BRN Discussion Ongoing

[Diogenese]

Uploaded a couple of hours ago with Doug McLelland via Frontgrade channel.

Think Frontgrade got month wrong .. should be April. What is it with tech companies and simple errors

The RISC-V in Space Workshop, organized by the European Space Agency in collaboration with RISC-V International, was held on May 2–3, 2025, in Gothenburg, Sweden.

Hi FMF,

This is the link to the conference.

https://www.risc-v.space/wp/

Some interesting speakers and sponsors.

Some really terrific information packed into a few minutes.

This is exciting news. A RISC-V + Akida device is on its way:

It was interesting to see that the NPUs can be reused to improve processing/number of NPUs per layer with negligible time penalty even though weights need to be reloaded each frame. The more NPUs per layer, the faster the data to the layer can be processed.

 

[Diogenese]

Steve talks about Edge Impulse. I wonder when this was recorded.

Neuromorphic Computing Delivers Low-Power AI

BrainChip’s Akida platform offers low-power, neuromorphic machine-learning capabilities.

www.electronicdesign.com

Some e-shorthand notes:

Prophesee and Edge Impulse are discussed - Edge Ai box + Prophesee DVS plug and play.

Akida 2 upgraded visual transformers improved accuracy

State space models improved accuracy, smaller model size, lower power computation

transformer based models transition to SSM

Akida 1 chip as learning tool for S/w developers on PCIe - M.2 box has fan for CPU

MetaTF freely available - Edge Impulse now doing TENNs models. Training CNN2SNN - State space = RNN performance.

On device learning

TENNs (SSM) real time time series data reduce parameters 1/3 model size SoTA accuracy > CNN

SSM reduce computation 10x to 100x reducing the number of multiplies. Jensen Wang GPU calculations explode by square SSM is linear.

Open source model is bad for privacy. - proprietary LLM models retain IP

Edge Impulse training - use GPU for training.

FPGA platform can run edge LLM 25 NPUs

Felxible scalable platform + subscription for model.

New product categories - Transformer models slow to reach edge - Transformer LLM run hot slow at edge.

 

[Diogenese]

Hi FMF,

This is the link to the conference.

https://www.risc-v.space/wp/

Some interesting speakers and sponsors.

Some really terrific information packed into a few minutes.

This is exciting news. A RISC-V + Akida device is on its way:

View attachment 83008


It was interesting to see that the NPUs can be reused to improve processing/number of NPUs per layer with negligible time penalty even though weights need to be reloaded each frame. The more NPUs per layer, the faster the data to the layer can be processed.

View attachment 83007

The GRAIN RISC-V + Akida will not be limited to space applications.

Will this be the first RISC-V+SNN chip? (Where is SiFive?)

This could be a first to market product.

The RISC-V community has 24 premier members, 172 strategic members (including FG and SiFive), and 212 community members:
https://riscv.org/members/

We may get some good prospects from this - floodgates/hockeysticks ... ?

I get that old 2018 anticipation when I learned PvdM was about to do the first ever US presentation, but this time it's different

 

[Fullmoonfever]

Official news release from about a hour ago to the Andes link up @Frangipani posted the other day.

Now we know where the M.2 carda are also going.

BrainChip's Game-Changing AI Chip Delivers 10x Efficiency Boost Through RISC-V Partnership

New integration enables 0.7 TOPS performance at just 250mW power draw. Discover how this breakthrough partnership transforms edge AI computing. See specs now.

www.stocktitan.net

BrainChip Extends RISC-V Reach with Andes Technology Integration​

04/23/2025 - 12:00 PM
LAGUNA HILLS, Calif.--(BUSINESS WIRE)-- BrainChip Holdings Ltd (ASX: BRN, OTCQX: BRCHF, ADR: BCHPY), the world’s first commercial producer of ultra-low power, fully digital, event-based, brain-inspired AI, today announced the integration of its NPUs with RISC-V cores from Andes Technology, the industry leading provider of RISC-V embedded cores. The companies will demonstrate BrainChip’s Akida™ AKD1500 on Andes’ QiLai Voyager Board and AndesCore™ AX45MP 64-bit multicore CPU IP at Andes RISC-V Con 2025 in San Jose, Calif. April 29 and in Hsinchu, Taiwan June 10.

