BRN Discussion Ongoing
- Thread starter TechGirl
- Start date
Does anyone know anything about Dimensity 9300 chip?? From Bloomberg - “Dimensity 9300 is the most powerful mobile system-on-chip" 2nd half of the story
AI Chip Boom Fuels Taiwan Firm’s 40% Rally, Beating Qualcomm and Peers
- Company’s Dimensity 9300 chip is viewed as a game changer
- Valuation relative to growth prospect is attractive, GAM says
By Charlotte Yang
27 November 2023 at 11:40 am AEST
A five-month rally in MediaTek Inc. looks set to extend as booming demand for smartphones and a promising new AI chip brighten the outlook.
The Taiwanese tech firm’s stock has soared almost 40% since end-June, outperforming a 2% advance in the Philadelphia Stock Exchange Semiconductor Index and a 7% rise in its US-based rival Qualcomm Inc. The gains have been fueled by robust appetite for mobile devices that utilize the company’s chips, especially in China.
The buzz surrounding MediaTek casts the spotlight on the growing competition between semiconductor firms that are exploiting the use of AI to grab a larger share of the smartphone business. Investors see the Taiwanese firm’s Dimensity 9300 chip as a game-changer that will help it to steal a march over Qualcomm, the current leader in the high-end mobile market.
The bullishness surrounding MediaTek’s shares is growing. Analysts have boosted estimates for the company’s earnings per share by 13% since the end of July while short interest on the stock has fallen from a June peak. They have also lifted its estimated price target by 27% since a July low.
MediaTek’s “valuation relative to its growth prospect into next year still looks attractive,” said Robert Mumford, an investment manager at GAM Hong Kong Ltd. “The handset cyclical recovery and MediaTek’s new system-on-chip — a cutting edge product released at the right time — plus potential AI product rollout, are all tailwinds.”
A lot of the optimism is centered around the Dimensity 9300 chip, which is a direct competitor of Qualcomm’s Snapdragon 8 Gen 3. Both firms are racing to unleash AI technology on devices, part of the so-called “edge AI” phenomenon that multiple Wall Street banks say will be a key investing theme heading into 2024.
“Dimensity 9300 is the most powerful mobile system-on-chip and has started to gain traction with Vivo’s new phone,” Morgan Stanley analysts including Charlie Chan and Daisy Dai wrote in a note. The bank expects MediaTek’s market share in the flagship smartphone segment to increase from 20% in 2023 to 30%-35% in 2024, or up from 13 million units to around 20 million.
The Dimensity 9300 chip is designed to address more complex workloads of new generative AI and gaming applications and its launch follows several earlier designs which failed to excite investors. Chinese smartphone Vivo X100, which was released this month, is the first model to use the chip, and this will be followed by the OPPO Find X7 series.
The stock is trading at about 17 times forward estimated earnings, roughly in line with its five-year average. With a market capitalization of over $47 billion, MediaTek ranks as the the island’s second-largest listed company, after only Taiwan Semiconductor Manufacturing Co Ltd.
Smartphone Boom
The company’s sales are expected to grow 14% this quarter from last year, snapping four straight quarters of year-on-year decline, according to data compiled by Bloomberg.
MediaTek said in its earnings call that its inventory has fallen for five straight quarters and the days of inventory reached a healthy level of 90 at the end of the third quarter. Mobile phone sales, which contributed about half of the firm’s revenue in the three months through September, gained 19% on quarter, thanks to restocking demand and new 4G and 5G launches, management said.
The rally in MediaTek’s shares also reflects the recovery in China’s smartphone industry, which registered double-digit growth in October. Phone sales in the world’s biggest mobile market were up 11% in the first four weeks of last month compared to the same period a year ago, according to Counterpoint Research data.
MediaTek, you don't know nutthin':
US2021334586A1 EDGE LEARNING DISPLAY DEVICE AND METHOD
An image processing circuit stores a training database and models in memory. The image processing circuit includes an attribute identification engine to identify an attribute from an input image according to a model stored in the memory. By enhancing the input image based on the identified attribute, a picture quality (PQ) engine in the image processing circuit generates an output image for display. The image processing circuit further includes a data collection module to generate a labeled image based on the input image labeled with the identified attribute, and to add the labeled image to the training database. A training engine in the image processing circuit then re-trains the model using the training database.
mediaTek uses MACs, not SNNs:
[0038] The CNN accelerator 612 includes hardware components specialized for accelerating neural network operations by convolutional operations, fully-connected operations, activation, pooling, normalization, element-wise mathematical computations, etc. In some embodiments, the CNN accelerator 612 includes multiple compute units and memory (e.g., Static Random Access Memory (SRAM)), where each compute unit further includes multipliers and adder circuits, among others, for performing mathematical operations such as multiply-and-accumulate (MAC) operations to accelerate the convolution, activation, pooling, normalization, and other neural network operations. The CNN accelerator 612 may perform fixed and floating-point neural network operations. In connection with the picture quality enhancement described herein, the CNN accelerator 612 may perform training and inference operations described in connection with FIG. 1.
