Energy awareness for a specific program can be a difficult aspect to measure. This element is becoming increasingly important in an increasingly energy conscious world. In software, enerrgy usage is a difficult measure to quantify mainly because of the large amount of asynchronous and synchronous aspects that make up an Operating System. With advanced schedulers, it is difficult to isolate exactly how much energy can be attributed to a specific program. This is manly through the use of various techniques that buffer and consolidate data across different programs. As such, isolating the energy consumption of a specific program over it's execution can be difficult.

This project explores some of the current techniques that are available to profile energy consumption of programs. Within this project, I created a technique to profile the energy consumed by the CPU, RAM, GPU, battery and network components of a given program. This was subsequently tested across a wide range of benchmarks and energy efficient/inefficient programs that explore the effectiveness of my profiler.

I perform this comparison against a large amount of schedulers, governers and hardware along with general program comparisons between Firefox and Chrome. Further details of my program can seen in my thesis.

A supervisor at Imperial tasked us with creating a platform for the lending and borrowing of items. The key idea was that the items aren't being used by their owner all the time, and so could be lent out to others who need them. Sharify was created to bring about the closeness that neighbours in smaller towns could have, where they would simply go next door to borrow something they needed.

Sharify was developed across three platforms: a website, an Android app and an iOS app. These were serviced by a backend that handled the users, items and transactions. For this project, I was responsible for the web-app for the site. For this site, I helped develop routes using NodeJS + Express and MongoDB. I was also responsible for setting up authentication and developing a test framework using Supertest and MochaJS. Finally, I helped design the client side iteractions using twitter bootstrap and JQuery.

An example prototype video for the site can be found here. Source code and reports for this project are available on request.

WikiMaster is a quiz app using questions that are automatically generated from Wikipedia articles. In this way, the app could quiz a user on almost any topic in the world, which we feel would be rather unique.

In many ways, this project layed the foundation for our architecture and project flow that proved successful for future projects. For this project, I worked with others on the backend which was written using NodeJS + express. I also helped build the website using Bootstrap and JQuery along with design the database interactions.

This project was chosen to be demonstrated for the department's open day. Source code is available for this project on request.

This was a group project to implement Pintos under the specification listed previously. For this project, we needed to implement many aspects that are critical for Operating System designs. This includes a expansive virtual memory system that expands to multiple pages, user programs, scheduling alogirthms including priority donation and multi-level-feedback-queues. This was a full x86 Operating System that was designed to be run on QEMU.

This project tested time management, C99 skills and skills managing concurrent access.

This was a group project to implement WACC. This project was split up into multiple stages based on syntax, semantic and code generation.

For implementing this code, we used ANTLR4 for parsing. This provided a useful tree structure that was useful for traversing the code. Each of the three stages of syntax, semantic and code generation were implemented using this tree. This is because the tree enabled the use of the visitor pattern which was an elegant approach to only visiting the nodes that we care about.

As an extension to this project, we wrote a toy implementation of battleships that successfully ran on QEMU.