Distributed Computing

[@ahmeddraz962]

5 years passed and still sharp and clear Explanation.

[@NERDfirst]

Hello and thank you very much for your comment! Glad you liked the video =)

[@preetisaroj2872]

the way you expalain the things is just amazing.....!!

[@NERDfirst]

Thank you very much! Very happy to be of help =)

[@shabirmalik6422]

Ofcourse @Preet Kaur

[@sithidethbouasavanh5558]

Subscribed. I wonder why our videos have only thousands of views, they deserve millions.

[@NERDfirst]

+hoppy BOUASAVANH Hello and thank you very much for your comment and subscription! I think I'll get more views if I advertised, but I'm currently taking a very "indie" approach to getting my work out there (read: I'm lazy xD).

[@adinarhas]

I love your voice, and your face, somehow you make it easy to listen to you.

[@NERDfirst]

Thank you very much! Glad you liked the video =)

[@Privacy-LOST]

Very, very clearly conveyed. Kudos

[@NERDfirst]

Hello and thank you for your comment! Glad to be of help =)

[@welligtonsantos3754]

Excellent explanation. I'm impressed! Thank you.

[@NERDfirst]

You're welcome! Glad you liked the video =)

[@DreamfireFilms]

Thanks man, I do VFX and some Blender 3D work and this was exactly the information I was looking for. Subbed!

[@NERDfirst]

Dreamfire Films Thank you very much for your comment and support! Glad you liked the video =)

[@hindj9142]

7 years later Thank you that was really great explanation

[@NERDfirst]

You're welcome! Very happy to be of help =)

[@gnaneshwarrao174]

Hey Keep it up. The explanation is clear with relevant examples. Thanks for the explanation. Language is clear. Make more videos.

[@NERDfirst]

+Gnaneshwar Rao Hello and thank you very much for your comment! Glad you found the video useful =) More videos to come three times a week!

[@kehcat1]

this is great! thanks. I didn't understand until this vid!

[@NERDfirst]

You're welcome! Very happy to be of help =)

[@chapparfellow]

Excellent explanation. Thank you.

[@NERDfirst]

Hello and thank you for your comment! Glad you liked the video :)

[@arnoldkyambadde]

Great breakdown. Thank you!

[@NERDfirst]

You're welcome! Very happy to be of help =)

[@MuhidAbid14point75]

Excellent video!

[@NERDfirst]

Hello and thank you for your comment! Glad you liked the video =)

[@Mooseydog]

Very well stated, I'm impressed!

[@NERDfirst]

+Mooseydog Thank you very much! Glad you liked the video =)

[@bitelogger]

Great Explanation Bro!!!

[@NERDfirst]

Hello and thank you very much for your comment! Glad to be of help =)

[@HarryPotter-fb2np]

Thank man! I want to know how we can speed up distributed computing using coding techniques?

[@NERDfirst]

Hello and thank you for your comment! This question is very broad, so to answer as broadly, I would say to use more efficient algorithms.

[@navbot1351]

Thanks, your explanation was great!

[@NERDfirst]

You're welcome! Very happy to be of help =)

[@SandeepKaur-su4it]

Excellent explanation....thx 👍

[@NERDfirst]

You're welcome! Glad to be of help =)

[@morampudiakhil516]

very good explanation brother

[@NERDfirst]

Hello and thank you very much for your comment! Glad you liked the video =)

[@dhoteliers9284]

When he said "a random wednesday episode" 6 years ago and I'm watching it for the first time today which is also a wednesday. LOL what a coincidence.

[@NERDfirst]

Hah! Glad it lined up nicely for you =)

[@Jdogrey1]

I wonder what this guy thinks of real time ray tracing.

[@NERDfirst]

Hello and thank you for your comment, this guy here. Real time raytracing was actually one of my big interest areas in university (that was like 8 years ago before it became a consumer technology). I think it's cool, I'm glad it's become the consumer tech we see today. In video games I'm not so sure if it has a really significant value-add over the more traditional shader-based approaches (and shaders have gotten incredible these past couple of years), but I feel having that hardware acceleration for ray tracing is still great for say 3D modelling.

[@Jdogrey1]

@@NERDfirst yeah, it is crazy to see how far we have come in the years since this video came out. Crazy to see you mention ray-tracing as something you have to have multiple computers work on for a single frame, and now we have single computers making hundreds of ray-traced frames in a second (though the ray-tracing done in that kind of time pales in comparison to the kind I think you were referring to, but it is still insane). Also, the way I see ray-tracing for video games is that it opens up a really great way for developers to more simply set up lighting in their games. As far as I am aware, things like rasterization are far more complicated to set up and therefore require more work. Using ray-tracing often makes games look better not because it is better quality, but because less has to be set up for it to work.

