Lab

Experiments

Small experiments, benchmarks and technical tests from projects I've worked on.

Interactive experiment / 005

Try the adaptive states.

Switch between the three rule states used in the prototype and see what the project actually measured. I keep the claims tied to this benchmark instead of turning them into universal rules.

LOW_POWER_FASTEnergy + throughput

Low power / fast

This state favors situations where processing cost and energy matter most.

SPECK was about 5.39x faster than the AES software baseline

SPECK measured energy was about 75% lower than the AES software baseline

These numbers come from this project benchmark

Benchmark set

SPECK / SIMON / ASCON / AES software baseline

Row-level dataset benchmark pipeline used in the adaptive encryption IoT project

001

Adaptive Benchmark Pipeline

The pipeline I used to move from public and local sensor data into ESP32 measurements, algorithm profiles and the adaptive rules used in the project.

Time and energy comparison from the IoT benchmark

002

Time / Energy Trade-off

A project-specific comparison of computation time and measured energy. I wanted the decision to come from the actual benchmark, not from a generic assumption about the algorithms.

Measured throughput comparison from the adaptive encryption IoT benchmark

003

Throughput Benchmark

The throughput comparison for SPECK, SIMON, ASCON and the AES software baseline from the same project benchmark.

004

Searchable PDF Pipeline

An OCR workflow that finds text regions, recognizes the content and places searchable text back at the matching coordinates in the document.