GPU Articles
Android Transformer Optimization: Attention and Mobile GPU Bottlenecks
A walkthrough of porting FlashAttention into an OpenGL ES Compute Shader on Android, using tiled computation and Online Softmax to cut attention memory traffic from O(n²) to O(n). Measured on the Mali-G710, latency dropped from 120ms to 18ms.
Read Post
Android Bitmap Memory Model: From Java Heap to Hardware Bitmap
A deep dive into how Android Bitmap pixel storage moved between native heap, Java heap, and GPU memory, and what that means for OOM prevention.
Read Post