The AKD1500 demonstrates the benefits of Akida’s pure digital, extremely energy-efficient, event-based AI computation for at-sensor or sensor-balanced solutions for AI, application processors, automotive electronics and security markets. The QiLai SoC and the Voyager development board further accelerates the development and porting of large RISC-V applications. Integrating BrainChip’s Akida technology and Andes’ high-performance QiLai SoC and Voyager development board provides a highly efficient solution for integrated edge AI compute and further expands the RISC-V ecosystem.

The BrainChip AKD1500 device is integrated into the Voyager development board using an M.2 card form factor. It delivers over 0.7 TOPS of event-based computing while consuming less than 250mW, achieving performance comparable to conventional CNN processing using 3–10× less compute. This demonstrates a cost and power-efficient path for integrating RISC-V SoCs, operating at a fraction of the power required by traditional AI accelerators. Akida is an event-based technology that is inherently lower power than conventional neural network accelerators, providing energy efficiency with high performance for partners to deliver AI solutions previously not possible on battery-operated or fan-less embedded, edge devices.

The Andes QiLai SoC chip incorporates a high-performance quad-core RISC-V AX45MP cluster. The AndesCore AX45MP is a superscalar, multicore design featuring a shared Level-2 cache, a coherence manager and a Memory Management Unit (MMU) to support Linux-based applications. Equipped with IOCP (I/O Coherency Port) interface, the AX45MP enables external hardware DMA to interact directly with the cache/memory subsystem, facilitating seamless communication between the AX45MP and high-speed modules like NPUs, GPUs and Gigabit Ethernet. Built on TSMC's 7nm process technology, the AX45MP achieves clock speeds of up to 2.2 GHz on QiLai SoC. With higher Specint2006 performance than Cortex A55, high-clock frequency and multi-core Linux capabilities, the AX45MP has been very popular as Linux AP on various applications.

“We consistently search for ways to grow adoption of RISC-V solutions across an ever-increasing number of use cases,” said Dr. Charlie Su, CTO and President at Andes Technology. “By working with BrainChip and integrating the Akida IP into our QiLai Voyager platform, we better enable developers with opportunities to develop RISC-V-based solutions optimized for edge AI workloads and capabilities.”

“As a pioneer in ultra-low power AI processing, our Akida technology integrates with RISC-V to enable efficient, intelligent compute at the edge,” said Steve Brightfield, CMO at BrainChip. “Showcasing these attributes through the integration of the AKD1500 with Andes QiLai platform at RISC-V Con 2025, ideally demonstrates how these solutions can be used in production SoCs to provide next-generation breakthroughs.”

Andes RISC-V Con 2025 brings together developers and technologists from around the world to discuss advancements in RISC-V technology, exchange ideas and collaborate on the future of open computing architectures. BrainChip will showcase its latest innovations in ultra-low power AI solutions, demonstrating how its technology enhances edge computing applications. Attendees are invited to visit the BrainChip booth and attend the presentation to learn more about the role of AI acceleration in RISC-V-based designs. Interested parties can visit https://bit.ly/brainchip-andes to register.

About Andes Technology

As a Founding Premier member of RISC-V International and a leader in commercial CPU IP, Andes Technology (TWSE: 6533; SIN: US03420C2089; ISIN: US03420C1099) is driving the global adoption of RISC-V. Andes’ extensive RISC-V Processor IP portfolio spans from ultra-efficient 32-bit CPUs to high-performance 64-bit Out-of-Order multiprocessor coherent clusters. With advanced vector processing, DSP capabilities, the powerful Andes Automated Custom Extension (ACE) framework, end-to-end AI hardware/software stack, ISO 26262 certification with full compliance, and a robust software ecosystem, Andes unlocks the full potential of RISC-V, empowering customers to accelerate innovation across AI, automotive, communications, consumer electronics, data centers, and mobile devices. Over 16 billion Andes-powered SoCs are driving innovations globally. Discover more at www.andestech.com and connect with Andes on LinkedIn, X (formerly Twitter), Bilibili and YouTube.

