BRN Discussion Ongoing

Afternoon Chippers ,

She's all happening next week ,

1 , 22 Jan ,Trump gets to wear the big hat & wield the baseball club .

2 , 23 jan , private jets aplenty & gallons of conyac will be consumed ,
DAVOS , all the big wigs rubbing shoulders figuring out world domination.
AI , is the theme for this year's fancy dress party.


3 , After 13 years this possum has come out of rehabilitation, sugar addiction , now the hard work starts, bit like BRN.


Regards,
Esq.

All I know is I love trump until he dumps as I brought some of his meme coin today @7für7
 
  • Thinking
Reactions: 1 users

7für7

Top 20

Diogenese

Top 20
Maybe someone saw your post. When I hover over "Products" to get the popup, I notice that the "Shop" item appears and takes me to the page with the M.2 device listed.

However, navigation to the "Akida Enablement Platforms" page still does not list the M.2 device. If anything, they are consistently inconsistent.
Thanks JD,

I missed that.

PS: I had my screen less than half width, and the Drop down menu does not appear.
 
Last edited:

JDelekto

Regular
And while we're on M2, as Tom Waits says: "get you a little something you can't get at home"

https://www.officeworks.com.au/shop...o-m-2-nvme-and-sata-enclosure-10gbps-cocmnv01

Comsol USB-C to M.2 NVME and SATA Enclosure 10Gbps​


This Comsol USB-C Enclosure allows you to connect an existing M.2 NVMe or SATA SSD to your laptop of desktop. Supporting up to 10 Gbps for NVMe and up to 5 Gbps for SATA, this enclosure is compatible with 2230, 2242, 2260 and 2280 SSDs. It is host powered, and features a screwless design for your convenience.

  • This enclosure can be used to connect an existing M.2 NVMe or SATA SSD to your laptop or desktop.
  • It has a USB-C male connector.
  • It supports NVMe transfer speeds of up to 10 Gbps and 5 Gbps for SATA.
  • The M.2 slot accepts both NVMe M-Key and SATA B+M-Key SSDs.
  • It's compatible with 2230, 2242, 2260 and 2280 SSDs.
  • It includes a 20 cm USB-C to USB-C 10 Gbps cable.
  • It features a screwless design for convenience.
  • It's host powered, so no additional power is required.
  • It's suitable for use with Windows, MacOS and ChromeOS devices.
An adapter like this, with suitable interface, would enable an Akida M2 to connect to any USB3 computer via a USB 3 socket.

This looks tempting, but my only concern would be whether or not the driver would interface with it. Currently, BrainChip only supports drivers for Linux (with the Ubuntu distribution being officially supported).

I am concerned that BrainChip's Linux driver would not work with the USB-C enclosure which will treat the device like an NVMe SSD instead of the specialized hardware.
 

JDelekto

Regular
PC motherboards, these days, tend to have more than 1 M.2 slot. One is for the main OS drive and the other can be used as storage and sometimes to even connect a wifi card etc. I wonder if the AKIDA M.2 would work in one of these onboard slots?
Depending upon how the slot is keyed (and in some cases configured in the BIOS), the Akida M.2 device is designed to work in these slots. The keying determines where the notch is cut in the board to fit in the slot. The motherboard manual should give you the specifics for each of the slots.

The M.2 is a specification for the slot interface, and it's important to note that some M.2 slots can take both NVMe and SATA devices (like solid-state drives). On a side note, if given a choice between an NVMe or SATA solid-state drive that will fit in an M.2 slot, go for the NVMe drive as they are significantly faster.

Now NVMe and SATA are specific protocols. NVMe uses four PCIe lanes, and SATA uses a single lane. According to BrainChip's specifications for the NVMe device, it only uses 2 lanes. That being said, I don't think it would work exactly like a NVMe or SATA drive and thus requires a custom driver:
1737283662796.png

Also, when you see numbers associated with the devices, like 2230, 2260, and 2280 (the Akida is 2260), those numbers usually indicate the width and length, respectively, in millimeters of the board.

