BRN Discussion Ongoing

Yes thanks @Slade Datalogic do barcode monitoring in smart distribution centres and are new Prophesee partners. 😎

Very cool.

My opinion only DYOR
FF

AKIDA BALLISTA
 
  • Like
  • Fire
  • Love
Reactions: 24 users

cassip

Regular
Hint to Podcast in German:
Quote Rob Telson:
"I can't wait to share with the audience how, by combining the neuromorphic platforms of PROPHESEE and BrainChip, we are creating a unique vertical stack that adds a whole new dimension to commercial artificial intelligence implementation with remarkable inference performance and ultra-low power consumption."

btw: there exists a Suabian proverb "nothing said is praise enough" ;)
 
  • Like
  • Love
  • Fire
Reactions: 35 users

TopCat

Regular
An interesting article to go with the discussion on Datalogic


Finally, the company has also revealed that it's working on an AI technology that can automatically scan packages. Currently, workers have to scan barcodes on packages using hand scanners — this technology will eliminate the need to do that. With this scanning capability in place, human workers don't even need to pause while sorting packages: The system can quickly recognize a package the passes its camera. Amazon explained that its camera runs at 120 frames per second and is powered by computer vision and machine learning technology.
 
  • Like
  • Fire
  • Love
Reactions: 26 users

Foxdog

Regular
But i dont think we are limited to the Hey Mercedes. I read that they have sensors that only act on a spikes which saves car energy, so basically I believe that the smart sensors in the vehicles will be equipped with Akida, Thats my thoughts and im sticking to it.
Merc have regularly mentioned 'systems' when talking about AKIDA.....simple. My opinion, it's ubiquitous within the EQXX
 
  • Like
  • Fire
  • Love
Reactions: 12 users

Fox151

Regular
Moderate me for what exactly? It’s this mentality that has stopped great posters like Uiux posting here
Uiux was being a prick and got called out on it.
 
  • Haha
  • Like
Reactions: 8 users

ndefries

Regular
Merc have regularly mentioned 'systems' when talking about AKIDA.....simple. My opinion, it's ubiquitous within the EQXX
Merc is taking on the legal liability of passengers of their level 3 autonomous car. They are not going to rely on cloud and latency which Tesla seem to be comfortable with. We are so in their sensors (what's the word "sensor fusion") and and lidar via valeo it's going to be a great partnership hey Mercedes!
 
  • Like
  • Love
  • Fire
Reactions: 20 users

Foxdog

Regular
Have you considered disinfectant? Or I believe one of our local genius's suggested a nice Kero bath......🤣🤣🤣


Ahh the good ol days huh? Imagine how many MOAB's Trumpy would have dropped on Putin by now 😆👌
 
  • Haha
  • Like
  • Fire
Reactions: 7 users

TopCat

Regular
Event-based reference system

One, two or a thousand eyes for your machines ?​


From the Prophesee website….do you think they know about us??? 🤣
 
  • Haha
  • Like
  • Fire
Reactions: 28 users

Diogenese

Top 20
  • Haha
  • Like
  • Love
Reactions: 11 users
An interesting article to go with the discussion on Datalogic


Finally, the company has also revealed that it's working on an AI technology that can automatically scan packages. Currently, workers have to scan barcodes on packages using hand scanners — this technology will eliminate the need to do that. With this scanning capability in place, human workers don't even need to pause while sorting packages: The system can quickly recognize a package the passes its camera. Amazon explained that its camera runs at 120 frames per second and is powered by computer vision and machine learning technology.
This is clearly interesting and shows that the market is begging for this technology solution but the moment they speak in frames per second based on how I understood Luca Verre it rules out Prophesee as frames per second do not exist for it. Perhaps @Diogenese can shed some frames on this and rule it back in???


My opinion only DYOR
FF

AKIDA BALLISTA
 
  • Like
Reactions: 11 users

Diogenese

Top 20
I agree I struggle with how they categorise their relationships. I would have thought that after todays podcast Prophesee might be better styled as a partner but I suppose that line “you can call me anything you like except late for dinner” sort of fits.

After todays podcast we know much more than a label used on the website would tell us.

My opinion only DYOR
FF

AKIDA BALLISTA
If it were to be classified as a joint venture, then the profits are shared in an agreed ratio according to the value of each party's contribution.

For example if we assume the Prophesee DVS IP is equivalent to the Akida IP, and that Prophesee undertakes the manufacture, and Prophesee undertakes the marketing, then maybe the split would be three quarters to Prophesee and one quarter to Brainchip.

But then there would be Brainchip's contribution to arranging the technical compatibility, so it could be three fifths to two fifths, ... or not.
 
