BRN Discussion Ongoing

TechGirl

Founding Member
Not sure if this has been shared but it’s great to see Global Foundries, GM, Renesas, Oculi and Brainchip mentioned in the same article.

I particularly liked this:
“GF is also championing the computing sector, namely with BrainChip's Akida neuromorphic chip built on 22 nm fully depleted silicon-on-insulator (FD-SOI)technology.”



Bloody Brilliant article

Thanks Slade (y)

Donald Duck Money GIF
 
  • Like
  • Love
  • Fire
Reactions: 15 users

TopCat

Regular
Has anyone here ever done any research on Dolphin Designs. Partners with Global Foundries and has designs with FD-SOI at 22nm and claims mW power consumption for sensors at the VERY edge! Clients now exceeding 500!


Tiny RAPTOR is a complete Neural Processor solution to deploy AI at the very edge combining Software and Hardware approaches, resulting from more than three years of development, together with CEA-List in a joint laboratory.


CEA-Leti & Dolphin Design Report FD-SOI Breakthrough that Boosts Operating Frequency by 450% and Reduces Power Consumption by 30%​

The architecture enables reducing energy consumption of processors in 22nm FD-SOI technology by up to 30 percent and increasing the operating frequency up to 450 percent compared to a technique in which body biased technique is not used. It also improves the manufacturing yield.

About Dolphin Design

Headquartered in France, Dolphin Design, previously known as Dolphin Integration, is a semiconductor company employing 160 people, including 140 highly qualified engineers. They provide differentiating platform solutions built on state-of-the-art IPs and architectures, customized by unique system level utilities to deliver fast and secure ASICs, either designed by or for their clients. These platforms are available for various technological processes and optimized for Energy Efficient SoC Design. Alongside their clients, now exceeding 500 companies, they focus on human, inventive and long-term collaboration to enable them to bring products, powered by innovative and accessible integrated circuits that minimize environmental impact, to the hands of billions of people every day. In consumer markets including IoT, AI and 5G, and in high reliability markets, they unleash SoC designer creativity to deliver differentiation.

F04F8303-E1E6-4290-BBE6-45D59660E4F6.jpeg
D1C65107-DEAC-4D2B-B34F-094FE712D877.jpeg
 
Last edited:
  • Like
  • Fire
  • Thinking
Reactions: 13 users
Has anyone here ever done any research on Dolphin Designs. Partners with Global Foundries and has designs with FD-SOI at 22nm and claims mW power consumption for sensors at the VERY edge! Clients now exceeding 500!


Tiny RAPTOR is a complete Neural Processor solution to deploy AI at the very edge combining Software and Hardware approaches, resulting from more than three years of development, together with CEA-List in a joint laboratory.


CEA-Leti & Dolphin Design Report FD-SOI Breakthrough that Boosts Operating Frequency by 450% and Reduces Power Consumption by 30%​

The architecture enables reducing energy consumption of processors in 22nm FD-SOI technology by up to 30 percent and increasing the operating frequency up to 450 percent compared to a technique in which body biased technique is not used. It also improves the manufacturing yield.

About Dolphin Design

Headquartered in France, Dolphin Design, previously known as Dolphin Integration, is a semiconductor company employing 160 people, including 140 highly qualified engineers. They provide differentiating platform solutions built on state-of-the-art IPs and architectures, customized by unique system level utilities to deliver fast and secure ASICs, either designed by or for their clients. These platforms are available for various technological processes and optimized for Energy Efficient SoC Design. Alongside their clients, now exceeding 500 companies, they focus on human, inventive and long-term collaboration to enable them to bring products, powered by innovative and accessible integrated circuits that minimize environmental impact, to the hands of billions of people every day. In consumer markets including IoT, AI and 5G, and in high reliability markets, they unleash SoC designer creativity to deliver differentiation.

View attachment 31236 View attachment 31235
Going back a bit in the old memory banks to when Mr. Dinardo was CEO someone asked about Raptor and the response from Peter van der Made was it is just an accelerator.