As long as you are using Ubuntu Linux with BrainChip's PCIe driver, you should be able to use one of the M.2 slots on the motherboard.
 

Diogenese

Top 20
This looks tempting, but my only concern would be whether or not the driver would interface with it. Currently, BrainChip only supports drivers for Linux (with the Ubuntu distribution being officially supported).

I am concerned that BrainChip's Linux driver would not work with the USB-C enclosure which will treat the device like an NVMe SSD instead of the specialized hardware.

Hi JD,

I see you've looked at this in depth back in May 2022:

I sent an e-mail to support and inquired if they had a driver for Windows, but unfortunately, they only provide PCIe drivers for Ubuntu 18.04 and 20.06.

That's why one of my future goals is to see if I could analyze the source for the Linux PCIe driver (which BrainChip does make available on GitHub) and see if I can create a driver for Windows as well.

I'm not sure if there is some type of "shim" that needs to be written to use this with Python, but that may have to be done as well. However, with a driver, I see no reason why the Akida board couldn't be used wit
h Windows as well.

Looks like they read your post:

https://doc.brainchipinc.com/
Inspired by the Keras API, MetaTF provides a high-level Python API for neural networks. This API facilitates early evaluation, design, final tuning, and productization of neural network models.

PS:
MetaTF:
https://doc.brainchipinc.com/installation.html

Installation

Supported configurations

  • Operating systems:
    • Windows 10, Windows 11
    • Any Linux variant compatible with manylinux 2.28 (Ubuntu 20.04, Ubuntu 22.04, …)
  • Python versions: 3.9 to 3.11
  • TensorFlow versions: 2.15
 
Last edited:
  • Like
Reactions: 7 users
I Love this story, read it a long time ago, it relates to me, how those who may consider themselves "open minded" are actually trapped within the restraints placed upon their mind's through the lessons of "education" and what they are "told" is Fact or Fiction, without the breadth of Knowledge, to determine for themselves, what is possible.


The Elephant Rope (Belief)

A gentleman was walking through an elephant camp, and he spotted that the elephants weren’t being kept in cages or held by the use of chains.

All that was holding them back from escaping the camp, was a small piece of rope tied to one of their legs.

As the man gazed upon the elephants, he was completely confused as to why the elephants didn’t just use their strength to break the rope and escape the camp. They could easily have done so, but instead, they didn’t try to at all.

Curious and wanting to know the answer, he asked a trainer nearby why the elephants were just standing there and never tried to escape.

The trainer replied;
“when they are very young and much smaller we use the same size rope to tie them and, at that age, it’s enough to hold them. As they grow up, they are conditioned to believe they cannot break away. They believe the rope can still hold them, so they never try to break free.”

The only reason that the elephants weren’t breaking free and escaping from the camp was that over time they adopted the belief that it just wasn’t possible.



Another story I Love about elephants, is the one about the sleeping elephant, that tramples everything, when it awakes.

That's BrainChip! 😉
 
  • Haha
Reactions: 1 users

JDelekto

Regular
Hi JD,

I see you've looked at this in depth back in May 2022:

I sent an e-mail to support and inquired if they had a driver for Windows, but unfortunately, they only provide PCIe drivers for Ubuntu 18.04 and 20.06.

That's why one of my future goals is to see if I could analyze the source for the Linux PCIe driver (which BrainChip does make available on GitHub) and see if I can create a driver for Windows as well.

I'm not sure if there is some type of "shim" that needs to be written to use this with Python, but that may have to be done as well. However, with a driver, I see no reason why the Akida board couldn't be used wit
h Windows as well.

Looks like they read your post:

https://doc.brainchipinc.com/
Inspired by the Keras API, MetaTF provides a high-level Python API for neural networks. This API facilitates early evaluation, design, final tuning, and productization of neural network models.