[@NERDfirst]

Yes - Even though real-time ray tracing is a thing, render farms aren't going to go away because of that. The kind of 3D you see in Hollywood blockbusters or your Pixar animated film do _a lot_ more raytracing than what you'd find in games (essentially everything you see onscreen is raytraced and every ray bounces and spawns a bunch of rays etc, so you're still looking at minutes per frame on fast hardware). Real-time raytracing may help accelerate these processes, but I think they really shine in making it easier to work with. Back then you had no idea what a raytraced scene looks like until you rendered it, but now you can get previews that get pretty close. And yeah, good point about the ease of setup as well. Shader-based techniques can sometimes be quite hacky, raytracing is straightforward and intuitive. Though of course, the drawback is that the user must have supported hardware!

[@appliedscience5880]

great video........and fine elaboration, thnx.

[@NERDfirst]

+Applied Science Hello and thank you for your comment! Glad to be of help =)

[@nikhilfadnis8009]

Good explanation my man!

[@NERDfirst]

Cheers! Glad to be of help =)

[@markyap3865]

Excellent explanation my friend

[@NERDfirst]

Hello and thank you very much for your comment! Very happy to be of help :)

[@berry5035]

Excellent explanation 😊

[@NERDfirst]

Thank you very much! Very happy to be of help :)

[@nnwccc]

simple and accurate explanations!!

[@NERDfirst]

Thank you very much! Glad you liked the video :)

[@MrWalterljohnson]

an excellent explaination

[@NERDfirst]

Thank you very much! Glad you liked the video =)

[@s.v.arulshiju507]

Can you suggest me few url links to find the right materials to create a distributed computing program (tiled (with ray tracing) as you said) ...? I have a plan to create one for blender 3d ...? Which programming language is best suitable? Can you say the advantage and disadvantage of using different programs like erlang, python, java etc. for distributed computing ... It will be nice if i can create everything in python since blender is made with python ...

[@NERDfirst]

Hello and thank you for your comment! This project sounds like it's going to be on a very big scale, and I'm afraid I don't really have the know-how to help you with this. At the same time, render farm management software already exist for Blender - You can look towards software like Loki or Flamenco. Unless you're building for exploratory reasons (and not to address a need) the solution already exists.

[@saulgoodman980]

My takeaway: Distributed computing is all about horizontal scaling. To get desirable results, the computation / task has to be parallelizable i.e. the computations should be able to run simultaneously and shouldn't depend on any previous step results.

[@NERDfirst]

Hello and thank you for your comment! Great summary! I would say, as much as possible we would want to avoid dependencies on previous step results, though there sometimes are ways to reformulate this problem such that it happens infrequently. For example, if I wanted to sum 100000 numbers, the approach of number1 + number2 + number3 + ... would be serial as each step is dependent on the previous step. However, I could do (number1 + number2), and (number3 + number4), and so on as independent steps. While each step wouldn't rely on each other, I would have to do this in multiple passes to get the final answer. Still dependent, but in a far more efficient way!

[@saulgoodman980]

@@NERDfirst Ahhh yes, thank you :)

[@athiahs2074]

Nice video ! Do you have any idea for any simple sequential system project that can be convert into a parallel system ?

[@NERDfirst]

+athiah s Hello and thank you for your comment! Parallelism isn't exactly my forte, and it's not immediately clear what exactly you need to do (I'm guessing this is homework of some kind), but one really simple idea might be sorting. There are a few trivial ways to approach sorting in a parallel context, and also some really involved ways, so that might be a good place to start.

[@VoodooChile69]

Thanks alot dude !🤘🏻

[@NERDfirst]

You're welcome! Happy to be of help =)

[@AlejandroPerea]

Hello there, I want to do a microcontroller cluster, and I was wonder about some task to perform on it. What are the most common examples (hello world) for parallel programing? Thanks in advance.

[@NERDfirst]

Alejandro Perea Rosales Hello and thank you for your comment! I don't actually know of any popular examples that are commonly used like in academia or anything, but some ideas off the top of my head: 1) Bucket Sort: Instead of sorting a list as it is, break it up into contiguous partitions and sort each partition on its own thread, then put everything together 2) Matrix Multiplication: Every row/column pair can be done in its own thread since the result of multiplying each pair doesn't affect the other pairs. Of course, these are "Embarassingly Parallel" tasks (en.wikipedia.org/wiki/Embarrassingly_parallel). Actual examples of tasks in distributed computing are more challenging because the tasks aren't independent, unlike what we have above.

[@emmanuelellis4996]

Good video! Thank you!

[@NERDfirst]

You're welcome! Glad to be of help =)

[@27010466]

Great Job! Keep It up

[@NERDfirst]

+Jonaseu Thank you very much for your comment! Glad you liked the video =)

[@vishaltanawade7637]

Too much helpful

[@NERDfirst]

Hello and thank you for your comment! Glad to be of help =)

[@achin4140]

sir pleasee tell about Distributed Deadlock Detection: system model, resource Vs communication deadlocks, deadlock prevention, avoidance, detection & resolution, centralized dead lock detection, distributed dead lock detection, path pushing algorithms, edge chasing algorithms. Agreement Protocols: Introduction, System models, classification of Agreement Problem, Byzantine agreement problem, Consensus problem, Interactive consistency Problem, Solution to Byzantine Agreement problem, Application of Agreement problem, Atomic Commit in Distributed Database system

[@NERDfirst]

Hello and thank you for your comment! Wow, that's a... pretty long list! Unfortunately almost everything here is in the domain of concurrency, which isn't my strong suit by any stretch of the imagination. I'll see what I can do!

