Using Recursion in ARM Assembly to compute the Fibonacci Sequence

In this video, we implement a recursive algorithm in pure assembly. We write a method in ARMv7 assembly code to compute nth term of the Fibonacci sequence.

---

Timestamp:
00:00 Intro
03:01 Base case vs Recursive Case
03:49 Base Case Overview
04:31 Recursive Case Overview
05:51 Filling out Base Case
07:05 Filling out Recursive Case
08:51 Saving (n-1) to the stack
10:17 Calculating Fibonacci (n-2)
11:49 Sum of calculated values
12:43 Assembling Program
14:05 Visualizing Output

---

Links Mentioned in Video:

---
laurieWIRED Twitter:

laurieWIRED Website:

laurieWIRED Github:

laurieWIRED HN:

laurieWIRED Reddit:

[CJHF]

The quality of the video editing just keeps going up! Great stuff

[adailtonjn]

It's really nice to see projects written in assembly. More videos like this, for ARM and RISC-V, , are very welcome

[ankk]

I can never put this crazy efforts on aesthetics.
Pure respect for that!

[platypii]

Incredible editing, love the aesthetic.

[JulietteLloyd-ro]

Thank you SO MUCH for this tutorial. You have SAVED my life!

[ntcarver]

The retro aesthetics of your videos are so cool. I look forward to watching more!

[jonathanalonso]

I have arrived to this video by the power of the algorithm gods, and _my god_ the production value is off the charts!
Subscribed _immediately._

[polaris_babylon]

I computed Fibonacci Sequence in MIPS Assembly, pretty cool.

Btw great channel, the level of production and aesthetic is amazing!

[JTMartineau]

This was great. Thank you so much. :) I also really appreciate the nice animations you've made. They're veeeery clean and nice looking. I love the clicks.

[guillermogarciamanjarrez]

Damn loved this, got PTSD from when I had to write assembly code for a PIC32 in university, suffered so much to make a program to calculate 2's exponentials and displaying it in 8 segments display, writing the delay function just killed me. Nice remembering those times, now my brain can´t function if I´m not using C# or python XD

[stalecu]

I'm genuinely impressed at the gold mine that is this channel, you should be having at least 10x your current sub count.

By the way, will you continue your RISC-V assembly series?

[MeloettaMarmalade]

The SEL intro was an instant subscribe

[flymetothemoon]

I'd love to know more about how you create and edit these videos, I love the style and animations. great stuff!

[ah_jay]

Love the Serial Experiments Lain vibes
😊

[mfreeman]

your videos are the best.. even if i don't understand shit

[CandyHam]

How do you only have 17, 000 subscribers.... That'll probably change quickly

[chitlitlah]

Correct me if I'm wrong, but when you call fib(16), it has to call fib(15) and fib(14) to return the result, and calling fib(15) means calling fib(14) and fib(13), and both times you call fib(14), it calls fib(13) and fib(12), etc. So to do fib(n), it has to call that routine 1 + 2 + 3 + ... + n times in total. It works, but that's going to get pretty intense for big numbers, which is probably why even just fib(16) took a while to calculate. It's just for teaching purposes so I'm not criticizing, but it's definitely not the most efficient way to do it. It kind of makes me want to convert it to x86 assembly to see how quickly it gets out of control on my computer.

[AdventuresOfPepero]

coquetry + low level programming passion!

I never saw this combination

[DavidConnerCodeaholic]

my first instinct was to use a cycle of subroutines each using 2 registers to retrieve and a third to store. The setup writes to A & B, then calls the first sub, which reads A & B, writes to C, then calls the second sub. That reads B & C, writes to A. Each subroutine needs to check N, and either return or decrement, calculate and call next sub. This is more efficient.

[xorlop]

where did this awesome person come from