BRN Discussion Ongoing

[Diogenese]

Google Pixel 7 Unblur function?????

Why not?

 

[alwaysgreen]

Why not?

It would be amazing but I haven't seen anything that would allude to Google having incorporated Akida in their chip (Tensor 2).

That being said, the photo unblur feature is specific to Pixel 7 (which has Googles Tensor 2 chip as opposed to last years phone which has the Tensor 1 chip). So maybe Akida was implemented specifically in their new chip. I bought one in October so it will show as revenue in the half year report if it's the case.

 

[Fact Finder]

It s Japanese

Thanks @Andi85

So my view is that if I was required to make an objective and logical prediction as to which mobile phone AKIDA will find its way into via MegaChips then Sony Xperia is the prime candidate.

Prophesee two lense sensor may make an appearance at the same time.

My opinion only DYOR
FF

AKIDA BALLISTA

 

[alwaysgreen]

Why not?

See below re Googles Tensor 2 chip. Forgive my lack of technical know how but it is a 4nm chip.My understanding is that Akida doesn't scale down to 4nm? Happy to be corrected.

Google Tensor G2: All the details on the Pixel 7's custom chipset

Google is hitting the custom silicon game hard! 🔥

www.androidpolice.com

​

Google Tensor G2 specifications​

For those that want a table, here you go:

Dev board code name	Cloudripper
Model number	GS201, Tensor G2
Cores	2x super-big ARM Cortex-X1, 2x big A78, 4x small Cortex-A55
GPU	Mali-G710
Manufacturing node	4nm Samsung PLP
Modem	Samsung Exynos 5300 5G

The Tensor G2 is made by Samsung on its 4nm node using panel-level packaging. This is a complicated way of saying the chips are carved out of a square wafer rather than a round one, reducing waste. This likely doesn't have much impact on the chip's performance in actual devices, but it's nifty and might reduce costs. Plus, it's potentially useful when we're still in the middle of a chip shortage.


The Tensor G2 keeps the 2+2+4 core cluster configuration that the original Tensor GS101 used, with two "super-big" cores, two more typical big cores, and four small cores. One thing that changes across generations is the frequency and one small tweak to the big clusters. The A76 cluster is replaced by an A78 cluster that's 100MHz faster at 2.35GHz. The other components remain the same, though. The X1 cluster has been bumped up by 50MHz, which gives it a frequency of 2.85GHz. This translates to a 10% to 15% better result in Geekbench, though you will be hard-pressed to notice much of this difference in real life.


Google has significantly upgraded the GPU, though. The Pixel 7 and 7 Pro are switching to the Mali-G710 GPU rather than the G78. That provides about 20% better performance and efficiency. The new GPU also helps the onboard machine-learning-focused TPU, giving it an up to 35% boost in applicable processes. The TPU is also seeing an upgrade.

The G2 is again paired with a Samsung-made modem, this time around, the Exynos S5300 5G. Mobile connectivity was poor on the Pixel 6 series and one of the biggest gripes many owners had with it. Based on initial reports from Pixel 7 owners, the situation is greatly improved with the new modem.


Overall, this small upgrade compared to the first-gen Tensor might be disappointing on paper, but it could make a lot of sense in the performance-to-power usage ratio. Newer processors are found to improve performance at the cost of energy consumption, so sticking with the older generation might leave more room for better efficiency. It also helps that Google has experience with this setup for a whole generation, making it simpler to optimize the system further. This is somewhat reminiscent of the company sticking with the same camera for multiple generations of Pixel phones, improving how the software interacts with the hardware with each iteration.

 

[Diogenese]

See below re Googles Tensor 2 chip. Forgive my lack of technical know how but it is a 4nm chip.My understanding is that Akida doesn't scale down to 4nm? Happy to be corrected.

