Thanks Fmf,That would be these.
HiSilicon owned by Huawei.
They do have what's called their DaVinci architecture tied in with the AI.
HISILICON
KUNPENG
In a are four dual processor nodes (Huawei TaiShan 200) in the eX3 cluster with the HiSilicon KunPeng 920 processor (64-bit, 64-core, ARM). This makes a total of eight processors with 512 cores. Each Huawei node has 1 TB DDR4 main memory and 4.2 TB local SSD scratch storage
AKIDA NEURAL PROCESSOR
The KunPeng CPU nodes (see above) hosts four Akida Neural Processors from BrainChip. These processors are designed specifically for neuromorphic computing.
What's Apple's turnover?Apple heading for the edge….. (no Brainchip mention)
These Technologies Could Play a Big Role in the iPhone's Future — CNET
The iPhone just turned 15, and these technologies could shape where it goes next.apple.news
Hopefully it's not going to be Apple crumble.What's Apple's turnover?
That was so funny. Well lets hope its row only.
Our relationship with T2MC Is still active per someone’s comment in a prior post (can’t remember who sorry) and these guys also manufacture apple components so there is a slight chance that Brainchip and Apple have been made aware of each other for quite some time now. I can’t quite remember when the T2MC agreement was dated.Apple heading for the edge….. (no Brainchip mention)
These Technologies Could Play a Big Role in the iPhone's Future — CNET
The iPhone just turned 15, and these technologies could shape where it goes next.apple.news
Hopefully it's not going to be Apple crumble.
DThanks Fmf,
I'd forgotten about this.
eX3 is a Norwegian research project into exascale high performance computing (HPC).
https://www.ex3.simula.no/mission
Together with partners, Simula Research Laboratory has established eX3 to become the previously missing national resource that can prepare researchers for exascale computing in Norway. The eX3 infrastructure is currently funded for five years through the RCN program for national research infrastructures. In addition to the host institution Simula, the project consortium also counts the national HPC management body Sigma2, HPC research groups from the University of Tromsø, NTNU, the University of Bergen, and OsloMet, as well as the HPC technology providers Graphcore, Dolphin Interconnect Solutions, and Numascale.
The eX3 infrastructure is not an exascale computer by itself, but it is a carefully curated ecosystem of technology components that will be crucial for embracing exascale computing. It will allow HPC researchers throughout Norway and their collaborators to experiment hands-on with emerging HPC technologies – hardware as well as software.
The eX3 project is hosted by Simula Research Laboratory and
funded by the Research Council of Norway
At first glance, I thought the reference to 4 Akida neural processors was referring to 4 Akida NPUs or 4 Akida nodes, because 4 Akida 1000s would be a supercomputer on its own, but having 4 Akida 1000s in a research project allows for several simultaneous implementations of experiments to be run in parallel, or one very big project.
D
I did some digging way back when first found eX3.
What I found interesting is that Huawei have their own Ascend 910 AI chip, DaVinci AI architecture (mentioned prev) and what are called Atlas server range which use the 910 but also some models have the KunPeng 920 which is what eX3 are coupling with Akida.
Particularly the Atlas 800 uses KunPeng.
Why? I'm curious what the scope and desired result sought is via the eX3 project.
Something to dig through on the Huawei side too maybe?
product-solution-l1
e.huawei.com
So running as a coprocessor as part of what we know Akida can be used for?Hi Fmf,
I think the word "hosts" is determinative:
AKIDA NEURAL PROCESSOR
The KunPeng CPU nodes (see above) hosts four Akida Neural Processors from BrainChip. These processors are designed specifically for neuromorphic computing.
I guess Akida is used as an accelerator to do the heavy lifting with the CNN-type workloads for the separate KunPeng CPU.
https://en.wikichip.org/wiki/hisilicon/kunpeng/920-6426
Kunpeng 920-6426 is a tetrahexaconta-core 64-bit ARM server microprocessor introduced by HiSilicon in early 2019. Fabricated by TSMC on a 7nm HPC process based on the TaiSHan v110 microarchitecture, this chip incorporates 64 cores operating at 2.6 GHz with a TDP of 180 W. This chip supports up to 2 TiB of octa-channel DDR4-2933 memory.
What's Apple's turnover?
Well I reckon that having 4 Akida coprocessors would turn any half-decent CPU into a supercomputer.So running as a coprocessor as part of what we know Akida can be used for?
Guess still curious on the project given Huawei have their own available AI.
Wonder what outcome / results eX3 would aim for out of the project other than just a direct comparison of performance or capabilities of throwing a SNN processor in and why just KunPeng ad they have other available products in their project program by the looks.
Guess time will tell maybe.
"Delivering parcels inside your house" - I'm going to patent the concept of a house with a chute through the roof for parcel delivery ... or has that been done before?What I really love about this article is that when you click on the link shown below, it takes you to another article in which Apple describes precisly why it needs to needs to adopt Akida technology to maintain it's edge (so to speak, hehehe). @Dhm It might be time to fire off an email to old mate Tim Cook!
View attachment 10316
Extract of linked article.
View attachment 10317
"Delivering parcels inside your house" - I'm going to patent the concept of a house with a chute through the roof for parcel delivery ... or has that been done before?
Autonomous arms race is underway.
Fire and forget weaponry is being born... :/
"Only AI can keep up with AI against drone swams" DW news...
Hey DWell I reckon that having 4 Akida coprocessors would turn any half-decent CPU into a supercomputer.
I guess Simula were just breadboarding a few COTS processors to try to give the students an eXacomputer experience without having a one-piece eXacomputer.
This could well be a demonstration of how Akida can fix the cloud. (Intel - are you listening?)