The reason I mention this was in the time of Mr. Dinardo he was hammering away to shareholders and the market that AKIDA was not an accelerator but a processor, a tiny Ai computer on a chip.

When he departed Rob Telson moved the language back towards describing AKIDA as an accelerator and I suggested at the time that this change in language may have been driven by customers inability to understand just what AKIDA was all about whereas they had some understanding of accelerators.

By describing AKIDA as an accelerator it was easier to get through the customers front door where they could then be educated as to what AKIDA was and why it should be chosen over a simple accelerator.

My opinion only DYOR
FF

AKIDA BALLISTA
 
  • Like
  • Love
  • Fire
Reactions: 23 users

Steve10

Regular
Another company to keep an eye on is SiTime which is owned by MegaChips.


SiTime Transforms Precision Timing Market for Edge Networks​

DATA CENTER AND CO-LOCATIONNEWS
By AIT News Desk On Mar 17, 2022

SiTime Corporation , a market leader in precision timing, introduced the Elite X Super-TCXO for edge networks such as data centers, 5G front haul, connected cars and industrial IoT. In these applications, Elite X solves critical timing problems and enables delivery of new services.

“Our Elite X Super-TCXO shines by delivering 2x better stability and 30x higher reliability than quartz. Consequently, we believe that Elite X will now be the precision timing device of choice in edge equipment,” said Piyush Sevalia, SiTime.

“Twenty-two million autonomous vehicles on the road by 2025, the rapid adoption of 5G, and the continued growth of data centers will require a large buildout in outdoor, decentralized edge networks,” said Piyush Sevalia, executive vice president of marketing at SiTime. “These new edge devices and their precision timing heartbeat must perform reliably in the presence of extreme temperatures, thermal shock, and vibration. Here, our Elite X Super-TCXO shines by delivering 2x better stability and 30x higher reliability than quartz. Consequently, we believe that Elite X will now be the precision timing device of choice in edge equipment.”

To enable new time-sensitive services such as ADAS, virtual reality, and smart manufacturing, edge devices will need to deliver faster speeds with lower latency. Precision timing plays a crucial role in their success. Legacy timing devices such as quartz TCXOs and mini-OCXOs have well-known weaknesses such as sensitivity to thermal shock and vibration, stability degradation at high temperatures, and shorter operating life. By using silicon MEMS technology that is inherently more robust and reliable, Elite X overcomes these problems and enables the next generation of time-sensitive services everywhere.

In edge networks, additional specifications such as aging, warm-up time, and power consumption are important for overall performance. Even here, SiTime Elite X outperforms legacy timing devices, simplifies system design, and offers dependable, consistent performance.

With Elite X (±10-20 ppb stability), Emerald (±5 ppb), and Elite (±50-500 ppb), SiTime offers a broad portfolio of best-in-class TCXOs and OCXOs that meet various application requirements in communications, networking, automotive, and industrial IoT.

SiT5501 Elite X Super-TCXO Features and Benefits

The SiT5501 Elite X Super-TCXO includes the following features and benefits:

  • ±10 ppb frequency stability
  • -40°C to +105°C operating temperature. More applications require 105°C operation.
  • 110 mW typical power consumption
  • Small package, 7.0 mm x 5.0 mm
  • ±0.5 ppb/°C dF/dT (frequency slope), resistant to thermal shock and airflow
  • ±0.5 ppb daily aging
  • Any frequency from 1 to 60 MHz with up to 6 decimal places of accuracy


Blog​

JAN092018
Timing's Critical Role in the AI Revolution
Posted By: Mark Bajus
Small timing devices run a big part of our lives, and many times, we fail to realize it. I sat down with Aaron Partridge, the chief scientist at SiTime, to discuss his prognostications for the future of the electronics industry and the role that timing will play.