About BrainChip Holdings Ltd (ASX: BRN, OTCQX: BRCHF, ADR: BCHPY)

BrainChip is the worldwide leader in Edge AI on-chip processing and learning. The company’s first-to-market, fully digital, event-based AI processor, AkidaTM, uses principles that mimic the human brain, analyzing only essential sensor inputs at the point of acquisition, processing data with unparalleled efficiency, precision, and economy of energy. Akida uniquely enables Edge learning local to the chip, independent of the cloud, dramatically reducing latency while improving privacy and data security. Akida Neural processor IP, which can be integrated into SoCs on any process technology, has shown substantial benefits on today’s workloads and networks, and offers a platform for developers to create, tune and run their models using standard AI workflows like Tensorflow/Keras. In enabling effective Edge compute to be universally deployable across real world applications such as connected cars, consumer electronics, and industrial IoT, BrainChip is proving that on-chip AI, close to the sensor, is the future, for its customers’ products, as well as the planet. Explore the benefits of Akida at www.brainchip.com.

 

[Fullmoonfever]

The GRAIN RISC-V + Akida will not be limited to space applications.

Will this be the first RISC-V+SNN chip? (Where is SiFive?)

This could be a first to market product.

The RISC-V community has 24 premier members, 172 strategic members (including FG and SiFive), and 212 community members:
https://riscv.org/members/

We may get some good prospects from this - floodgates/hockeysticks ... ?

I get that old 2018 anticipation when I learned PvdM was about to do the first ever US presentation, but this time it's different

Wonder if Akida IP will be part of the FG development supported by ESA to go into their 7nm RISC-V?

From a Dec article I recalled.

Gaisler to build 7nm RISC-V chip for space

Frontgrade Gaisler in Sweden has signed a deal with ESA to build a 7nm rad hard chip for space systems using the RISC-V architecture.

www.eenewseurope.com

Gaisler to build 7nm RISC-V chip for space​

Business news | December 16, 2024
By Nick Flaherty
RISC-V MPUs/MCUs AI space
Cette publication existe aussi en Français

---

Frontgrade Gaisler in Sweden has signed a deal with the European Space Agency (ESA) to build a 7nm rad hard chip for space systems using the RISC-V architecture.


The EEE Space Component Sovereignty for Europe programme will see Frontgrade Gaisler working with Belgian process technology expert imec and IMST in Germany on high-performance microprocessors, advanced semiconductor libraries, and high-speed memory interfaces.

Frontgrade was formerly Cobham Advanced Electronic Solutions (CAES) Space Systems until it was bought by private equity firm Veritas Capital in 2023 and includes the Cobham Gaisler space business.

The consortium’s initial focus is to establish radiation-hardened libraries and Intellectual Property (IP) cores that will serve as the foundation for highly reliable and efficient chips down to 7nm. This will be used for a RISC-V microprocessor prototype will be developed and tested for performance and radiation.


The project will incorporate additional state-of-the-art technologies from adjacent developments and partners, including high-speed serial interfaces, die-to-die interconnect, and System-in-Package concepts.

Now called Frontgrade Gaisler, the company has also licensed the Akida neuromorphic AI IP from Brainchip for a space-grade, fault-tolerant system-on-chip. NASA has also been working with the the Brainchip IP for space systems.

This commercial license agreement paves the path forward for AI chips to be deployed in space. The Akida neuromorphic computing solution was selected based upon providing real time data stream processing with resiliency, autonomy and accuracy at the lowest power, mass and volume. The European Space Agency (ESA) has been leading the efforts to deploy neuromorphic computing.

• ST, CAES team on octacore RISC-V space chip
• 16 core RISC-V chip for space designs

“While many fields no doubt can benefit from neuromorphic computing, one essential area is computer vision applications where current FPGA or GPU technology does not bring satisfactory results for edge deployments when considering mass, volume and power constraints,” said Laurent Hili, microelectronics and data handling engineer at the ESA.

“This programme has demonstrated the superiority of neuromorphic technology through several on-going activities with satellite primes and IP/components suppliers such as Frontgrade Gaisler and BrainChip.”

Gaisler is one of Europe’s leading supplier of space chips, with the NOEL RISC-V processor and the LEON SPARC_based processor.