Well they can do training:
An image processing circuit stores a training database and models in memory. The image processing circuit includes an attribute identification engine to identify an attribute from an input image according to a model stored in the memory. By enhancing the input image based on the identified attribute, a picture quality (PQ) engine in the image processing circuit generates an output image for display. The image processing circuit further includes a data collection module to generate a labeled image based on the input image labeled with the identified attribute, and to add the labeled image to the training database. A training engine in the image processing circuit then re-trains the model using the training database.
... but it will cut your milage in half.
CrabmansFriend
Regular
Although way over my paygrade,
it seems like there is some serious effort made in improving devlopment for neuromorphic systems cross hardware platform/software (easier to port between and compare). Allthough Brainchip is mentioned in the PDF (linked below), they are currently not part of the 7 listed hardware platforms.
Neuromorphic Intermediate Representation: A Unified Instruction Set for Interoperable Brain-Inspired Computing
github.com
it seems like there is some serious effort made in improving devlopment for neuromorphic systems cross hardware platform/software (easier to port between and compare). Allthough Brainchip is mentioned in the PDF (linked below), they are currently not part of the 7 listed hardware platforms.
Neuromorphic Intermediate Representation: A Unified Instruction Set for Interoperable Brain-Inspired Computing
sources:
GitHub - neuromorphs/NIR: Neuromorphic Intermediate Representation reference implementation
Neuromorphic Intermediate Representation reference implementation - neuromorphs/NIR
github.com
Fullmoonfever
Top 20
Get the feeling MB still working out their blends of tech innovations to meet overall processing requirements, cloud connections as necessary, latency and retaining the data cybersecurity.
Given the ChatGPT use in the Hey Merc function is driven by LLM, I suspect the need for us to bed down the LLM processing ability too for ease of integration.
community.nasscom.in
November 17, 2023
Re-engineering tomorrow's hyperconnected autonomous world
Author: Rajaganapathy Rajendran, Vice President - Cloud and Connected Services, MBRDI
Let’s begin with what a future exchange with Mercedes’ voice assistant ‘Hey Mercedes’ would look like.
You: Hey Mercedes!
Hey Mercedes: Hello! Good morning! You have an upcoming appointment at 9 AM and I have mapped the fastest route to your destination. You should arrive by 8:50 AM. Your favourite playlist is being brought up for you to enjoy the ride. Enjoy your ride!
….
Hey Mercedes: Due to traffic congestion, there is going to be a delay and as per the new estimate, your arrival time has been updated to 9:10 AM. Other participants in the meeting have been notified of your delay.
…
Hey Mercedes: Would you like me to connect you to your meeting via voice call?
Please be mindful of your driving.
This is a glimpse of what the hyperconnected future offers but before we understand the nuances, let’s take a short walk down memory lane.
Since the advent of the Internet of Things (IoT), the automotive industry has revolutionised and moulded vehicles to operate as more than just a means of transportation. We have taken it a step further by marrying IoT and Autonomous Vehicles to redefine mobility and luxury.
Advancing with ADAS
To decipher the mercurial trajectory of the automotive industry, we must go back to the start. When the concept of an Advanced Driver Assistance System (ADAS) first seeped into the industry in the 1970s, it was about diminishing road fatalities through systems that attempted to negate human error. What started with Anti-lock Braking Systems (ABS) has today evolved to using Radar and LiDAR to evaluate optimal deceleration, even without human intervention.
A changing paradigm
While the intention is to remain uncompromising with safety, enhancing the driving and passenger experience is what the future firmly points towards. IoT, cloud connectivity and data sharing are all growing trends, and increasing the capabilities of the connected car is a collective vision all manufacturers share.
The first step towards achieving this has been transitioning from Electronic Control Units (ECU) to High-Performance Computers (HPC). HPCs are crucial in order to handle the advanced architecture modern vehicles possess.
5G at the core
Trends are increasingly moving towards making vehicles more autonomous. The connected car and 5G will undeniably have a crucial play in this transformation.
But we aren’t quite there yet. For autonomous vehicles to boast better reliability, connectivity and safety, extensive data from numerous sensors must be processed. Thus, the need for 5G, considering 4G hasn’t been able to handle the load and efficiently process the data.