As you look ahead, what trends in electronics do you see?​

One giant trend that I see is the development of AI systems. Artificial intelligence will usher in the next wave of the industrial revolution. The first wave used steam power to mechanize production and travel in the 1780s and extended this with fossil fuel in the 1920s. The second used electronics and software to automate communications in the 1980s. Today, we need to process the data from our increasingly digitized world. AI organizes this data to augment our mental limits.
One can think of technology as extending our human limits. The first wave amplified our muscles, in that we can fly, drive, and move heavy objects. The second wave amplified our senses, in that we could hear people thousands of miles away or see minute objects. And the third wave is now amplifying our ability to think.

You use the present tense to talk about the next wave of the industrial revolution. Why is that?​

That’s the thing with technological revolutions. When we are in the midst of them, many of us don’t notice that change is happening. The technology slowly creeps into our lives. We won’t wake up one day and suddenly live in the world of the Jetsons. Right now, the personal assistants in our mobile devices, the self-driving car I just bought, they both use artificial intelligence.

What is timing’s role in AI?​

Timing will play a critical role in the communications infrastructure that AI leverages. When many people think about artificial intelligence, they think of robots, but it is about more than that. It’s also about processing massive amounts of data. So, AI relies on data centers to store information and on high-speed radio and fiber to transport the information – all of which relies on timing. Without high-performance oscillators, artificial intelligence systems wouldn’t be able to process the data reliably and without latency.

How do you see AI fitting into the evolution of timing?​

That’s a pretty big question. What did timing mean to the Romans? The Romans could precisely measure sunrise, midday, and sunset. Their sundials divided the day into 12 hours of daylight and 12 hours of nighttime, and the length of the hours depended on the length of the day. So, if Romans agreed to meet at 1:00, they were probably accurate to maybe 20 minutes.
Skip ahead a couple millennia to today. Our phones know what time it is to a microsecond. Our GPS receivers know the accurate time to within thirty billionths of a second. So, we’ve gone from 1,000 seconds to a millionth-of-a-second accuracy. That is a billion time more accurate.
And we are not done. The upcoming 5G phones will hold time at least 10x more accurately than 4G, and soon GPS receivers will keep track of time down to one billionth of a second. So, that is a trillion times more accurate than the Romans.

In addition to accuracy, what else will be needed for AI systems?​

People are now talking about “the internet of things.” All of those “things” will be networked, and networking requires timing. So, in that sense, we are dealing with the “internet of time-synchronized things.”

In a world of artificial intelligence and time-synchronized things, there will be a massive number of sensors everywhere, and they will need to be smaller, consume less power, and continue to perform reliably as they age. MEMS-based timing solutions, like the ones from SiTime, fit perfectly into that world.

 
  • Like
  • Thinking
Reactions: 15 users

toasty

Regular
Yep. You put used trays in it and when it's full it drives into the kitchen, dodging people, to be unloaded. Just couldn't resist taking a photo for the 1000 eyes!
There's one of these operating at Mugen House, a Japanese restaurant in Adelaide..................
 
  • Like
Reactions: 5 users

buena suerte :-)

BOB Bank of Brainchip
ANNOUNCEMENT ⏰⏰


$$$$$$$ BrainChip Introduces Second-Generation Akida Platform $$$$$$$


Drives extremely efficient and intelligent edge devices for Artificial Intelligence of Things (AIoT) solutions
• New generation of Akida allows designers and developers to incorporate features that were not possible before
• Adds capabilities that are critically needed in industrial, automotive, digital health, smart home and smart city applications


