Breaking News
Comments
Oldest First | Newest First | Threaded View
Page 1 / 3   >   >>
MichaelSB123
User Rank
Rookie
chip technology?
MichaelSB123   10/11/2013 7:18:54 PM
NO RATINGS
What does it mean exactly "implemented in hardware"?

Is NPU using regular CMOS technology as the rest of the processor, or is it some kind of specialized analog circuitry?

prabhakar_deosthali
User Rank
CEO
Re:
prabhakar_deosthali   10/12/2013 1:30:47 AM
NO RATINGS
It will be interesting to track how these new NPUs evolve to imitate those brain functions which make the learning process a natural one - a kids play.

junko.yoshida
User Rank
Blogger
NPU for automotive?
junko.yoshida   10/12/2013 5:39:54 AM
NO RATINGS
How far are we exactly to use NPU for vision procesing in automotive applications? What needs to get done before we achieve such a goal? Curious mind wants to know.

Amedeo87
User Rank
Rookie
Re: chip technology?
Amedeo87   10/12/2013 4:19:29 PM
NO RATINGS
maybe is builds with memristor?

_hm
User Rank
CEO
Good news but add optical processing to it
_hm   10/12/2013 7:46:49 PM
NO RATINGS
It is almost two decades. But now, it is a good news.

Can Qualcomm add optical processing to it? It will give real edge for future of computing.

R_Colin_Johnson
User Rank
Blogger
Re: chip technology?
R_Colin_Johnson   10/13/2013 12:46:08 PM
NO RATINGS
Qualcomm is not revealing many details yet, but is using a spiking neural network model--a digital representation that is compatible with standard CMOS processing.

R_Colin_Johnson
User Rank
Blogger
Re: NPU for automotive?
R_Colin_Johnson   10/13/2013 12:54:44 PM
NO RATINGS
@ NPU for automotive? 
junko.yoshida 

Next year Qualcomm will release its suite of software tools that work with an FPGA emulator for developers to use when creating applictions for its NPU. Regarding automotive, Purdue University professor Eugenio Culurciello has already shown that recognition of roadside scenes can result in realtime classification into pedestrians, vehicles, buildings, etc., but I suspect it will take a year or two for developers to start making good use of this type of information in collision avoidance and similar automotive applications.

Terry.Bollinger
User Rank
Manager
The role of federal research funding
Terry.Bollinger   10/13/2013 1:48:54 PM
NO RATINGS
When I first saw this I thought, "Rats! Why can't all of the fantastic and very similar work that Eugenio Culurciello has been doing for years under Office of Naval Research (ONR) funding get this kind of press...?" This also shows why one should read figure captions. From said caption on the familiar-looking first figure I finally realized that this _is_ the next phase for Dr Culurciello's amazing chips! Since shutting down the government seems to be the theme-du-jour, it's worth pointing out just how huge a role federal funding plays in giving folks like Dr Culurciello a chance to move the ball on some wild new idea to the point where a large company like Qualcomm can see the potential and catch the pass. That's the kind of teaming where everyone benefits.

rick merritt
User Rank
Author
Re: The role of federal research funding
rick merritt   10/13/2013 3:27:36 PM
NO RATINGS
Good to see Qcomm pumping some of its profits into long term research

moloned
User Rank
Rookie
Re: chip technology?
moloned   10/14/2013 4:44:19 AM
NO RATINGS
More hype than substance I'd say at least when it comes to the claims about a neural network processing unit (NPU) in hardware.  

Looking at the braincorp jobs page http://braincorporation.com/index.php/category/opportunities/ these guys are hosted inside Qualcomm and are focused on building robots, robotics algorithms and tools.  

They are hosted inside Qualcomm and appear to be leveraging Qualcomm's GPUs using OpenGL and OpenCL, as well as C/C++, Python and Matlab to to the NN processing rather than some kind of specialised NN processor.  

I'd say most of the NN work is being done in Matlab with some kind of back-end code generation of C/C++ code or potentially OpenGL/OpenCL ES code, the smart approach would be to have libraries for the GPU which are optimised at low level and be able to call identical libraries within Matlab or Python for rapid development rather than designing an esoteric compiler.

I'd guess this may lead to adding NPU support in the ISA of future GPUs at some point if they really need it

Page 1 / 3   >   >>


Most Recent Comments
Flash Poll
Top Comments of the Week
Like Us on Facebook
EE Times on Twitter
EE Times Twitter Feed

Datasheets.com Parts Search

185 million searchable parts
(please enter a part number or hit search to begin)
EE Life
Frankenstein's Fix, Teardowns, Sideshows, Design Contests, Reader Content & More
Engineering Investigations

Air Conditioner Falls From Window, Still Works
Engineering Investigations
Post a comment
It's autumn in New England. The leaves are turning to red, orange, and gold, my roses are in their second bloom, and it's time to remove the air conditioner from the window. On September ...

Max Maxfield

Feast Your Orbs on My Jiggly Exercise Machine
Max Maxfield
54 comments
Last weekend, I was chatting with my mother on the phone. She's all excited that I'm coming over to visit for a week in November. "I'll be seeing you in only seven weeks," she trilled ...

David Blaza

RadioShack: The End Is Nigh!
David Blaza
124 comments
I'm feeling a little nostalgic today as I read about what looks like the imminent demise of RadioShack, at least as we currently know it. An old ubiquitous cartoon image popped into my ...

Larry Desjardin

Engineers Should Study Finance: 5 Reasons Why
Larry Desjardin
47 comments
I'm a big proponent of engineers learning financial basics. Why? Because engineers are making decisions all the time, in multiple ways. Having a good financial understanding guides these ...