“ESA is proud to support this groundbreaking initiative, which represents a critical step toward European sovereignty in advanced semiconductor technologies for space,” said Boris Glass, Technical Officer at ESA. “By investing in Ultra Deep Sub-Micron processes such as 7nm FinFET CMOS technology, we are ensuring that Europe remains at the forefront of space innovation and autonomy, securing the technology necessary for next-generation space exploration and satellite constellations. This collaboration with Frontgrade Gaisler and its industry partners is essential to meet the growing demands of the space sector and strengthen Europe’s capabilities in the global arena.”

In a future project, this microprocessor prototype will be advanced toward full functionality, production, and qualification, providing Europe with autonomous and highly competitive space computing capabilities, facilitating advanced AI and Edge computing to meet the demands of next-generation satellite constellations and deep-space missions.

“Frontgrade Gaisler has decades of experience supplying the space sector with advanced semiconductor products, which lends itself well to the work of EEE Space Component Sovereignty for Europe,” said Sandi Habinc, General Manager at Frontgrade Gaisler. “Through this program, we’re leveraging our expertise – along with the other participants – to advance deep submicron technology and to strengthen our position in this industry.”

“For the last 20 years, Gaisler’s space-grade microprocessors have been successfully deployed to every planet of our solar system in missions from ESA as well as most other space agencies. Together with BrainChip, we evaluated the Akida IP and decided that the next step of licensing it would beneficially augment our future space processors with neuromorphic AI.”

“This collaboration with Frontgrade Gaisler to license Akida IP for implementation into space SoCs represents an important step in satisfying the market demand for space-based AI deployments, turning into reality what once was considered unattainable,” said Sean Hehir, CEO of BrainChip. “We are pleased to expand on our trusted relationship with Frontgrade as they push the boundaries of space computing.”

 

[Diogenese]

Official news release from about a hour ago to the Andes link up @Frangipani posted the other day.

Now we know where the M.2 carda are also going.

BrainChip's Game-Changing AI Chip Delivers 10x Efficiency Boost Through RISC-V Partnership

New integration enables 0.7 TOPS performance at just 250mW power draw. Discover how this breakthrough partnership transforms edge AI computing. See specs now.

www.stocktitan.net

BrainChip Extends RISC-V Reach with Andes Technology Integration​

04/23/2025 - 12:00 PM
LAGUNA HILLS, Calif.--(BUSINESS WIRE)-- BrainChip Holdings Ltd (ASX: BRN, OTCQX: BRCHF, ADR: BCHPY), the world’s first commercial producer of ultra-low power, fully digital, event-based, brain-inspired AI, today announced the integration of its NPUs with RISC-V cores from Andes Technology, the industry leading provider of RISC-V embedded cores. The companies will demonstrate BrainChip’s Akida™ AKD1500 on Andes’ QiLai Voyager Board and AndesCore™ AX45MP 64-bit multicore CPU IP at Andes RISC-V Con 2025 in San Jose, Calif. April 29 and in Hsinchu, Taiwan June 10.

The AKD1500 demonstrates the benefits of Akida’s pure digital, extremely energy-efficient, event-based AI computation for at-sensor or sensor-balanced solutions for AI, application processors, automotive electronics and security markets. The QiLai SoC and the Voyager development board further accelerates the development and porting of large RISC-V applications. Integrating BrainChip’s Akida technology and Andes’ high-performance QiLai SoC and Voyager development board provides a highly efficient solution for integrated edge AI compute and further expands the RISC-V ecosystem.

The BrainChip AKD1500 device is integrated into the Voyager development board using an M.2 card form factor. It delivers over 0.7 TOPS of event-based computing while consuming less than 250mW, achieving performance comparable to conventional CNN processing using 3–10× less compute. This demonstrates a cost and power-efficient path for integrating RISC-V SoCs, operating at a fraction of the power required by traditional AI accelerators. Akida is an event-based technology that is inherently lower power than conventional neural network accelerators, providing energy efficiency with high performance for partners to deliver AI solutions previously not possible on battery-operated or fan-less embedded, edge devices.