Sydney – 06 March 2023: BrainChip Holdings Ltd (ASX: BRN, OTCQX: BRCHF, ADR: BCHPY), the world’s first commercial producer of ultra-low power, fully digital, neuromorphic AI IP, today announced the launch of the second generation of its Akida™
platform that drives extremely efficient and intelligent edge devices for the Artificial Intelligence of Things (AIoT) solutions and services market that is forecast to be worth US$1T+ by 2030 according to an industry report published by Fortune Business Insights. Sean Hehir, BrainChip CEO said, “Our customers wanted us to enable expanded predictive intelligence, target tracking, object detection, scene segmentation, and advanced vision capabilities.
This new generation of Akida allows designers and developers to do things that were not possible before on an Edge device. By inferring and learning from raw sensor data, we take a substantial step toward a cloudless Edge AI experience. With this launch, we have significantly extended our competitive advantage in neuromorphic AI.” This hyper-efficient, yet powerful neural processing system, architected for embedded Edge AI applications, now adds efficient 8-bit processing to go with advanced capabilities such as time domain convolutions and vision transformer acceleration, for an unprecedented level of performance in sub-watt devices, taking them from perception towards cognition. The second-generation of Akida now includes Temporal Event Based Neural Nets (TENN) spatial-temporal convolutions that supercharge the processing of raw time-continuous streaming data, such as video analytics, target tracking, audio classification, analysis of health monitoring data such as heart rate and respiratory rate for vital signs prediction, Page | 2 and time series analytics used in forecasting, and predictive production line maintenance.

These capabilities are critically needed in industrial, automotive, digital health, smart home and smart city applications. The TENNs allow for radically simpler implementations by consuming raw data directly from sensors - drastically reducing model size and operations performed, while maintaining very high accuracy. This can shrink design cycles and dramatically lower the cost of development. Mr Hehir added, “The development of the second generation of Akida was strongly influenced by our customers’ feedback and driven by our extensive market engagement. We have recently expanded our sales organisation to become truly global and we are focused on executing more IP licence agreements and generating revenue growth over coming years. Another addition to the second generation of Akida is Vision Transformers (ViT) acceleration, a leading-edge neural network that has been shown to perform extremely well on various computer vision tasks, such as image classification, object detection, and semantic segmentation.

The Akida IP platform has a unique ability to learn on the device for continuous improvement and data-less customization that improves security and privacy. This, combined with the efficiency and performance available, enable very differentiated solutions that until now have not been possible. These include secure, small battery powered devices like hearing aids and wearable electronic devices such as watches, medical devices for monitoring vital signs, and consume only microwatts of power. This can scale up to HD-resolution vision solutions delivered through high-value, batteryoperated or fan-less devices enabling a wide variety of applications from surveillance systems to factory management and augmented reality to scale effectively. Akida’s software and tooling further simplifies the development and deployment of solutions and services with these features: • An efficient runtime engine that autonomously manages model accelerations completely transparent to the developer • MetaTF™ software that developers can use with their preferred framework, like TensorFlow/Keras, or development platform, like Edge Impulse, to easily develop, tune, and deploy AI solutions. • Supports all types of Convolutional Neural Networks (CNN), Deep Learning Networks (DNN), Vision Transformer Networks (ViT) as well as Spiking Neural Networks (SNNs), future-proofing designs as the models get more advanced. Akida comes with a Models Zoo and a burgeoning ecosystem of software, tools, and model vendors, as well as IP, SoC, foundry and system integrator partners. Page | 3 General availability will follow in Q3’ 2023.

A formal platform launch press release will take place on 7 March at 1:00am AEDT, 6 March at 6:00am US PST. This announcement is authorised for release by the BRN Board of Directors 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 neuromorphic principles to mimic the human brain, analysing 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 onchip AI, close to the sensor, is the future, for its customers’ products, as well as the planet. Explore the benefits of Essential AI at www.brainchip.com. Follow BrainChip on Twitter: https://www.twitter.com/BrainChip_inc Follow BrainChip on LinkedIn: https://www.linkedin.com/company/779200
 
