Crossware Contact Us
Platform engineering

From first power-on
to production-ready.

Crossware has delivered working embedded software across application processors, automotive MCUs and AI accelerators. Board bring-up, BSP development, GPU enablement and memory-optimised production builds — on the hardware your product actually ships on.

Board bring-up BSP & device tree GPU acceleration Memory optimisation Yocto / OpenEmbedded
What Crossware does on your platform

The engineering that makes hardware ship software.

01

Board Bring-Up

First power-on through a bootable, verified system. Crossware validates hardware, brings up peripherals and establishes a stable foundation for software development.

  • Power sequencing and reset validation
  • UART, SPI, I²C, USB, Ethernet bring-up
  • Display and camera interface validation
  • Debug infrastructure setup (JTAG, OpenOCD)
  • Bootloader configuration (U-Boot, SPL)
02

BSP Development

Linux and RTOS board support packages built for the hardware, not ported from a generic reference. Device trees, kernel drivers and platform integration that reflects your actual board design.

  • Linux device tree authoring and maintenance
  • Kernel driver development and upstreaming
  • Yocto / OpenEmbedded layer creation
  • FreeRTOS and bare-metal BSP for MCUs
  • Boot-time and suspend/resume optimisation
03

GPU Acceleration

Enabling hardware GPU compositing and OpenGL ES / Vulkan rendering for HMI and multimedia workloads. Crossware configures the full display pipeline — from DRM/KMS through Qt's platform plugin to the application layer.

  • DRM / KMS display pipeline configuration
  • OpenGL ES and Vulkan driver bring-up
  • Qt EGLFS and Wayland platform plugin setup
  • Hardware compositor integration (ivi-compositor)
  • Benchmark-driven GPU performance validation
04

Memory & Build Optimisation

Reducing flash and RAM footprint to fit production memory budgets. Crossware profiles, strips and restructures builds to meet the constraints of resource-limited embedded platforms without compromising function.

  • Yocto image size profiling and reduction
  • Qt module selection and dead-code stripping
  • Shared library and symbol optimisation
  • RAM footprint reduction and heap analysis
  • OTA-friendly partition and image design
Hardware platforms

Delivered across the platforms your product ships on.

Application Processors · Linux / Android
NXP
i.MX8M Plus
Cortex-A53 · NPU 2.3 TOPS · i.MX8M Plus
Yocto Linux Qt GStreamer Edge AI WebRTC
Work done
  • CrossPlay multimedia stack deployment
  • DMS inference on integrated NPU
  • GPU-accelerated Qt HMI at 60fps
  • Yocto BSP and eIQ ML integration
NXP
i.MX95
Cortex-A55 · Enhanced NPU · i.MX95
Yocto Linux Android Qt Edge AI
Work done
  • Sign recognition vision pipeline
  • Automotive-grade DMS integration
  • Android IVI bring-up and HMI
  • Board bring-up and BSP development
ST Microelectronics
STM32MP1 / MP2
Cortex-A7 / A35 · Linux-capable MPU
Yocto Linux Qt OpenGL ES
Work done
  • Linux BSP and display pipeline
  • Qt EGLFS GPU-accelerated HMI
  • Industrial HMI applications
Automotive MCUs · Real-time / Graphics
Infineon
TRAVEO™ T2G
Dual Cortex-M7 · 2D GPU · Automotive MCU
Qt for MCUs FreeRTOS Bare-metal AUTOSAR
Work done
  • Qt for MCUs instrument cluster (10+ SOPs)
  • 2D GPU driver integration and optimisation
  • Memory-optimised QUL runtime builds
  • Infineon Preferred Design House engagement
Renesas
RH850
Multi-core · Automotive real-time MCU
Bare-metal AUTOSAR FreeRTOS
Work done
  • AUTOSAR integration and BSP
  • Real-time safety-critical firmware
  • Automotive cluster and body control
NXP
RT1176
Dual Cortex-M7/M4 · Crossover MCU
Qt for MCUs FreeRTOS Bare-metal
Work done
  • Qt for MCUs HMI with GPU acceleration
  • Inter-core communication (M7 / M4)
  • Production firmware and BSP
General Purpose MCUs & Graphics
ST Microelectronics
STM32
Cortex-M · STM32H7S78 · resource-constrained embedded
Bare-metal FreeRTOS Azure RTOS Qt for MCUs LVGL
Work done
  • Azure RTOS porting on STM32H7S78
  • Resource-constrained HMI with Qt for MCUs
  • Industrial panel and control interfaces
  • Memory-optimised production builds
Nuvoton
Nuvoton MCU
Arm Cortex-M · Embedded MCU platforms
Bare-metal FreeRTOS LVGL
Work done
  • Embedded HMI firmware
  • BSP and peripheral driver development
  • Industrial control panel applications
AXELL
AG903
Graphics-focused MCU
Graphics HMI Bare-metal
Work done
  • GPU pipeline and display bring-up
  • Graphics-accelerated HMI development
AI Accelerators · NPU / Edge Inference
DeepX
DX-M1 NPU
Dedicated edge-AI accelerator
Driver integration Model deployment Vision pipeline
Work done
  • NPU driver and runtime integration
  • DMS and people counting deployment
  • Co-processor integration alongside i.MX SoC
  • INT8 model quantisation and optimisation
NXP eIQ
i.MX Integrated NPU
NPU inside i.MX8M Plus & i.MX95
eIQ ML stack TensorFlow Lite ONNX
Work done
  • eIQ toolchain integration in Yocto
  • TFLite and ONNX model deployment
  • DMS inference on-SoC alongside HMI
Software & operating systems
Embedded Linux Yocto / OpenEmbedded Android (AOSP) FreeRTOS Azure RTOS Bare-metal C/C++ Qt Qt for MCUs LVGL U-Boot OpenGL ES Vulkan Wayland GStreamer WebRTC AUTOSAR
Start with your hardware

Tell us the board. We'll take it to production.

Share your SoC, memory constraints and target software stack. Crossware will propose the right bring-up, BSP and software architecture for your platform and timeline.

Discuss your platform