Portfolio

How we helped others in the past to make their projects a success.

We have extensive experience with different customers, project sizes, hardware, software, architecture, and algorithms. In the following you can find some projects we successfully completed as a team at our previous employer.

Runtime System Memory DSP Porting Emulation AVX Vectorization AI NPU ARM Neon

Runtime System

Overall System Optimization

System-level architectural optimizations and optimization of most expensive applications to reduce resource usage by one CPU cluster.

Runtime Memory DSP Porting

Radar Perception on DSP

Helping the customer to pick the right SoC and porting critical core software to DSP.

Runtime System

Efficient Middleware

Analysis and optimization of middleware with 2× speedup in typical use cases.

Memory

RAM and ROM Reduction For Critical Cost Savings

Reduced ROM consumption by ⅓, several hundred of kilobytes of RAM saved on MCU

Runtime System

Faster Mutexes

Optimized mutexes 2×, resulting in less kernel contention.

DSP Emulation AVX Vectorization

C6000 DSP Support Libraries

Real-time resimulation of C6000™ DSP code on x86-64 and bit-identical math functions.

Runtime AI NPU Vectorization

Custom CNN Inference Layers on Hexagon NPU

Optimized custom CNN inference layer 30× on Hexagon™ DSP.

Runtime ARM Neon Vectorization

Feature Detection for Classical Computer Vision

Speedup of core feature detection algorithm by 7× on Arm® Cortex®-A76.

Runtime AI ARM

Faster Convolutional Layers for Embedded CNNs

Speedup of convolutional layers in CMSIS 5 by 3×.