The Andes QiLai SoC chip incorporates a high-performance quad-core RISC-V AX45MP cluster. The AndesCore AX45MP is a superscalar, multicore design featuring a shared Level-2 cache, a coherence manager and a Memory Management Unit (MMU) to support Linux-based applications. Equipped with IOCP (I/O Coherency Port) interface, the AX45MP enables external hardware DMA to interact directly with the cache/memory subsystem, facilitating seamless communication between the AX45MP and high-speed modules like NPUs, GPUs and Gigabit Ethernet. Built on TSMC's 7nm process technology, the AX45MP achieves clock speeds of up to 2.2 GHz on QiLai SoC. With higher Specint2006 performance than Cortex A55, high-clock frequency and multi-core Linux capabilities, the AX45MP has been very popular as Linux AP on various applications.

“We consistently search for ways to grow adoption of RISC-V solutions across an ever-increasing number of use cases,” said Dr. Charlie Su, CTO and President at Andes Technology. “By working with BrainChip and integrating the Akida IP into our QiLai Voyager platform, we better enable developers with opportunities to develop RISC-V-based solutions optimized for edge AI workloads and capabilities.”

“As a pioneer in ultra-low power AI processing, our Akida technology integrates with RISC-V to enable efficient, intelligent compute at the edge,” said Steve Brightfield, CMO at BrainChip. “Showcasing these attributes through the integration of the AKD1500 with Andes QiLai platform at RISC-V Con 2025, ideally demonstrates how these solutions can be used in production SoCs to provide next-generation breakthroughs.”

Andes RISC-V Con 2025 brings together developers and technologists from around the world to discuss advancements in RISC-V technology, exchange ideas and collaborate on the future of open computing architectures. BrainChip will showcase its latest innovations in ultra-low power AI solutions, demonstrating how its technology enhances edge computing applications. Attendees are invited to visit the BrainChip booth and attend the presentation to learn more about the role of AI acceleration in RISC-V-based designs. Interested parties can visit https://bit.ly/brainchip-andes to register.

About Andes Technology

As a Founding Premier member of RISC-V International and a leader in commercial CPU IP, Andes Technology (TWSE: 6533; SIN: US03420C2089; ISIN: US03420C1099) is driving the global adoption of RISC-V. Andes’ extensive RISC-V Processor IP portfolio spans from ultra-efficient 32-bit CPUs to high-performance 64-bit Out-of-Order multiprocessor coherent clusters. With advanced vector processing, DSP capabilities, the powerful Andes Automated Custom Extension (ACE) framework, end-to-end AI hardware/software stack, ISO 26262 certification with full compliance, and a robust software ecosystem, Andes unlocks the full potential of RISC-V, empowering customers to accelerate innovation across AI, automotive, communications, consumer electronics, data centers, and mobile devices. Over 16 billion Andes-powered SoCs are driving innovations globally. Discover more at www.andestech.com and connect with Andes on LinkedIn, X (formerly Twitter), Bilibili and YouTube.

About BrainChip Holdings Ltd (ASX: BRN, OTCQX: BRCHF, ADR: BCHPY)

BrainChip is the worldwide leader in Edge AI on-chip processing and learning. The company’s first-to-market, fully digital, event-based AI processor, AkidaTM, uses principles that mimic the human brain, analyzing only essential sensor inputs at the point of acquisition, processing data with unparalleled efficiency, precision, and economy of energy. Akida uniquely enables Edge learning local to the chip, independent of the cloud, dramatically reducing latency while improving privacy and data security. Akida Neural processor IP, which can be integrated into SoCs on any process technology, has shown substantial benefits on today’s workloads and networks, and offers a platform for developers to create, tune and run their models using standard AI workflows like Tensorflow/Keras. In enabling effective Edge compute to be universally deployable across real world applications such as connected cars, consumer electronics, and industrial IoT, BrainChip is proving that on-chip AI, close to the sensor, is the future, for its customers’ products, as well as the planet. Explore the benefits of Akida at www.brainchip.com.

Wonder if Akida IP will be part of the FG development supported by ESA to go into their 7nm RISC-V?

From a Dec article I recalled.

Gaisler to build 7nm RISC-V chip for space

Frontgrade Gaisler in Sweden has signed a deal with ESA to build a 7nm rad hard chip for space systems using the RISC-V architecture.

www.eenewseurope.com

Gaisler to build 7nm RISC-V chip for space​

Business news | December 16, 2024
By Nick Flaherty
RISC-V MPUs/MCUs AI space
Cette publication existe aussi en Français

---

Frontgrade Gaisler in Sweden has signed a deal with the European Space Agency (ESA) to build a 7nm rad hard chip for space systems using the RISC-V architecture.


