A structured visual guide to the major mathematical areas and their relationships.
Search by code, branch, topic, subtopic, or a keyword from the descriptions.
This subtopic introduces the core ideas in parallel and high-performance computing, including foundational concepts, standard methods, and the main questions used to organize the area. Typical uses include building mathematical background, framing related research problems, and supporting applications in neighboring fields where these concepts provide useful structure.
This topic studies how numerical algorithms can be decomposed and executed on parallel hardware to improve throughput and scalability. It addresses communication, load balancing, and synchronization in large-scale scientific computation.
Parallel computing (Wikipedia)
Used in large-scale simulation, PDE solvers, and numerical linear algebra where time-to-solution is critical.
This topic examines algorithms tailored to particular hardware architectures such as GPUs, vector processors, or distributed-memory systems. The emphasis is on matching mathematical structure to the communication and execution model of the machine.
Parallel computing (Wikipedia)
Used when numerical software must exploit GPUs, multicore processors, or distributed clusters efficiently.
This topic concerns numerical software packages and libraries that package standard algorithms for solving scientific and engineering problems. It emphasizes reproducibility, robustness, and ease of use when implementing sophisticated numerical methods.
Library (computing) (Wikipedia)
Used when researchers and engineers need reliable, tested numerical routines without re-implementing the full algorithm stack.
This topic studies the theoretical and empirical cost of numerical algorithms, including operation counts, memory traffic, scaling, and practical bottlenecks. It connects mathematical design with computational efficiency on real hardware.
Algorithmic complexity (Wikipedia)
Used to compare algorithms, predict scaling, and choose suitable methods for large-scale numerical computation.