Projects
The followings are collection of my research and personal projects. You can click on them to see the details. If you are interested in one of those, just click on the below email address to contact me.
CosyWake
A WatchOS smart alarm clock app
Human sleep is structured in cycles, primarily comprising two main stages: light sleep and deep sleep. CosyWake uses advanced sensors and algorithms to track sleep stages accurately. By identifying light sleep, it wakes you at the ideal time, ensuring a smoother, more refreshing start to your day.
Efficient Cuda SDOT Kernel
The fastest CUDA SDOT product kernel.
This project demonstrates an optimised reduction method in parallel computing to apply the dot product to vector x and y. By testing, this GPU-optimized kernel boosts the efficiency of the SDOT operation by minimizing data movement, which runs faster than the cuBLAS SDOT kernel and the traditional sequential reduction method.