The EEE Space Component Sovereignty for Europe programme will see Frontgrade Gaisler working with Belgian process technology expert imec and IMST in Germany on high-performance microprocessors, advanced semiconductor libraries, and high-speed memory interfaces.

Frontgrade was formerly Cobham Advanced Electronic Solutions (CAES) Space Systems until it was bought by private equity firm Veritas Capital in 2023 and includes the Cobham Gaisler space business.

The consortium’s initial focus is to establish radiation-hardened libraries and Intellectual Property (IP) cores that will serve as the foundation for highly reliable and efficient chips down to 7nm. This will be used for a RISC-V microprocessor prototype will be developed and tested for performance and radiation.


The project will incorporate additional state-of-the-art technologies from adjacent developments and partners, including high-speed serial interfaces, die-to-die interconnect, and System-in-Package concepts.

Now called Frontgrade Gaisler, the company has also licensed the Akida neuromorphic AI IP from Brainchip for a space-grade, fault-tolerant system-on-chip. NASA has also been working with the the Brainchip IP for space systems.

This commercial license agreement paves the path forward for AI chips to be deployed in space. The Akida neuromorphic computing solution was selected based upon providing real time data stream processing with resiliency, autonomy and accuracy at the lowest power, mass and volume. The European Space Agency (ESA) has been leading the efforts to deploy neuromorphic computing.

• ST, CAES team on octacore RISC-V space chip
• 16 core RISC-V chip for space designs

“While many fields no doubt can benefit from neuromorphic computing, one essential area is computer vision applications where current FPGA or GPU technology does not bring satisfactory results for edge deployments when considering mass, volume and power constraints,” said Laurent Hili, microelectronics and data handling engineer at the ESA.

“This programme has demonstrated the superiority of neuromorphic technology through several on-going activities with satellite primes and IP/components suppliers such as Frontgrade Gaisler and BrainChip.”

Gaisler is one of Europe’s leading supplier of space chips, with the NOEL RISC-V processor and the LEON SPARC_based processor.

“ESA is proud to support this groundbreaking initiative, which represents a critical step toward European sovereignty in advanced semiconductor technologies for space,” said Boris Glass, Technical Officer at ESA. “By investing in Ultra Deep Sub-Micron processes such as 7nm FinFET CMOS technology, we are ensuring that Europe remains at the forefront of space innovation and autonomy, securing the technology necessary for next-generation space exploration and satellite constellations. This collaboration with Frontgrade Gaisler and its industry partners is essential to meet the growing demands of the space sector and strengthen Europe’s capabilities in the global arena.”

In a future project, this microprocessor prototype will be advanced toward full functionality, production, and qualification, providing Europe with autonomous and highly competitive space computing capabilities, facilitating advanced AI and Edge computing to meet the demands of next-generation satellite constellations and deep-space missions.

“Frontgrade Gaisler has decades of experience supplying the space sector with advanced semiconductor products, which lends itself well to the work of EEE Space Component Sovereignty for Europe,” said Sandi Habinc, General Manager at Frontgrade Gaisler. “Through this program, we’re leveraging our expertise – along with the other participants – to advance deep submicron technology and to strengthen our position in this industry.”

“For the last 20 years, Gaisler’s space-grade microprocessors have been successfully deployed to every planet of our solar system in missions from ESA as well as most other space agencies. Together with BrainChip, we evaluated the Akida IP and decided that the next step of licensing it would beneficially augment our future space processors with neuromorphic AI.”

“This collaboration with Frontgrade Gaisler to license Akida IP for implementation into space SoCs represents an important step in satisfying the market demand for space-based AI deployments, turning into reality what once was considered unattainable,” said Sean Hehir, CEO of BrainChip. “We are pleased to expand on our trusted relationship with Frontgrade as they push the boundaries of space computing.”

We were told quite some time ago when we learned about SiFive that RISC-V presented a good opportunity for Akida.

Now we can add Andes and Frontgrade Glaiser.

Given that all the RISC-V processor makers are in fairly early stage development compared with CPU?GPU makers, the possibility of broad adoption of Akida for RISC-V looks promising.