotive
We are not stopping there. Looking ahead, we see that systems are only going to get more complex, the need for safety will become even greater, and we’ll need to bring all of these compute components together for our partners, standardizing for consistency where it matters. Arm CSS for Automotive will deliver pre-integrated and validated configurations of Arm AE IP optimized for performance, power, and area using a leading-edge foundry process. Our first CSS for automotive is expected to be delivered in 2025.
Virtual prototypes will enable ecosystem to leverage Arm AE IP for software development ahead of silicon
Traditional automotive development cycles are linear: First processor IP is delivered, then silicon development begins, and only when hardware becomes available around two years later, can software developers start. Today we’re changing that with a new approach: virtual prototyping on industry-leading, latest-generation Arm AE IP, which allows software developers to start before the physical silicon is available. We’re partnering with industry leaders including Amazon Web Services (AWS), Cadence, Corellium, Siemens and others to deliver these virtual platforms and cloud solutions, which allow for earlier, more seamless development across the full software stack in collaboration with partners including Autoware Foundation, BlackBerry QNX, Elektrobit, Kernkonzept, LeddarTech, Mapbox, Sensory, Tata Technologies, TIER IV, Vector and more. Read more about the new virtual platforms here, and more about our collaborations at the foundational and application software layer here.
Today’s announcement marks a critical step in enabling the automotive ecosystem by unlocking new opportunities for developers and accelerating time to market for leading OEMs, automotive Tier 1 suppliers, silicon partners, and software providers that are all building the future of software-defined vehicles, on Arm.

Additional Resources:​

• Supporting Quotes from the Arm Automotive Ecosystem
• Arm’s Broadest Ever Automotive Enhanced IP Portfolio Designed for the Future of Computing in Vehicles
• Virtual Platforms from Arm and Partners Available Now to Accelerate and Transform Automotive Development
• Ecosystem Collaborations Bring Full Stack Software Solutions to Develop Leading-edge Automotive Applications From Day One
• Functional Safety Compute for the Software-defined Vehicle

Arm introduces new safety-enabled AE processors | Arm posted on the topic | LinkedIn

🆕 We're bringing Armv9 and Neoverse-based technologies to automotive for the very first time. With our new safety-enabled AE processors, industry leaders can… | 28 comments on LinkedIn

www.linkedin.com

Hi TTM,

This bit is encouraging:

• A range of configurable system IP to enable the Arm silicon ecosystem to deliver scalable, high-performance automotive SoCs

 

[cosors]

I wish you a good start in the day!
View attachment 58943

That looks different to me now as a non-TA. @Fullmoonfever ?
No more hammer down...?

 

[Tothemoon24]