5G is further essential with cloud computing, simplifying processing times and latency. This paves the path for effortless vehicle-to-everything (V2X) communications. V2X remains a vital cog in car-to-car and car-to-infrastructure exchanges.
Implications of data collection
Extensive data exchange necessitates extensive security measures. The future of mobility is powered by data collection and its efficient and intelligent non-intrusive use.
The strive towards positive impact isn’t bereft of its share of risks. Opening a communication channel with cloud devices exposes the ever-prevalent risk of cyber breaches. The more we innovate towards advanced autonomous levels of driving, the greater the risks of hackers introducing harmful elements into the vehicle architecture and inducing catastrophic consequences.
Now more than ever before, there is a need to fortify our cybersecurity infrastructure. This is where circling back to the start of this piece will give us a better understanding. Mercedes-Benz’s recent integration of ChatGPT and Hey Mercedes reaffirmed the necessity of having a watertight system to safeguard the privacy of our customers while allowing them the benefits of a truly delightful in-cabin experience.
The industry remains focused on navigating towards a hyperconnected, and hyper-personalised space.
Essentially, safety, convenience and luxury must be at the core of this journey. Innovating responsibly and using technology sensibly is the way forward, as we enter a new dawn with a ‘third living space’.
Rajaganapathy Rajendran
Given the ChatGPT use in the Hey Merc function is driven by LLM, I suspect the need for us to bed down the LLM processing ability too for ease of integration.
Re-engineering tomorrow's hyperconnected autonomous world
Re-engineering tomorrow's hyperconnected autonomous world Author: Rajaganapathy Rajendran, Vice President - Cloud and Connected Services, MBRDI Let’s begin with what a future exchange with Mercedes’ voice assistant ‘Hey Mercedes’ would look like. You: Hey Mercedes! Hey Mercedes: Hello...
RE-ENGINEERING TOMORROW'S HYPERCONNECTED AUTONOMOUS WORLD
Mercedes-Benz Research and Development India
@MBRDINovember 17, 2023
Re-engineering tomorrow's hyperconnected autonomous world
Author: Rajaganapathy Rajendran, Vice President - Cloud and Connected Services, MBRDI
Let’s begin with what a future exchange with Mercedes’ voice assistant ‘Hey Mercedes’ would look like.
You: Hey Mercedes!
Hey Mercedes: Hello! Good morning! You have an upcoming appointment at 9 AM and I have mapped the fastest route to your destination. You should arrive by 8:50 AM. Your favourite playlist is being brought up for you to enjoy the ride. Enjoy your ride!
….
Hey Mercedes: Due to traffic congestion, there is going to be a delay and as per the new estimate, your arrival time has been updated to 9:10 AM. Other participants in the meeting have been notified of your delay.
…
Hey Mercedes: Would you like me to connect you to your meeting via voice call?
Please be mindful of your driving.
This is a glimpse of what the hyperconnected future offers but before we understand the nuances, let’s take a short walk down memory lane.
Since the advent of the Internet of Things (IoT), the automotive industry has revolutionised and moulded vehicles to operate as more than just a means of transportation. We have taken it a step further by marrying IoT and Autonomous Vehicles to redefine mobility and luxury.
Advancing with ADAS
To decipher the mercurial trajectory of the automotive industry, we must go back to the start. When the concept of an Advanced Driver Assistance System (ADAS) first seeped into the industry in the 1970s, it was about diminishing road fatalities through systems that attempted to negate human error. What started with Anti-lock Braking Systems (ABS) has today evolved to using Radar and LiDAR to evaluate optimal deceleration, even without human intervention.
A changing paradigm
While the intention is to remain uncompromising with safety, enhancing the driving and passenger experience is what the future firmly points towards. IoT, cloud connectivity and data sharing are all growing trends, and increasing the capabilities of the connected car is a collective vision all manufacturers share.
The first step towards achieving this has been transitioning from Electronic Control Units (ECU) to High-Performance Computers (HPC). HPCs are crucial in order to handle the advanced architecture modern vehicles possess.
5G at the core
Trends are increasingly moving towards making vehicles more autonomous. The connected car and 5G will undeniably have a crucial play in this transformation.
But we aren’t quite there yet. For autonomous vehicles to boast better reliability, connectivity and safety, extensive data from numerous sensors must be processed. Thus, the need for 5G, considering 4G hasn’t been able to handle the load and efficiently process the data.
5G is further essential with cloud computing, simplifying processing times and latency. This paves the path for effortless vehicle-to-everything (V2X) communications. V2X remains a vital cog in car-to-car and car-to-infrastructure exchanges.
