emulators are cool

I started working hacking together a Game Boy emulator the other day and wanted to see how quickly I can get from absolute zero to something. I have achieved what I set out to do:

classic

I took the absolute minimalist route and wanted something that wasn’t abstracted away like most (good) emulators are these days. Since my goal was understanding, abstraction wasn’t a high priority.

First, I wrote the parser for the ROMs to ensure I handled all of the instructions correctly in the bootrom. When that was looking good, I had to do a bit of refactoring now to actually execute those instructions and modify my state. After about a week or so of digging in, I had the basics working. For now, I’ll be putting this project on hold because other things have captured my attention. The code is available at https://github.com/AugustKarlstedt/gameboyemu but I warn: it isn’t pretty. 🙂

I chose to use CMake for dependency management and I must say, THIS is the way to use C++ these days. Wonderful all around.

A higher-level discussion

Why do we write an emulator that gets compiled into machine code for a specific CPU? E.g. in my case I wrote my emulator in C++ and compiled it into x86 instructions using MVCC. What’s the point? Our CPUs already have instructions to do what these older CPUs did.

Why don’t we use LLVM to simply translate from e.g. Game Boy instruction set to e.g. x86 instruction set?

Luckily a smart dude already has worked on this: https://andrewkelley.me/post/jamulator.html

Going Through Old Source Code

After backing up some files, I remembered that I wanted to go through some of the old programs and code that I’ve written. It would be a fun experiment: seeing how far I’ve come, what I actually spent my time doing as a teenager, and probably get a few laughs. Well, here are some of the interesting bits I found.

The first thing I came across was a C++ program I made for my friend Jose’s 17th birthday using SFML and Bass. It consisted of a single main.cpp file and most likely completed the night before.

Jose Birthday project

I wasn’t very clever in naming projects. Something I called “Google” was actually a really cool start of a physics puzzle game. Worked on sometime in May through July 2011, it has a complete level format based on XML. The player drags a path for a ball to enter the designated “win” zone. It used Lua scripting, Box2d Physics, and SFML for rendering. Again C++.

“Google” game project

(30 minutes later…) wow, ok, I found something that I really enjoyed! It’s called ParticleSim, written in June of 2011. I’m beginning to think that the dates on the files might be incorrect, or I was just quite productive in 2011? Not sure yet. Anyhow, at the time I was trying to replicate another Facepunch forum member’s work that I found really fun. My take on it was a game where the player can place attractors and repellers and then create particles that will interact with them. I started with the math to get that all working, then added a quick feature to make the particles’ size based on the bass of the current song. I got to play around with FMOD, Lua, and Boost all in C++.

ParticleSim demo

That’s all for now! More to come later.

Long Short-Term Memory Neural Networks and Guitar Tablature

Just messing around with LSTM neural nets, I had the idea to train it on some guitar tabs and see what happened.

Data

My dataset was a collection of guitar tabs for The Who’s album, Tommy.

Some stats
Text total length: 277,155
Distinct chars : 91
Total sequences : 92,369

So yes, it’s a tiny little dataset.

Training

I’m using TFLearn’s implementation of an LSTM neural net. My setup is as follows:

Input data -> LSTM layer (512 units, returns full sequence (3d tensor)), -> dropout (50%) -> another LSTM layer (512 units, returns only output of last sequence (2d tensor)) -> dropout (50%) -> fully connected layer (softmax activation).

Results

On the first epoch with a temperature of 1.0 (meaning we’re trying to generate novel sequences, less likely to repeat the input data), we get:

MAT 140 & PGP Encryption

I was in MAT 140 – Discrete Mathematics last quarter and I found that the fundamental theorem of arithmetic is super interesting. When the professor was speaking about it, I asked if it was the base for PGP encryption and it turned out it was. I had an inkling for this, because I previously heard that PGP generates super large numbers and that’s where the security lies.

I’m kind of skeptical. Sure, with the current state of computing these cryptography keys are nearly impossible to crack, but the fact is that they still are possible to crack. I’m betting on sooner rather than later.

Well, I’ll be looking into how this works in greater detail whenever I have time.

Anyone know some places to learn more?

New Classes – Winter Quarter

Alright so I’m finally going to discuss the new courses I’m taking now.

CSC 395 – Research Colloquium – I’m pretty excited about this course. It’s completely online and every week there will be a lecture by a guest speaker on new work in areas of computing. Hopefully there are some talks relevant to my interests, but nonetheless I love TED talks, so this should be similar.

GAM 341 – Introduction to Level Design – Since I’m majoring in Computer Game Development and focusing on the programming portion of it, I figured I should also get a little more experience in the non-programming parts. We’re going to be using Unreal Engine 4, which I’m pretty stoked about!

GAM 376 – AI for Computer Games – I’ve always been interested in artificial intelligence. Actually, in the class I’ll describe next we’re discussing AI through a philosophical perspective. I’ll probably post more on this later. As for this course, I hope we don’t just use some already-implemented AI algorithms and we can actually write our own. I doubt it though.

