S

SIMD

The NAStJA framework provides an easy way to enable massively parallel simulations for a wide range of multi-physics applications based on stencil algorithms.

High-performance cryptographic suite featuring massive parallel execution of Secp256k1, SHA-256, and RIPEMD-160 using NVIDIA CUDA (GPU) and AVX2 (CPU). Includes a standalone CashAddr library.

🚀 IronFox browser optimized for ARM/Snapdragon processors

• NEON SIMD instructions for 25-40% faster text processing • Direct GPU text rendering bypassing Android stack (40-60% faster) • Mobile-optimized DOM algorithms (20-35% faster) • VP9 video decoding optimizations (15-25% faster) • jQuery-like performance enhancements

Targets: Qualcomm Snapdragon, ARM Cortex-A78/A77/A76, OpenGL ES 3.0+

An attempt to implement SSE/AVX instructions in Python, mainly for fun.

A NAStJA Application

Provide a good solution to the CFM 2020 Challenge the hard way (hand writing the optimizer, directly use the Linear Algebra routines, use the most out of the multicore hardware).

Image filters, implemented in Assembly with SSE instructions for multiple data processing.

Academic Project.