Skip to content
View Daviddedic2008's full-sized avatar
💭
arm64 compiler
💭
arm64 compiler
4 followers · 9 following

Block or report Daviddedic2008

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Daviddedic2008/README.md

David Dedić

Systems, Compiler, and High-Performance Computing Engineering

Focusing on low-level software architectures, spanning bare-metal embedded systems, custom compiler design, and hardware-accelerated parallel computing. Software is engineered with a focus on physical hardware limits, execution pipelines, and memory hierarchies.

Technical Stack:

  • Proficient: C, CUDA, ARM Assembly (AArch64 / Thumb-2), AVR Assembly
  • Familiar: C++, Python

Core Engineering Projects

Custom Language Toolchains and Compilers

  • Multi-Pass Subset-C to ARM64 Compiler: Architected a multi-pass compiler backend that lowers a functional subset of the C programming language into native AArch64 machine code. Features an optimization pipeline utilizing Chaitin-style graph coloring for global register allocation.
  • Self-Hostable Thumb-2 Compiler: Developed a single-pass compiler for a custom scripting language dialect. Implemented greedy register allocation integrated with loop-depth-aware spilling heuristics, lazy flag storage tracking, and register state snapshotting across conditional branch states.

Operating Systems and Bare-Metal Firmware

  • Custom Cortex-M4 Microkernel / RTOS: Developed a preemptive, bare-metal real-time operating system from scratch without a Hardware Abstraction Layer (HAL). Implemented a deterministic round-robin co-op scheduler via PendSV interrupts, low-latency mailbox IPC, hardware timer tracking, and a zero-fragmentation localized memory allocator.
  • Bare-Metal Device Drivers: Written native AVR Assembly and C drivers for LCD displays and bit-banged PS/2 keyboard matrices utilizing direct memory-mapped register manipulation and hardware interrupts.

Computer Graphics and High-Performance Simulation (HPC)

  • Global Illumination Path Tracers: Engineered two distinct ray-object intersection engines from the ground up:
    • CUDA Accelerated Path Tracer: A parallelized global illumination engine utilizing shared memory allocation and thread-coalescence techniques.
    • Library-Free C Path Tracer: A standalone, pure C ray tracer built with zero external framework dependencies, outputting raw image bytes directly via localized math routines.
  • Software Rasterization Pipelines: Developed discrete rendering pipelines to explore resource-constrained vs. massively parallel execution:
    • CUDA Parallel Rasterizer: A high-throughput, hardware-accelerated rasterization pipeline written completely in custom CUDA kernels.
    • Cortex-M4 Software Rasterizer: A highly resource-constrained software engine utilizing optimized floating-point math and frame-buffer optimizations to render geometry under strict microprocessing RAM limits.
  • CUDA Eulerian Fluid Dynamics Simulator: Developed a high-performance Computational Fluid Dynamics (CFD) simulation engine. Leveraged GPU parallelism to solve the Navier-Stokes equations natively across uniform grid structures.

Connect With Me

Fun Note: My profile picture is cool :)

Pinned Loading

  1. Cortex_M4_Compiler Cortex_M4_Compiler Public

    Simple compiler that targets & runs on the resource constrained Cortex M4. No dependencies required, emits raw binary files. Build/remake scripts included.

    C 3

  2. STM32F407xx_Renderer STM32F407xx_Renderer Public

    Realtime render pipeline for Elegoo 2.7 in. TFT module.

    C 1

  3. Cuda_RealtimeGraphicsPipeline Cuda_RealtimeGraphicsPipeline Public

    realtime render pipeline written in cuda by me. openGL is only used for final display of precalculated pixel colors.

    C 4

  4. STM32F407VGTX_Operating_System STM32F407VGTX_Operating_System Public

    Small baremetal RTOS for STM devboards in the F407VGTX category. Includes LCD/keyboard drivers, a file system, and more.

    C 1

  5. cudaCFD cudaCFD Public

    fluid dynamics sim written in cuda, uses openGL for drawing the output frames.

    C 1

  6. tinyDriverINO tinyDriverINO Public

    Driver for Elegoo UNO R3 2.8" TFT Touchscreen. Runs on Arduino UNO R3. Not cross-platform, and is based on AVR asm. Achieves slightly higher performance than stock drivers that arrive w/ the LCD unit.

    Assembly 1