PS:
MetaTF:
https://doc.brainchipinc.com/installation.html

Installation

Supported configurations

  • Operating systems:
    • Windows 10, Windows 11
    • Any Linux variant compatible with manylinux 2.28 (Ubuntu 20.04, Ubuntu 22.04, …)
  • Python versions: 3.9 to 3.11
  • TensorFlow versions: 2.15
I need to contact one of their support engineers inquiring about hardware support. The MetaTF package that can be installed on Windows and used on Windows has an Akida IP simulator, but I don't believe it will work with the physical hardware.

When I recently ordered the M.2 device and was contacted to choose which Key I wanted (they didn't have it as a selection on the order form), I asked if they had any other documentation I could use since I wanted to work on a Windows driver. I got the following response:

Our platform runs on Linux, ubuntu. It hasn’t been qualified on Windows platforms.
All the information you need to get up and running can be found on our MetaTF site.

I'm not quite sure what they mean by "qualified" (maybe they now have a Windows driver, but it is not Microsoft WHQL certified?) but I understood it couldn't currently be used on Windows. The GitHub repo only has the source for the PCIe driver to be compiled on Ubuntu.

I know that I had originally installed Windows 10 on my shuttle PC when I purchased the PCIe card, and it wouldn't even detect the hardware nor try to install any drivers. After adding a new SSD and installing Ubuntu, I successfully ran their scripts to install the kernel driver and use the tools to detect and test the Akida PCIe card.

Ultimately I want to be able to use the hardware on Windows operating systems, not the simulator that comes with MetaTF.
 
Last edited:

Diogenese

Top 20
I need to send an e-mail to one of their actual Engineers regarding hardware support. The MetaTF package that can be installed on Windows and used on Windows has an Akida IP simulator, but I don't believe it will work with the physical hardware.

When I recently ordered the M.2 device and was contacted to choose which Key I wanted (they didn't have it as a selection on the order form), I asked if they had any other documentation I could use since I wanted to work on a Windows driver. I got the following response:

Our platform runs on Linux, ubuntu. It hasn’t been qualified on Windows platforms.
All the information you need to get up and running can be found on our MetaTF site.

I'm not quite sure what they mean by "qualified" (maybe they now have a Windows driver, but it is not Microsoft WHQL certified?) but I understood it couldn't currently be used on Windows. I know that I had originally installed Windows 10 on my shuttle PC when I purchased the PCIe card, and it wouldn't even detect the hardware nor try to install any drivers. After adding a new SSD and installing Ubuntu, I successfully ran their scripts to install the kernel driver and use the tools to detect and test the Akida PCIe card.

Ultimately I want to be able to use the hardware on Windows operating systems, not the simulator that comes with MetaTF.
All above my paygrade, but here is the low down on WHQL:

https://www.intel.com/content/www/us/en/support/articles/000093158/graphics.html

There are two types of drivers listed on the Download Center: drivers with and without a WHQL Certification. What are the differences between the two?
Resolution
Drivers labeled as WHQL Certified have been thoroughly tested by Intel, have passed Windows Hardware Lab Kit testing on various platforms and configurations, and are signed by Microsoft as compatible with Windows* operating systems.
Drivers that do not have WHQL Certification are also thoroughly tested by Intel, are of the same functional quality as WHQL Certified drivers, and are signed by Microsoft. The key difference is that Non-WHQL drivers have not completed the full Windows Hardware Lab Kit testing prior to release. This is known as an attest-signed driver
.

... but this may only apply to things available from the Intel download centre.

But there is a requirement for the latest Visual C++ redistributable package when using Windows:

https://doc.brainchipinc.com/installation.html

Using Windows, the latest Visual C++ redistributable package is required. Please refer to this link for installation.

available from MS:

https://learn.microsoft.com/en-us/cpp/windows/latest-supported-vc-redist?view=msvc-170
 
Last edited:
Top Bottom