DevHeads IoT Integration Server•15mo ago

https://tenor.com/view/run-explosions-gif-13620594

Tenor

Ttechielew https://tenor.com/view/run-explosions-gif-13620594

techielewOP•10/6/24, 5:51 AM

^^Umesh during his next Engineering Hour

PPhoenix emotional damage

ZacckOsiemo•10/6/24, 5:51 AM

remember this is just. the IDE and we are the benevolent here, we reserve the right to explode when overloaded

Phoenix•10/6/24, 5:52 AM

for ~50 concurrent users we'd still need servers with enough CPU and RAM to handle that
specially if we’re running a web-based IDE
it would be around $100–$300/month on AWS

techielewOP•10/6/24, 5:53 AM

100-300 more than current? i mean we could get that on patreon

Phoenix•10/6/24, 5:53 AM

$20–$50/month for load balancing if we want to reduce downtime

Ttechielew 100-300 more than current? i mean we could get that on patreon

Phoenix•10/6/24, 5:53 AM

yes

ZZacckOsiemo we can time slice or do priviledge based, then it becomes a sharing issue.

Ming•10/6/24, 6:27 AM

My generation solved that one too - Mainframes

ZacckOsiemo•10/6/24, 6:28 AM

knew it

MMing My generation solved that one too - Mainframes 😄

techielewOP•10/6/24, 6:52 AM

First day on the job: https://images.fineartamerica.com/images/artworkimages/mediumlarge/2/early-mainframe-computer-bettmann.jpg

MMing My generation solved that one too - Mainframes 😄

techielewOP•10/6/24, 6:53 AM

Actually did you run into any punch card systems at uni?

Ttechielew Actually did you run into any punch card systems at uni?

Ming•10/6/24, 6:57 AM

I was just after punch cards - but we still had to submit jobs through the control. The jobs would be run at some point later and we could pick up the results the next day.

MMing I was just after punch cards - but we still had to submit jobs through the contr...

techielewOP•10/6/24, 7:07 AM

So, what do you think about the live team coding idea? Tried something like that before?

Ttechielew So, what do you think about the live team coding idea? Tried something like that...

Ming•10/6/24, 7:12 AM

Never tried it - but would be up for it.

PPhoenix we don't want stuff to explode mid event

Umesh Lokhande•10/6/24, 7:31 AM

That is why we use AWS EC2 Instance to deploy client's IoT dashboards to handle higher traffic generated by iot devices (to save server from crash). They are very scalable. I'm not suggesting the use of EC2 for this case but just saying that the similarity between device-to-server and client-to-server communication!

bruce_wayne•10/6/24, 8:00 AM

Hey guys 4 years experienced Embedded Developer here

Experienced in Hardware + firmware design complete projects.

Any project idea suggestions to get into Embedded AI space ? (Training myself to fetch big bucks 1-2 years down the lane). Please suggest

Bbruce_wayne Hey guys 4 years experienced Embedded Developer here ✌️ Experienced in Hardware ...

techielewOP•10/6/24, 9:13 AM

Hey Bruce, are you familiar with TinyML?

MMing Never tried it - but would be up for it.

ZacckOsiemo•10/6/24, 9:13 AM

Oh wow I have something to teach you then

Bbruce_wayne Hey guys 4 years experienced Embedded Developer here ✌️ Experienced in Hardware ...

Edison_ngunjiri•10/6/24, 9:17 AM

Hi tinyML would be a good place to start . Also, if you are looking at speech to text or something related to audio in AI, you can try Picovoice. For the two options, they have specific devices supported. I did a project using Portenta H7 on controlling things via voice commands.

Bbruce_wayne Hey guys 4 years experienced Embedded Developer here ✌️ Experienced in Hardware ...

Edison_ngunjiri•10/6/24, 9:28 AM

Demo: https://www.linkedin.com/posts/edison-ngunjiri_embeddedsystems-embeddedsoftware-firmwareengineer-activity-7060947881603530752-W0la?utm_source=share&utm_medium=member_desktop

EDISON NGUNJIRI on LinkedIn: #embeddedsystems #embeddedsoftware #fi...

Project: Control LED using voice commands(FULLY OFFLINE AND ON EDGE)
Board: Portenta H7
Steps:-
1: A…

Ttechielew Hey Bruce, are you familiar with TinyML?

bruce_wayne•10/6/24, 9:58 AM

Yes have been trying to research about moving into this space online. Overwhelming information available regarding entering the ML / AI space.