Implications of data collection
Extensive data exchange necessitates extensive security measures. The future of mobility is powered by data collection and its efficient and intelligent non-intrusive use.
The strive towards positive impact isn’t bereft of its share of risks. Opening a communication channel with cloud devices exposes the ever-prevalent risk of cyber breaches. The more we innovate towards advanced autonomous levels of driving, the greater the risks of hackers introducing harmful elements into the vehicle architecture and inducing catastrophic consequences.
Now more than ever before, there is a need to fortify our cybersecurity infrastructure. This is where circling back to the start of this piece will give us a better understanding. Mercedes-Benz’s recent integration of ChatGPT and Hey Mercedes reaffirmed the necessity of having a watertight system to safeguard the privacy of our customers while allowing them the benefits of a truly delightful in-cabin experience.
The industry remains focused on navigating towards a hyperconnected, and hyper-personalised space.
Essentially, safety, convenience and luxury must be at the core of this journey. Innovating responsibly and using technology sensibly is the way forward, as we enter a new dawn with a ‘third living space’.
Rajaganapathy Rajendran
AARONASX
Holding onto what I've got
I think it was on my end on the computer, tried again on mobile and fixed it, i thinkg
buena suerte :-)
BOB Bank of Brainchip
Last edited:
Frangipani
Top 20
Stumbled across this short video just now and honestly don’t know why I even bothered to watch it, as I actually suspected it to be another one of those boring AI-generated clips mushrooming everywhere these days… Much to my surprise, though, this slide came up right at the beginning:
… and even though the rest of the video doesn’t live up to the expectations this intro raises (delving into PVDM’s insights in less than three minutes sounded massively overpromising anyway
), as it covers very little except for some basic statements about the future of AI, the importance for companies to embrace it and a brief explanation of what neuromorphic processing is about (however, without once referring to Akida)…… it is still free publicity for Brainchip! And many a mickle makes a muckle. Right, James?! It cuts both ways, you know…
The person behind this YouTube channel is a Japan-based Pakistani lady called Hinza Asif (https://nftstudio24.com/hinza-asif/) who had recently been nominated for a Women in Tech APAC Award. I’d imagine that at least some of her 3000 plus followers would now google “Peter van der Made” as well as “neuromorphic” to find out more about The Future of AI…
Bravo
If ARM was an arm, BRN would be its biceps💪!
Can someone please call Rob Telson ASAP to get him to speak to someone at Phillips Medical Systems?
I just stumbled upon this Power Point Preso which shows Phillips are working with IMEC on the denoising of medical XRAYs using SNN's (SENECA) and they're hoping to apply this same technology can be applied in WIDE RANGE OF CLINICAL APPLICATION AREAS.
I believe that SENECA is still only at research stage which could leave the door wide open for us to say hello to Phillips. It says in this research paper below, published 23 June 23 that " The SENECA platform and the tools used in this project are available for academic research upon request."
Then there's also this Linkedin response from 5 months ago where the researcher at IMEC says they can't compare SENECA with BrainChip which he differentiates as being "commercial neuromorphic architecture".
www.ncbi.nlm.nih.gov
I just stumbled upon this Power Point Preso which shows Phillips are working with IMEC on the denoising of medical XRAYs using SNN's (SENECA) and they're hoping to apply this same technology can be applied in WIDE RANGE OF CLINICAL APPLICATION AREAS.
I believe that SENECA is still only at research stage which could leave the door wide open for us to say hello to Phillips. It says in this research paper below, published 23 June 23 that " The SENECA platform and the tools used in this project are available for academic research upon request."
Then there's also this Linkedin response from 5 months ago where the researcher at IMEC says they can't compare SENECA with BrainChip which he differentiates as being "commercial neuromorphic architecture".
SENECA: building a fully digital neuromorphic processor, design trade-offs and challenges - PMC
Neuromorphic processors aim to emulate the biological principles of the brain to achieve high efficiency with low power consumption. However, the lack of flexibility in most neuromorphic architecture designs results in significant performance loss ...
Last edited:
AARONASX
Holding onto what I've got
Totally agree with you there
Buy side is strong and very little resistance on the sell
Good Morning Chippers ,
Thinking shes about to go stratospheric.
.
Esq.
Thinking shes about to go stratospheric.
.Esq.
Last edited:
Pom down under
Top 20
buena suerte :-)
BOB Bank of Brainchip
Yep she's steadily moving in a positive trajectoryGood Morning Chippers ,
Thinking shes about to go stratospheric.
Esq.
...Low volume but solid!Better looking Buy/Sell ratio also
HarryCool1
Regular
Good Morning Chippers ,
Thinking shes about to go stratospheric.
Esq.
DingoBorat
Slim
Similar threads