Last edited:
  • Like
  • Fire
  • Love
Reactions: 113 users
BrainChip Introduces Second-Generation Akida Platform • Drives extremely efficient and intelligent edge devices for Artificial Intelligence of Things (AIoT) solutions • New generation of Akida allows designers and developers to incorporate features that were not possible before • Adds capabilities that are critically needed in industrial, automotive, digital health, smart home and smart city applications Sydney – 06 March 2023: BrainChip Holdings Ltd (ASX: BRN, OTCQX: BRCHF, ADR: BCHPY), the world’s first commercial producer of ultra-low power, fully digital, neuromorphic AI IP, today announced the launch of the second generation of its Akida™ platform that drives extremely efficient and intelligent edge devices for the Artificial Intelligence of Things (AIoT) solutions and services market that is forecast to be worth US$1T+ by 2030 according to an industry report published by Fortune Business Insights. Sean Hehir, BrainChip CEO said, “Our customers wanted us to enable expanded predictive intelligence, target tracking, object detection, scene segmentation, and advanced vision capabilities. This new generation of Akida allows designers and developers to do things that were not possible before on an Edge device. By inferring and learning from raw sensor data, we take a substantial step toward a cloudless Edge AI experience. With this launch, we have significantly extended our competitive advantage in neuromorphic AI.” This hyper-efficient, yet powerful neural processing system, architected for embedded Edge AI applications, now adds efficient 8-bit processing to go with advanced capabilities such as time domain convolutions and vision transformer acceleration, for an unprecedented level of performance in sub-watt devices, taking them from perception towards cognition. The second-generation of Akida now includes Temporal Event Based Neural Nets (TENN) spatial-temporal convolutions that supercharge the processing of raw time-continuous streaming data, such as video analytics, target tracking, audio classification, analysis of health monitoring data such as heart rate and respiratory rate for vital signs prediction, Page | 2 and time series analytics used in forecasting, and predictive production line maintenance. These capabilities are critically needed in industrial, automotive, digital health, smart home and smart city applications. The TENNs allow for radically simpler implementations by consuming raw data directly from sensors - drastically reducing model size and operations performed, while maintaining very high accuracy. This can shrink design cycles and dramatically lower the cost of development. Mr Hehir added, “The development of the second generation of Akida was strongly influenced by our customers’ feedback and driven by our extensive market engagement. We have recently expanded our sales organisation to become truly global and we are focused on executing more IP licence agreements and generating revenue growth over coming years. Another addition to the second generation of Akida is Vision Transformers (ViT) acceleration, a leading-edge neural network that has been shown to perform extremely well on various computer vision tasks, such as image classification, object detection, and semantic segmentation. The Akida IP platform has a unique ability to learn on the device for continuous improvement and data-less customization that improves security and privacy. This, combined with the efficiency and performance available, enable very differentiated solutions that until now have not been possible. These include secure, small battery powered devices like hearing aids and wearable electronic devices such as watches, medical devices for monitoring vital signs, and consume only microwatts of power. This can scale up to HD-resolution vision solutions delivered through high-value, batteryoperated or fan-less devices enabling a wide variety of applications from surveillance systems to factory management and augmented reality to scale effectively. Akida’s software and tooling further simplifies the development and deployment of solutions and services with these features: • An efficient runtime engine that autonomously manages model accelerations completely transparent to the developer • MetaTF™ software that developers can use with their preferred framework, like TensorFlow/Keras, or development platform, like Edge Impulse, to easily develop, tune, and deploy AI solutions. • Supports all types of Convolutional Neural Networks (CNN), Deep Learning Networks (DNN), Vision Transformer Networks (ViT) as well as Spiking Neural Networks (SNNs), future-proofing designs as the models get more advanced. Akida comes with a Models Zoo and a burgeoning ecosystem of software, tools, and model vendors, as well as IP, SoC, foundry and system integrator partners. Page | 3 General availability will follow in Q3’ 2023. A formal platform launch press release will take place on 7 March at 1:00am AEDT, 6 March at 6:00am US PST
 
  • Like
  • Love
  • Fire
Reactions: 84 users

Bravo

If ARM was an arm, BRN would be its biceps💪!
images.jpeg



..."the second generation of its Akida™ platform that drives extremely efficient and intelligent edge devices for the Artificial Intelligence of Things (AIoT) solutions and services market that is forecast to be worth US$1T+ by 2030".
 