Google Tensor G2: All the details on the Pixel 7's custom chipset

Google is hitting the custom silicon game hard! 🔥

www.androidpolice.com

​

Google Tensor G2 specifications​

For those that want a table, here you go:

Dev board code name	Cloudripper
Model number	GS201, Tensor G2
Cores	2x super-big ARM Cortex-X1, 2x big A78, 4x small Cortex-A55
GPU	Mali-G710
Manufacturing node	4nm Samsung PLP
Modem	Samsung Exynos 5300 5G

The Tensor G2 is made by Samsung on its 4nm node using panel-level packaging. This is a complicated way of saying the chips are carved out of a square wafer rather than a round one, reducing waste. This likely doesn't have much impact on the chip's performance in actual devices, but it's nifty and might reduce costs. Plus, it's potentially useful when we're still in the middle of a chip shortage.


The Tensor G2 keeps the 2+2+4 core cluster configuration that the original Tensor GS101 used, with two "super-big" cores, two more typical big cores, and four small cores. One thing that changes across generations is the frequency and one small tweak to the big clusters. The A76 cluster is replaced by an A78 cluster that's 100MHz faster at 2.35GHz. The other components remain the same, though. The X1 cluster has been bumped up by 50MHz, which gives it a frequency of 2.85GHz. This translates to a 10% to 15% better result in Geekbench, though you will be hard-pressed to notice much of this difference in real life.


Google has significantly upgraded the GPU, though. The Pixel 7 and 7 Pro are switching to the Mali-G710 GPU rather than the G78. That provides about 20% better performance and efficiency. The new GPU also helps the onboard machine-learning-focused TPU, giving it an up to 35% boost in applicable processes. The TPU is also seeing an upgrade.

The G2 is again paired with a Samsung-made modem, this time around, the Exynos S5300 5G. Mobile connectivity was poor on the Pixel 6 series and one of the biggest gripes many owners had with it. Based on initial reports from Pixel 7 owners, the situation is greatly improved with the new modem.


Overall, this small upgrade compared to the first-gen Tensor might be disappointing on paper, but it could make a lot of sense in the performance-to-power usage ratio. Newer processors are found to improve performance at the cost of energy consumption, so sticking with the older generation might leave more room for better efficiency. It also helps that Google has experience with this setup for a whole generation, making it simpler to optimize the system further. This is somewhat reminiscent of the company sticking with the same camera for multiple generations of Pixel phones, improving how the software interacts with the hardware with each iteration.

Thanks ag,

My "Why not?" was really aspirational rather than based on any detailed information.

There is nothing stopping Akida being made in 4 nm aside from cost. A couple of years ago 4nm was still an embryonic tech, and quite risky, when there was no reason to adopt the latest tech for Akida. 28 nm was the proven reliable tech, so Anil chose it to be sure that no complications arose from the manufacturing process, as we were running on the smell of an oily rag and flying by the seat of our pants (is that why they called it the joy stick?) at the time. Also 28 nm is more radiation proof than 4 nm, so it would have been better for our NASA relations.

But back to google,
https://www.zdnet.com/article/how-to-use-photo-unblur-on-the-google-pixel-7-series/
The latest Google Photos feature, Photo Unblur, takes to artificial intelligence to scan previously-captured images and applies a sharpening filter that magically "unblurs" the subject(s).

This is quite different from the Sony/Prophesee system which combines a DVS image with a shuttered camera image to eliminate blur on the fly.

So google uses a post-capture software to unblur, while Sony/Prophesee use an on-the-fly hardware system.

 

[JK200SX]

See below re Googles Tensor 2 chip. Forgive my lack of technical know how but it is a 4nm chip.My understanding is that Akida doesn't scale down to 4nm? Happy to be corrected.

Google Tensor G2: All the details on the Pixel 7's custom chipset

Google is hitting the custom silicon game hard! 🔥

www.androidpolice.com

​

Google Tensor G2 specifications​

For those that want a table, here you go:

Dev board code name	Cloudripper
Model number	GS201, Tensor G2
Cores	2x super-big ARM Cortex-X1, 2x big A78, 4x small Cortex-A55
GPU	Mali-G710
Manufacturing node	4nm Samsung PLP
Modem	Samsung Exynos 5300 5G

The Tensor G2 is made by Samsung on its 4nm node using panel-level packaging. This is a complicated way of saying the chips are carved out of a square wafer rather than a round one, reducing waste. This likely doesn't have much impact on the chip's performance in actual devices, but it's nifty and might reduce costs. Plus, it's potentially useful when we're still in the middle of a chip shortage.