  • Like
  • Fire
Reactions: 12 users

alwaysgreen

Top 20
Does anyone know how much revenue Prophesee currently makes? I found a few figures but they were only estimates. They are not exactly setting the world on fire... yet.

Hopefully our joint solution has customers throwing $$$ at both of us.
 
  • Like
  • Haha
Reactions: 12 users

Diogenese

Top 20
This is clearly interesting and shows that the market is begging for this technology solution but the moment they speak in frames per second based on how I understood Luca Verre it rules out Prophesee as frames per second do not exist for it. Perhaps @Diogenese can shed some frames on this and rule it back in???


My opinion only DYOR
FF

AKIDA BALLISTA
... well Luca did talk about combining frames and DVS for mobiles to eliminate blurring ...
 
Last edited:
  • Like
  • Fire
Reactions: 16 users

Jefwilto

Regular
Does anyone know how much revenue Prophesee currently makes? I found a few figures but they were only estimates. They are not exactly setting the world on fire... yet.

Hopefully our joint solution has customers throwing $$$ at both of us.
Yes AG i had a look today,and saw it was about $25 mil PA,but i,m sure that now Brainchip is on board,that will improve dramatically 🙂
 
  • Like
  • Love
  • Fire
Reactions: 21 users

TopCat

Regular
This is clearly interesting and shows that the market is begging for this technology solution but the moment they speak in frames per second based on how I understood Luca Verre it rules out Prophesee as frames per second do not exist for it. Perhaps @Diogenese can shed some frames on this and rule it back in???


My opinion only DYOR
FF

AKIDA BALLISTA
Yes I think you’re right. I just listened to that section of the podcast again and it sounds like fps are not used
 
  • Like
  • Fire
Reactions: 8 users

Bravo

If ARM was an arm, BRN would be its biceps💪!
Just in case there was any doubt whatsoever, I wanted to make it crystal clear that I am definitely not a downpantser. No, no, no! I like to keep my pants tucked up under my chin, that way I know I haven't forgotten to put them on! You can call me Harriet High Pants from now on.

To clarify, I thought it would be interesting to ask the question whether BrainChip and Esperanto could be compatible technologies. Initially it was the links that both Esperanto and BrainChip have to SiFive (thus far separately as far as we know) that prompted me to ask the question.

The other thing I considered is that Esperanto and SiFive are big on this whole IFS ( Intel Foundry) ecosystem thing and they're also big on the CHIP Alliance which is going to see lot of different chip companies teaming up together making different chip-sets.

Anyway, I found a very interesting video, which I recommend because it's, well, interesting, which features Bob Brennan from IFS and Dave Ditzel from Esperanto. The whole interview is really interesting (third times a charm) but the bit that caught my attention was around the 4 minute mark when Dave talks about how teaming up with IFS means that Esperanto will be able to team their chiplets up with different types of chiplets to get new products to market in months rather than years. And that this this sort of collaboration will yield an explosion (yes, he said "explosion") of new products that haven't been thought of yet. Later in the interview there was a discussion that revolved around voice control and how prominently this will feature in the future (and we all know who does voice control 5-10 times better than conventional voice control).

I am unable to find anything to suggest Esperanto is into SNN and I can't find anything that would suggest they'd be any good with sensor fusion stuff and we know BrainChip's expertise is in sensory AI, with all 5 human senses which makes me see, hear, smell, taste and feel victory (hopefully)!

Just random thoughts on my behalf but I thought I'd share nonetheless.

B 💋


 
Last edited:
  • Like
  • Love
  • Fire
Reactions: 39 users

ItsKane

Regular
All these seagulls trying to get our delicious chip. One seagull arrives, 10 others follow. Have you noticed there's always a couple of loud ones screeching while the rest patiently wait. We'll it's the patient ones that I always gravitate towards and I make sure that they get their share of chip too, no matter how many chips I need to throw out to the continuous growing flock of sea birdies. Cheeky little gulls aren't they, they're everywhere. Everywhere
 
  • Like
  • Haha
  • Love
Reactions: 11 users

Diogenese

Top 20
Just in case there was any doubt whatsoever, I wanted to make it crystal clear that I am definitely not a downpantser. No, no, no! I like to keep my pants tucked up under my chin, that way I know I haven't forgotten to put them on! You can just call me Harry High Pants from now on.

To clarify, I thought it would be interesting to ask the question whether BrainChip and Esperanto could be compatible technologies. Initially it was the links that both Esperanto and BrainChip have to SiFive (thus far separately as far as we know) that prompted me too ask the question.

The other thing I considered is that Esperanto and SiFive are big on this whole IFS ( Intel Foundry) ecosystem thing and they're also big on the CHIP Alliance which is going to see lot of different chip companies teaming up together making different chip-sets.