EEdison_ngunjiri Hi tinyML would be a good place to start . Also, if you are looking at speech to...

bruce_wayne•10/6/24, 9:59 AM

Thank you, will check this out right away.

Bbruce_wayne Yes have been trying to research about moving into this space online. Overwhelmi...

techielewOP•10/6/24, 10:10 AM

@Renuel Roberts any tips for Bruce to get started?

Bbruce_wayne Yes have been trying to research about moving into this space online. Overwhelmi...

Renuel Roberts•10/6/24, 11:31 AM

Hello @bruce_wayne are u familiar with python

bruce_wayne•10/6/24, 11:32 AM

Yes absolutely. All my career I’ve been scripting using Python and sending commands over Serial or Ethernet to automate testing.

Renuel Roberts•10/6/24, 11:38 AM

Ahv u worked with any AI related libraries like decision trees, Tensorflow lite and pandas/ numpy or matplotlib for mL

Renuel Roberts•10/6/24, 11:43 AM

If you haven't, start by Data manipulation using pandas and NumPy, understanding algorithms, and working with libraries like Scikit-learn, TensorFlow, and PyTorch.
U can start by creating basic AI chat not till u understand the logic flow to move into complex projects

ke7c2mi•10/6/24, 4:35 PM

@techielew I was just reading the scroll back, this is quite interesting, it also sounds almost identical to a hardware in the loop ci setup with the key differences being we push from users instead of source control (I guess it would be encumbering for users to push git to flash the hardware?) And that users want to poke/shoogle/look at the hardware instead of automated tests.

I'd be interested to discuss further / help with this.

My gut feeling re the aws thing is I'm on the side of the $5 linode - I'm generally of the opinion that if the part talking to hardware needs to scale for users the architecture was built wrong - we usually run these things on something in the raspberry pi territory

P.s. - if you do use a PI, you get the gpios/busses so it's easy to hang off a couple of i2c/spi devices to stimulate various parts of the system, the linode would basically serve some rpc interface which should be very light and the "fancy web stuff" can be done client side - the linode is more there so that you arent exposing a pi in your house to the nasty world

Kke7c2mi @techielew I was just reading the scroll back, this is quite interesting, it als...

ZacckOsiemo•10/6/24, 4:57 PM

you need to teach me so many things please

ZacckOsiemo•10/6/24, 4:58 PM

wow I didnt know v2x was a thing already, this is really amazing.

ZZacckOsiemo you need to teach me so many things please 🙏🏾

ke7c2mi•10/6/24, 5:06 PM

In essence - look at hardware in the loop testing

Consider that a lot of the setup is identical

You could actually drive a system like this from git, people drop a link into some things, when it's their turn the system picks up the code flashes it and runs it.

The difficult part here is if you want people to be able to interact with the target in realtime - now just throwing jobs at it doesnt work, it becomes more like booking an appointment.

We were talking about building something like this on another server for an embedded advent of code, we hashed out some ideas but didnt have the time to commit and make it happen

Kke7c2mi In essence - look at hardware in the loop testing Consider that a lot of the se...

ZacckOsiemo•10/6/24, 5:12 PM

let me read up a bit, keen for this December maybe on this server let me see what resources I can gather, I'll come back with questions

Renuel Roberts•10/6/24, 8:09 PM

when working with Zephyr OS for embedded systems, what do you find most beneficial?

@Dtynin @Boss lady @melta101
I noticed you guys have work on some projects using zephyr can you please share some of your insights

RRenuel Roberts when working with Zephyr OS for embedded systems, what do you find most benefici...

ke7c2mi•10/6/24, 9:20 PM

Not tagged but the HAL and the use of device trees gives more hardware abstraction than other options. The closest thing in terms of completeness is the arduino hal

The general quality of the drivers is also good enough that you don't have to hack it up all the time

Kke7c2mi Not tagged but the HAL and the use of device trees gives more hardware abstracti...

Renuel Roberts•10/6/24, 9:34 PM

That's a great point, Zephyr’s HAL and device tree really do provide that extra level of hardware abstraction, which can simplify development significantly. How would you compare working with Zephyr’s device tree to other approaches you've used in the past? Do you find it as intuitive, or are there challenges you’ve had to overcome to fully leverage its potential?

Kke7c2mi In essence - look at hardware in the loop testing Consider that a lot of the se...

techielewOP•10/6/24, 10:19 PM

Yeah looking at it like scheduled HIL testing is a good point. There's a way to do this that could be much less expensive (ie, linode) and not reinventing the wheel if we're willing to do round robin 'appointment' scheduling. It's not what was originally envisioned, but perhaps a necessary next step.