The Tensor G2 keeps the 2+2+4 core cluster configuration that the original Tensor GS101 used, with two "super-big" cores, two more typical big cores, and four small cores. One thing that changes across generations is the frequency and one small tweak to the big clusters. The A76 cluster is replaced by an A78 cluster that's 100MHz faster at 2.35GHz. The other components remain the same, though. The X1 cluster has been bumped up by 50MHz, which gives it a frequency of 2.85GHz. This translates to a 10% to 15% better result in Geekbench, though you will be hard-pressed to notice much of this difference in real life.


Google has significantly upgraded the GPU, though. The Pixel 7 and 7 Pro are switching to the Mali-G710 GPU rather than the G78. That provides about 20% better performance and efficiency. The new GPU also helps the onboard machine-learning-focused TPU, giving it an up to 35% boost in applicable processes. The TPU is also seeing an upgrade.

The G2 is again paired with a Samsung-made modem, this time around, the Exynos S5300 5G. Mobile connectivity was poor on the Pixel 6 series and one of the biggest gripes many owners had with it. Based on initial reports from Pixel 7 owners, the situation is greatly improved with the new modem.


Overall, this small upgrade compared to the first-gen Tensor might be disappointing on paper, but it could make a lot of sense in the performance-to-power usage ratio. Newer processors are found to improve performance at the cost of energy consumption, so sticking with the older generation might leave more room for better efficiency. It also helps that Google has experience with this setup for a whole generation, making it simpler to optimize the system further. This is somewhat reminiscent of the company sticking with the same camera for multiple generations of Pixel phones, improving how the software interacts with the hardware with each iteration.

Most likely correct for the physical AKIDA chip, but I would assume that the AKIDA IP would be chip/wafer size agnostic?

 

[Fact Finder]

See below re Googles Tensor 2 chip. Forgive my lack of technical know how but it is a 4nm chip.My understanding is that Akida doesn't scale down to 4nm? Happy to be corrected.

Google Tensor G2: All the details on the Pixel 7's custom chipset

Google is hitting the custom silicon game hard! 🔥

www.androidpolice.com

​

Google Tensor G2 specifications​

For those that want a table, here you go:

Dev board code name	Cloudripper
Model number	GS201, Tensor G2
Cores	2x super-big ARM Cortex-X1, 2x big A78, 4x small Cortex-A55
GPU	Mali-G710
Manufacturing node	4nm Samsung PLP
Modem	Samsung Exynos 5300 5G

The Tensor G2 is made by Samsung on its 4nm node using panel-level packaging. This is a complicated way of saying the chips are carved out of a square wafer rather than a round one, reducing waste. This likely doesn't have much impact on the chip's performance in actual devices, but it's nifty and might reduce costs. Plus, it's potentially useful when we're still in the middle of a chip shortage.


The Tensor G2 keeps the 2+2+4 core cluster configuration that the original Tensor GS101 used, with two "super-big" cores, two more typical big cores, and four small cores. One thing that changes across generations is the frequency and one small tweak to the big clusters. The A76 cluster is replaced by an A78 cluster that's 100MHz faster at 2.35GHz. The other components remain the same, though. The X1 cluster has been bumped up by 50MHz, which gives it a frequency of 2.85GHz. This translates to a 10% to 15% better result in Geekbench, though you will be hard-pressed to notice much of this difference in real life.


Google has significantly upgraded the GPU, though. The Pixel 7 and 7 Pro are switching to the Mali-G710 GPU rather than the G78. That provides about 20% better performance and efficiency. The new GPU also helps the onboard machine-learning-focused TPU, giving it an up to 35% boost in applicable processes. The TPU is also seeing an upgrade.

The G2 is again paired with a Samsung-made modem, this time around, the Exynos S5300 5G. Mobile connectivity was poor on the Pixel 6 series and one of the biggest gripes many owners had with it. Based on initial reports from Pixel 7 owners, the situation is greatly improved with the new modem.


