Ten Things Every Computing Student Should Master¶
If you are comfortable with these ten ideas, you understand computers better than most beginners — and you are ready for systems, architecture, HPC or cloud work.
1. Think in binary and hex¶
Convert numbers, read two's complement, and know why 0.1 + 0.2 is not exactly 0.3 in floating point.
2. Trace a program from source to silicon¶
Compiler, linker, loader, then the fetch-decode-execute cycle on the CPU.
3. Explain the memory hierarchy¶
Registers, caches, DRAM and storage trade size for speed — and locality is why caches work.
4. Understand how modern CPUs go fast¶
Pipelines, superscalar and out-of-order execution, branch prediction, SIMD and multiple cores.
5. Know when a GPU or accelerator wins¶
Thousands of simple parallel lanes beat a few smart cores on matrix maths and AI.
6. Reason about performance¶
Use latency vs throughput, Amdahl's law and the roofline idea before you optimise anything.
7. Know what the operating system does¶
Processes, threads, scheduling, virtual memory, system calls, drivers and the boot chain.
8. Understand virtualisation and containers¶
Hypervisors run whole machines; containers share one kernel — and Kubernetes orchestrates them.
9. Scale out with parallel and distributed computing¶
OpenMP, MPI, Slurm, data/tensor/pipeline parallelism and the CAP theorem — and design for failure with ECC, checkpoints and health checks.
10. Measure, don't guess¶
Benchmark and profile with tools like perf, htop, nvidia-smi and Nsight — then fix the real bottleneck.