We'd certainly welcome the assist, @ke7c2mi. Thanks for the offer. I have a lot going on this upcoming week but hopefully we'll be able to regroup next weekend.

Also, I'm open to any ideas from anyone that would provide an alternative to the "scheduling/appointment" phenomenon, which would be annoying.

Ttechielew Yeah looking at it like scheduled HIL testing is a good point. There's a way to ...

techielewOP•10/6/24, 10:21 PM

Maybe it's an extension of the AMA we just did with Ming. You have to submit your question and code ahead of time and then we'll select ones that are of the broadest interest and let that user 'drive' under guidance of the instructor/expert.

RRenuel Roberts That's a great point, Zephyr’s HAL and device tree really do provide that extra ...

ke7c2mi•10/6/24, 11:12 PM

I don't think it's particularly intuitive and I think it's quite a steep learning curve coming from bare metal / reading datasheets and poking memory. It has felt cumbersome for me compared to things like freertos/rtic/baremetal.

Device trees themselves are neat, and I think pretty intuitive. It's just some descriptive templatey IDL. On the other hand they map to your entire system in an abstract way so actually doing things with them can be complicated - the device tree itself isn't the problem there though!

CA•10/7/24, 12:28 AM

Hey Guys Cherry here, just joined the group. Excited to be here. Little intro about myself. I am a recent grad with over an year experience as an Embedded Software Developer.

RRenuel Roberts If you haven't, start by Data manipulation using pandas and NumPy, understanding...

bruce_wayne•10/7/24, 2:08 AM

Thanks for the suggestion. @Edison_ngunjiri also suggested a tinyML project that involves hardware and code that runs on a microcontroller. Am I correct to assume that it would be the first thing I need to understand as the step towards AI/ML in Embedded Systems?

Kke7c2mi I don't think it's particularly intuitive and I think it's quite a steep learnin...

ZacckOsiemo•10/7/24, 2:45 AM

felt the same to me it was odd

CCA Hey Guys Cherry here, just joined the group. Excited to be here. Little intro ab...

ZacckOsiemo•10/7/24, 2:47 AM

Welcome Cherry, let us know what you are working and any way we can assist.

RRenuel Roberts when working with Zephyr OS for embedded systems, what do you find most benefici...

melta101•10/7/24, 3:27 AM

It's pretty neat
Was able to ship few modules with Zephyr on it.
Development is pretty direct especially since I like Linux kernel and the whole dts style of development.

One oof moment was the difference in opensource licence between the Linux kernel and Zephyr
Had to take down my PR due to the incompatibility

CCA Hey Guys Cherry here, just joined the group. Excited to be here. Little intro ab...

techielewOP•10/7/24, 3:31 AM

Welcome Cherry

Where do you work/what do you work on currently?

Bbruce_wayne Thanks for the suggestion. @Edison_ngunjiri also suggested a tinyML project that...

Edison_ngunjiri•10/7/24, 4:27 AM

If you are interested in ML/AI in embedded systems, I would suggest you go right into it.
Just repeat the already existing projects on tinyML/picovoice.

With these two platforms, it's more of a process than actual coding.
After several projects ,you may want to look at pandas,numpy and scikit-learn for more understanding.
This has always been my approach.
From known to unknown.

EEdison_ngunjiri If you are interested in ML/AI in embedded systems, I would suggest you go right...

bruce_wayne•10/7/24, 4:28 AM

Thanks a ton

Ttechielew Welcome Cherry 🙂 Where do you work/what do you work on currently?

CA•10/7/24, 5:56 AM

Hi, I currently work for locomotive manufacturing company its been year. My current work is mostly on testing and integrating the legacy projects to QNX.

Thank you

CCA Hi, I currently work for locomotive manufacturing company its been year. My curr...

techielewOP•10/7/24, 6:02 AM

interesting that its qnx. worked with other RTOS?

CA•10/7/24, 6:13 AM

No I haven’t worked with other RTOS yet.

ZacckOsiemo•10/7/24, 11:54 AM

https://tweedegolf.nl/en/blog/137/rust-is-rolling-off-the-volvo-assembly-line Rust in in production in Volvo ECUs

Rust is rolling off the Volvo assembly line - Blog - Tweede golf

In my job I get to speak to lots of people about Rust. Some are just starting out, some have barely ever heard of it, and then some people are running Rust silently in production at a very large c ...

ZacckOsiemo•10/7/24, 11:55 AM

@32bitSaviour

https://tenor.com/view/run-explosions-gif-13620594

Similar Threads

Similar Threads

Similar Threads