PHL 242 – Philosophy & Technology – We’re discussing what technology actually is. Admittedly, when asked this question I thought it was going to be easy to define. I was horribly wrong. While it isn’t required, I may read the recommended material, “War in the Age of Intelligent Machines” by Manuel De Landa. We read the first chapter and thinking about how a robot historian would describe its history is something that made me extremely intrigued.

SOC 281 – Sociology of Rock Music – This is kind of weird, yes, and I know that. It fulfills some requirement that I needed to complete, so hey why not? We’re going to trace Rock Music from the beginning to now, and I’m thinking I’ll learn a lot. The professor asked in the first day of class what we knew about rock. While I do know some stuff, and I listen to rock all the time, I wrote that I had no knowledge of rock. I think this will benefit me because starting with a blank-ish slate should be fun.

Anyhow, I wasn’t going to take these classes and considered switching majors. I’ll explain why in the next post.

GPH 321 – Completed

In this class we created a ray tracer from the ground up. I by “we” I mean our professor wrote most of the code and we had to fill in certain class functions. It was pretty terrible, but overall the end result was good. Oh and we had to zip our source code and send it in, instead of simply using something like GitHub or Bitbucket. The midterm and final SUCKED, simply because it’s 2015 (well, it was 2014) and written exams for programming finals are absurd. Writing code on paper completely lacks the feel of typing it out using an IDE. Plus, come on, who doesn’t use IntelliSense, Resharper, Visual Assist X, or some kind of code-completion tool? A fool! And that’s exactly what we are when we write code by hand.

Anyhow, here are some screenshots from the course.

First we simply had to get rays working and detecting collisions.

Rays & Intersections
Rays & Intersections

Then we had to ray-trace our first object, a sphere. This is because it’s pretty simple to represent in math.

First Ray-traced Object
First Ray-traced Object

Then we had to add some shading to that sphere. I know,Γö¼├íit isn’t 100% correct. Spot the error!

Phong shading
Phong Shading

Next came shadows. These are awesome.

Ray-traced Shadows
Ray-traced Shadows

Next was reflections. Super easy thanks to how ray-tracing works.

Ray-traced Reflections
Ray-traced Reflections

Then we had to create a checkerboard pattern.

Checkerboard
Checkboard

Finally we were to implement Fresnel equations to properly simulate glass. I didn’t finish this one fully.

Ray-traced Fresnel
Ray-traced Fresnel

And that was the end of the course. This course was online only, which probably had an impact on my view of it. The course notes weren’t that great. I repeatedly found myself Googling things that were phrased in strange ways to figure out how to implement them. And variable names like V, N, n1, n2, etc. used throughout the given code made things unnecessarily complex.

Oh well, it’s over and done with!

Updates!

Obviously I haven’t posted anything in a while. But it’s time for a change – and change is good!

I am now currently in three courses that are at least semi-relevant to this blog.

  • ANI 230 – 3D Design & Modeling
  • CSC 298 – Internship
  • GPH 321 – Computer Graphics Development I

In ANI 230, we’re learning Maya 2015. Starting with low-poly and now (it’s about midterms time) we’re onto animation with keyframes. I’m particularly interested in ray tracing, because we are building our very own in C++ in GPH 321.

As for CSC 298, it basically runs alongside the internship that I already have. Here, I’ve been working with a ton of Java stuff lately interfacing with a USB HID. Through this, I’ve learned to use Maven – let me just say it’s awesome.

Anyhow, I’ll be posting some images of the stuff I’ve been working on soon.

Also, I may move this entire blog to a domain/host I pay for now. Or maybe even use GitHub Pages and Jekyll, those look fun.

Oh and I read a few Android books, great stuff. Loving Android dev.

Finished with GAM 374

Well, I chose to turn in exactly what you see in that last blog post. Everything that was required for the assignment was there and it earned me a 95%.

I am now focusing my efforts into work with a near-full-time schedule of 38 hours a week. I will be posting stuff about happenings relating to that from here on out.

I’m also working on an Android app, so there will be some Java tossed in.

And last but not least, my love for C++ is growing as I discover C++11 features. There’s a little code I wrote to test out the functionality of some of the new stuff if you read on.

Continue reading →

Player Tank!

I’ve finally done some work on this and got the player tank working. Azul doesn’t handle mouse input very well (I’m spoiled by Unity), so it’s all keyboard controls.

WASD controls the tank movement

LEFT/RIGHT control the turret

I’m thinking the up arrow will fire a shell and the down arrow will lay a landmine.

DePaul Azul Engine - Tanks!
DePaul Azul Engine – Tanks!

Screenshot taken with Greenshot.

First Day of Final Project

Today marks the first day I’ve started working on the final project for my GAM 374: Fundamentals of Game Programming I class. I’ve decided to recreate Tank – well actually, the Wii Play version called “Tanks!“.

We’re using the DePaul game engine called Azul – which I think could definitely use an engine overhaul since it’s based on the singleton design pattern. Gross.

Click to learn about Game Programming Patterns