Anyway, I found a very interesting video, which I recommend because it's, well, interesting, which features Bob Brennan from IFS and Dave Ditzel from Esperanto. The whole interview is really interesting (third times a charm) but the bit that caught my attention was around the 4 minute mark when Dave talks about how teaming up with IFS means that Esperanto will be able to team their chiplets up with different types of chiplets to get new products to market in months rather than years. And that this this sort of collaboration will yield an explosion (yes, he said "explosion") of new products that haven't been thought of yet. Later in the interview there was a discussion that revolved around voice control and how prominently this will feature in the future (and we all know who does voice control 5-10 times better than conventional voice control).

I am unable to find anything to suggest Esperanto is into SNN and I can't find anything that would suggest they'd be any good with sensor fusion stuff and we know BrainChip has all 5 sensors which makes me see, hear, smell, taste and feel victory (hopefully)!

Just random thoughts on my behalf but I thought I'd share nonetheless.

B 💋



Esperanto is an arrangement of RISC-V cores. Basically it is a RISC-V MPU/CPU.

https://www.esperanto.ai/technology/
ET-SoC-1 features:

  • 1088 energy-efficient ET-Minion 64-bit RISC-V in-order cores, each with a custom vector/tensor unit optimized for ML applications
  • 4 high-performance ET-Maxion 64-bit RISC-V out-of-order cores for running an OS in self-hosted mode
  • Over 160 million bytes of on-chip SRAM
  • Interfaces for large external memory with low-power LPDDR4x DRAM and eMMC FLASH
  • PCIe Gen4 x8 and other common I/O interfaces

Minimum energy consumption across a range
of datacenter inference workloads.​


Our approach employs an array of over a thousand of our low-power ET-Minion cores, delivering energy efficiency with high integer and floating-point throughput, including tensor and vector acceleration optimized for ML / DL workloads. Our distributed memory architecture puts on-chip memory where it’s needed, improving processing utilization and relieving memory bandwidth bottlenecks.


The Esperanto ET-Minion
on-chip RISC-V core.​


The ET-Minion core, based on the open RISC-V ISA, adds proprietary extensions optimized for machine learning. This general-purpose 64-bit microprocessor executes instructions in order, for maximum efficiency, while extensions support vector and tensor operations on up to 256 bits of floating-point data (using 16-bit or 32-bit operands) or 512 bits of integer data (using 8-bit operands) per clock cycle. Multiply-accumulate operations always work with 32-bit accumulators to avoid loss of precision for large arrays, a real problem for competing products that accumulate to smaller representations. The longest tensor operations execute for 512 cycles without consuming memory bandwidth or power to fetch and decode instructions. These operations are perfectly matched to the execution units; both run at full speed for the full duration of each operation.
 
  • Like
  • Fire
  • Love
Reactions: 19 users

Bravo

If ARM was an arm, BRN would be its biceps💪!
Esperanto is an arrangement of RISC-V cores. Basically it is a RISC-V MPU/CPU.

https://www.esperanto.ai/technology/
ET-SoC-1 features:

  • 1088 energy-efficient ET-Minion 64-bit RISC-V in-order cores, each with a custom vector/tensor unit optimized for ML applications
  • 4 high-performance ET-Maxion 64-bit RISC-V out-of-order cores for running an OS in self-hosted mode
  • Over 160 million bytes of on-chip SRAM
  • Interfaces for large external memory with low-power LPDDR4x DRAM and eMMC FLASH
  • PCIe Gen4 x8 and other common I/O interfaces

Minimum energy consumption across a range​

of datacenter inference workloads.​


Our approach employs an array of over a thousand of our low-power ET-Minion cores, delivering energy efficiency with high integer and floating-point throughput, including tensor and vector acceleration optimized for ML / DL workloads. Our distributed memory architecture puts on-chip memory where it’s needed, improving processing utilization and relieving memory bandwidth bottlenecks.


The Esperanto ET-Minion​

on-chip RISC-V core.​


The ET-Minion core, based on the open RISC-V ISA, adds proprietary extensions optimized for machine learning. This general-purpose 64-bit microprocessor executes instructions in order, for maximum efficiency, while extensions support vector and tensor operations on up to 256 bits of floating-point data (using 16-bit or 32-bit operands) or 512 bits of integer data (using 8-bit operands) per clock cycle. Multiply-accumulate operations always work with 32-bit accumulators to avoid loss of precision for large arrays, a real problem for competing products that accumulate to smaller representations. The longest tensor operations execute for 512 cycles without consuming memory bandwidth or power to fetch and decode instructions. These operations are perfectly matched to the execution units; both run at full speed for the full duration of each operation.



dumb-and-dumber-jim-carrey.gif
 
  • Haha
  • Like
  • Love
Reactions: 15 users

Pmel

Regular
  • Like
  • Fire
Reactions: 4 users
Top Bottom