Teaching

2026/27, 2025/26: Computer Architecture is a first-year course on how computers work, from the representation of information to the execution of programs by the hardware. It covers digital representation of data (integers and their operations, floating point, text and Unicode), C programming, the RISC-V instruction set and assembly language (including procedure calls and the call stack), and the translation from C to assembly and machine code. It then turns to processor organization (the datapath, pipelining and its hazards), the memory hierarchy (caches and virtual memory), and input/output. In the practical classes, students write weekly exercises and programming projects in C and RISC-V assembly. Assessment combines regular Moodle quizzes, the projects and tests about them, and written theory tests. The course aims to give students a solid understanding of the hardware/software interface and of how architecture affects program behaviour and performance.

2026/27, 2025/26: Lecturing a module on Concurrency.

2026/27: Concurrent Systems covers the principles and practice of shared-memory multiprocessor programming, based on The Art of Multiprocessor Programming. The theory runs from mutual exclusion and the correctness and progress of concurrent objects (linearizability, lock- and wait-freedom), through consensus and the relative power of synchronization primitives, to practical techniques: spin locks and contention, fine-grained and lock-free data structures, data races and deadlocks, and transactional memory. In a semester-long group project, students design, implement, test and evaluate a concurrent system, from lock-free data structures and synchronization primitives to memory reclamation and analysis tools. They develop it in checkpointed phases, review each other’s work, present it publicly, and defend it individually. The course aims to make students able to reason rigorously about concurrent algorithms and to build, measure and justify efficient, correct concurrent software.

2025/26, 2024/25: Concurrent Systems covers the principles and practice of shared-memory multiprocessor programming, based on The Art of Multiprocessor Programming. The theory runs from mutual exclusion and the correctness and progress of concurrent objects (linearizability, lock- and wait-freedom), through consensus and the relative power of synchronization primitives, to practical techniques: spin locks and contention, fine-grained and lock-free data structures, data races and deadlocks, and transactional memory. In the practical classes, students solve exercises that put these concepts into practice and develop a project in groups of two: implementing and evaluating a concurrent algorithm or system. The course aims to make students able to reason rigorously about concurrent algorithms and to build correct, efficient concurrent software.

Co-lectured with José Alferes and João Leitão.