[@pwrsoft]

That is actually a Map-Reduce Principle!

[@NERDfirst]

Hello and thank you for your comment! Map-reduce is certainly one way to do distributed computing, though not the only way =)

[@rafaelruales6871]

thanks dude

[@NERDfirst]

You're welcome! Glad to be of help =)

[@TheManifestationnnnnnnnnn]

nice video. would be more better with diagrams.. My visual memory works more significantly than verbal.

[@NERDfirst]

+Ashwini Mahindrakar Hello and thank you very much for your comment. Your feedback has been noted. I try my best to add as many visual aids as possible, but because I'm a full-time student doing this on the side, I don't always have the time to create visual aids except where absolutely necessary.

[@georgewaititu]

thank you you teach great

[@NERDfirst]

You're welcome! Glad you found the video useful =)

[@roshandahane464]

thanks

[@NERDfirst]

You're welcome! Happy to be of help =)

[@sitesh1613]

amazing

[@NERDfirst]

Thank you! Glad you liked the video =)

[@mdalaminmahamud3775]

can you please upload a video of ray tracing. i know it's a huge topic. a basic understanding would be fine.

[@NERDfirst]

Md. Alamin Mahamud Thank you for your comment! I'll see what I can do - I did learn a lot about raytracing at school, but I wonder how all that can be summarized / reduced for a video. No promises, but I'll try!

[@skaniver1]

thanks ! :D

[@NERDfirst]

skaniver1 You're welcome! Glad you liked the video =)

[@KevinYap91]

+lcc0612 hye my name is kevin great explanation by the way. I need your help, my lecturer has given me a task to create a project related to distributed system, im interested in doing a project regarding the distribution of task for a 3d graphic modeling or something similar to it. Im still a newbie in DS, so how do i actually make it work do i need to create a certain program related to parrallel programming to distribute the task from the main PC to the other or, there is a software or a command i can use. My lecturer said make it simple but focus on DS ... please help much appreciated.

[@NERDfirst]

+KVinTv Hello and thank you for your comment! Sorry for the late reply. I'm not familiar myself with actual implementations of distributed computing, but it depends heavily on what exactly you want to do. For example, the free open-source 3D program Blender has built in capabilities to render with a render farm. As far as I understand, there is GUI that can help you do this. It's not impossible to program it yourself either. As long as you set up a reasonable API for both client and server, it shouldn't be too hard to perform execution of code remotely, and have the clients send the result back after computation.

[@KevinYap91]

+lcc0612 ooo i see, so the Blender 3d software could help me in splitting the 3d rendering task to other pc using GUI eh thnx a lot ill try it ... do you have any idea for a simple distributed system project ... plus do you know any good software or OS for load/bandwidth balancing ...

[@NERDfirst]

+KVinTv Yes! Blender is a fully featured 3D program. You may have to model and set up a scene within the program first, before you can get the render to begin. Media is probably the best thing to do for distributed computing, be it rendering or video encoding. In the worst case, take a large abstract problem (eg. Sorting, or some graph problem like Travelling Salesman), find a way to break it down into parallelizable chunks, and have each computer work on one of the chunks, reporting the result back. Sorry, but I have no experience or knowledge of load balancing.

[@solesolver5740]

the first 4 digit of your channel name, is it your ATM password?

[@NERDfirst]

Heh, you wish!

[@amitkakade3335]

hmm nice video.i like it.

[@NERDfirst]

+Amit Kakade Hello and thank you very much for your comment! Glad you liked the video =)

[@alhassanadamu6979]

wow interesting

[@NERDfirst]

Thank you very much! Hope you found the video useful =)

[@obiaozukelechi1568]

cool.

[@NERDfirst]

Thank you very much =)

[@koolguy5248]

Who's here for ITCE course assingment???

[@Rohan2127]

do my exam for me please 👍

[@evadou1076]

it will be better to do it with image when doing the explaining.

[@NERDfirst]

Hello and thank you for your comment! I'm sorry you've had a bad experience! More visual aids would have certainly been helpful, though unfortunately back then, the focus was on turn-around time, so we had to reduce complexity. Do check out the newer videos, hopefully they're easier to learn from =)

[@97ynoT91]

#@(k the Algorithm!

[@elimargarabillo6216]

999 like 69 dislikes and 69 comments

[@itachiujumaki8047]

Thanks dude

[@NERDfirst]

You're welcome! Glad to be of help :)