Overall, this small upgrade compared to the first-gen Tensor might be disappointing on paper, but it could make a lot of sense in the performance-to-power usage ratio. Newer processors are found to improve performance at the cost of energy consumption, so sticking with the older generation might leave more room for better efficiency. It also helps that Google has experience with this setup for a whole generation, making it simpler to optimize the system further. This is somewhat reminiscent of the company sticking with the same camera for multiple generations of Pixel phones, improving how the software interacts with the hardware with each iteration.

Hi @alwaysgreen

I have no idea about the bigger issue but Anil Mankar actually mentioned AKIDA being implemented in 4nm. Quite a long time ago but I am positive about this as a while after that Peter van der Made referred to AKIDA and 5nm being potentially a thing.

I also understood that there was actually no restriction on how low AKIDA technology can be scaled.

My opinion only DYOR
FF

AKIDA BALLISTA

 

[BRNlongterm]

Hey fellow BRN investors.

Can anybody help me out to find a good Stock screener? I am pretty new to investing. And currently I am learning how to understand a companies evaluation.

 

[VictorG]

View attachment 21546
MegaChips Quarter Results

https://pdf.irpocket.com/C6875/aG1f/lbpj/UtUa.pdf

IR Releases | MegaChips Corporation

We aim at sustainable growth through profitability and growth opportunity.

www.megachips.co.jp

Excellent post @Cardpro.

Clearly the vote of confidence by MegaChips in making Brainchip central to their growth strategy goes a very long way in justifying everyone's faith in Brainchip and its management.
Imagine how many other NDA holders are eyeing 2025 as the launch of Akida driven products.
Akida being ubiquitous is becomming a reality.

 

[wilzy123]

Sony Xperia

Lol... Sony mobile phones. Let's hope not...

 

[Diogenese]

Lol... Sony mobile phones. Let's hope not...

But what a camera!

 

[SERA2g]

Looking into the paradigm of growth (see pg. 12)it looks like 2023 would be the year of the commercialisation. This would be massive for brn too especially in 2024 where it looks like volume production starts. Wow!!!!

This is the point I think a lot of retail shareholders (that aren’t members of the Tsex cult) miss.

Akida is proven in silicon, we’ve licenced the IP to Renasas and Megachips.

We’re sitting here with a commercially available neuromorphic processor and are waiting for clients to finish designs and proof of concepts before they sign the dotted line that will create new licence agreements (or agreements with Renasas and Megachips).

Engineering takes time.
Manufacturing takes time.
Production takes time.
Integrating new products takes time.

For Brainchip, it is (IMO) no longer a matter of if, but when.

The first royalties are likely to come through in the first half of 2023 (minor amounts in Q4’22 if we’re lucky).

Once royalties have started, I am almost convinced that our quarterly Royalties will increase on the previous quarter continuously for the next decade, if not longer.

All we need to do is be patient.

IMO, DYOR.

 

[SERA2g]

Thanks @Andi85

So my view is that if I was required to make an objective and logical prediction as to which mobile phone AKIDA will find its way into via MegaChips then Sony Xperia is the prime candidate.

Prophesee two lense sensor may make an appearance at the same time.

My opinion only DYOR
FF

AKIDA BALLISTA

I’m soon to be in the market for a new phone.

Do you think my fiancé would appreciate dick pics taken from a Sony Xperia powered by next gen edge-AI neuromorphic technology or should I stick to an iPhone?

Xx

 

[Fullmoonfever]

Excellent post @Cardpro.

Clearly the vote of confidence by MegaChips in making Brainchip central to their growth strategy goes a very long way in justifying everyone's faith in Brainchip and its management.
Imagine how many other NDA holders are eyeing 2025 as the launch of Akida driven products.
Akida being ubiquitous is becomming a reality.

Is also good to see that both Megachips and Renesas are suppliers to Apple.

We're not on the list in 2021 unfortunately and doesn't mean much yet but is always an opportunity at least through our relationships.