  • Like
  • Fire
  • Love
Reactions: 55 users
Never before have so few delivered so much to so many - Apologies to Winston Churchill

This new generation AKIDA has taken Brainchip's technology lead out to at least the five years Peter van der Made predicted.

It has also added 8 bits to catch out Qualcomm, Intel, Nvidia and Samsung at the game they were trying to play.

It is already set to be made available by Edge Impulse so those 60,000 developers will be extremely busy.

It has introduced Transformers as Rob Telson alluded to in his podcast with Carnegie Mellon.

It has literally blown everyone else out of the water.

My opinion only of course so DYOR
FF

AKIDA BALLISTA
 
  • Like
  • Fire
  • Love
Reactions: 143 users

Boab

I wish I could paint like Vincent
 
  • Like
  • Love
  • Fire
Reactions: 13 users

AC/DC

Member
2nd Generation - when no-one else can offer a first generation competitor...
 
  • Like
  • Fire
  • Love
Reactions: 50 users

buena suerte :-)

BOB Bank of Brainchip
Never before have so few delivered so much to so many - Apologies to Winston Churchill

This new generation AKIDA has taken Brainchip's technology lead out to at least the five years Peter van der Made predicted.

It has also added 8 bits to catch out Qualcomm, Intel, Nvidia and Samsung at the game they were trying to play.

It is already set to be made available by Edge Impulse so those 60,000 developers will be extremely busy.

It has introduced Transformers as Rob Telson alluded to in his podcast with Carnegie Mellon.

It has literally blown everyone else out of the water.

My opinion only of course so DYOR
FF

AKIDA BALLISTA
This is MASSIVE NEWS ....... EXCITING!!!!!!! GO BRN and great to see a Price Sensitive announcement 👏👏👏🙏🙏🙏🚀🚀🚀🚀
 
Last edited:
  • Like
  • Love
  • Fire
Reactions: 64 users
D

Deleted member 118

Guest
 
  • Haha
  • Like
Reactions: 28 users

BaconLover

Founding Member
  • Haha
  • Like
  • Love
Reactions: 30 users

FJ-215

Regular
'This hyper-efficient, yet powerful neural processing system, architected for embedded
Edge AI applications, now adds efficient 8-bit processing to go with advanced
capabilities such as time domain convolutions and vision transformer acceleration, for an
unprecedented level of performance in sub-watt devices, taking them from perception
towards cognition.'

'Mr Hehir added, “The development of the second generation of Akida was strongly
influenced by our customers’ feedback
and driven by our extensive market engagement.'



Hi Dio,
Am I right in thinking that the AKD1500 doesn't have enough 'grunt' to run 8 bit efficiently. Is this possibly the AKD2000?

And what customers need 8 bit?
 
  • Like
  • Thinking
Reactions: 13 users

toasty

Regular
Be interesting to see how the market reacts. My guess is that they will generally not understand the significance.....no $$$ attached to the announcement.

Don't get me wrong. For those of us who follow this technology and the BRN story this is a breakthrough achievement. Congrats to Peter, Anil and the team!!!
 
  • Like
  • Fire
  • Love
Reactions: 31 users
  • Haha
  • Like
Reactions: 12 users
D

Deleted member 118

Guest
  • Haha
Reactions: 2 users
Think about this:

"This hyper-efficient, yet powerful neural processing system, architected for embedded Edge AI applications,

now adds efficient 8-bit processing to go with

advanced capabilities such as time domain convolutions and vision transformer acceleration,

for an unprecedented level of performance in sub-watt devices,

taking them from perception towards cognition.

The second-generation of Akida now includes Temporal Event Based Neural Nets (TENN)

spatial-temporal convolutions that supercharge the processing of raw time-continuous streaming data,

such as video analytics,
target tracking,
audio classification,
analysis of health monitoring data such as heart rate and respiratory rate for vital signs prediction, and

time series analytics used in forecasting, and predictive production line maintenance."

And there is more.

My opinion only DYOR
FF

AKIDA BALLISTA
 
  • Like
  • Love
  • Fire
Reactions: 56 users
Top Bottom