 

[Fullmoonfever]

Is also good to see that both Megachips and Renesas are suppliers to Apple.

We're not on the list in 2021 unfortunately and doesn't mean much yet but is always an opportunity at least through our relationships.

View attachment 21577
View attachment 21575 View attachment 21576

Appears Apples camera CMOS sensor is supplied by Sony.

Now....if we can succeed with Prophesee onto Sony sensors then....who knows hey.


iPhone 14 Pro Max Bom cost exposure: the proportion of parts in the United States has increased significantly, and China has declined – yqqlm
October 8, 2022

Recently, Nikkei Shimbun cooperated with Tokyo research firm Fomalhaut Techno Solutions to dismantle the three models of the iPhone 14 series launched by Apple in September. Models on sale rose by about 20%, setting a new record. In terms of countries/regions, the proportion of US-made parts has made a major leap forward, surpassing the Korean factory, and the proportion of mainland China and Taiwan has shrunk.

The performance of the camera components has also improved, with the main camera CMOS sensor made by Sony Group, the largest of the three rear cameras, 30% larger in size and about 50% more expensive to $15.

Sony’s sensor has a unique layered structure that ensures the area of each pixel through a smaller sensor size, ensures brightness and suppresses noise, you can take high-quality photos.

iPhone 14 Pro Max Bom cost exposure: the proportion of parts in the United States has increased significantly, and China has declined – yqqlm - Breaking Latest News

iPhone 14 Pro Max的零件成本较去年开卖的iPhone 13 Pro Max暴增逾60美元至501美元。零件成本总额、增加幅度皆为2018年以来最高，制造成本的上涨，意味着苹果的利润率可能下滑。另根据最新公布的苹果2021会计年度供应名单，新增大陆供应商七家，剔除六家；新增台湾厂商三家，剔除7家……

www.breakinglatest.news

 

[Slade]

I’m soon to be in the market for a new phone.

Do you think my fiancé would appreciate dick pics taken from a Sony Xperia powered by next gen edge-AI neuromorphic technology or should I stick to an iPhone?

Xx

You will find the powerful closeup Zoom lens very useful.

 

[alwaysgreen]

Lol... Sony mobile phones. Let's hope not...

They really are great phones. It's just nobody buys them!

 

[Sirod69]

I looked for contact between Apple and us on LinkedIn.
Antonio J Viana
ARM's Chairman of the Board and Rene Haas are in connection with Xuan Gu,
ASIC Design at Apple. I'll connect with him sometime

 

[Sirod69]

I just found an article from IMEC written by Steven Latre (IMEC),
I searched about apple neuromorphic chip

Prof. Steven Latré, is leading the artificial intelligence research at imec, the R&D hub for nano-electronics and digital technologies. His main expertise focuses on combining sensor technologies and chip design with AI to provide end-to-end solutions in sectors such as health and smart industries. Next to this, he is also a part-time professor at the University of Antwerp.

Now when I searched for him, I found a contact of mine named Amirreza Yousefzadeh, Neuromorphic Hardware Researcher at imec.

I have the following contacts in common with him.

https://www.computerweekly.com/feature/How-Imec-hopes-to-industrialise-artificial-intelligence

 

[FabricatedLunacy]

This is the point I think a lot of retail shareholders (that aren’t members of the Tsex cult) miss.

Akida is proven in silicon, we’ve licenced the IP to Renasas and Megachips.

We’re sitting here with a commercially available neuromorphic processor and are waiting for clients to finish designs and proof of concepts before they sign the dotted line that will create new licence agreements (or agreements with Renasas and Megachips).

Engineering takes time.
Manufacturing takes time.
Production takes time.
Integrating new products takes time.

For Brainchip, it is (IMO) no longer a matter of if, but when.

The first royalties are likely to come through in the first half of 2023 (minor amounts in Q4’22 if we’re lucky).

Once royalties have started, I am almost convinced that our quarterly Royalties will increase on the previous quarter continuously for the next decade, if not longer.

All we need to do is be patient.

IMO, DYOR.

Let me fix that last sentence. All we need to do is to be patient and buy buy buy buy