Everyone has a desire to customize their tools, and there is no exception when it comes to development setups. One might write flawless, beautiful code using VSCode, but the joy and experience of customizing your coding setup is what led me into this endless journey. Neovim The idea of Vim/Neovim can be jarring for many people; navigating through code with only your keyboard seems extremely inefficient. However, the reality is quite the opposite. Read more...

Navigating the Challenges of #![no-std] in Eerie The quest for #![no-std] support in Eerie has proven to be more challenging than anticipated, particularly when delving into embedded development nuances that were not initially apparent. Unraveling Linker errors Rust brings many advantages, one of which is freedom from dealing with traditional C/C++ linker errors. However, linking Rust and C in an embedded environment has presented a unique set of challenges. A particularly vexing issue arose when encountering a linker error that looked like this: Read more...

How did I encounter this problem? While building Eerie, a Rust binding for IREE, I encountered a seemingly precarious aspect of a C API: // Creates a new session forced to use the given |device|. // This bypasses any device enumeration performed by the loaded modules but // the loaded modules will still verify that the device matches their // requirements. // // A base set of modules may be added by the runtime during creation based on // |options| and users may load additional modules - such as the one containing // their user code - by using the iree_vm_context_t provided by // iree_runtime_session_context. Read more...

What is Eerie? Eerie is a Rust binding for IREE, an end-to-end compiler and runtime based on MLIR (Multi-Level Intermediate Representation). It serves as a transformative bridge for Machine Learning (ML) models, converting them into a unified Intermediate Representation (IR). For those unfamiliar with the ML compiler space, this process involves exporting models from frontend frameworks such as PyTorch, TensorFlow, and JAX into MLIR as a series of tensor operations. Subsequently, these models are compiled into IREE’s own IR. Read more...

For passionate side project enthusiasts, documenting and sharing your projects can be a challenging task. Often, projects become a tangled mess with undocumented code, forgotten execution procedures, and hidden gems of experience buried in a labyrinthine GitHub repository. This isn’t an ideal scenario for showcasing your hard work and lessons learned. Why Choose a Custom Blog Setup? While there are popular blogging platforms like Medium, Blogger, or Wordpress, opting for a custom setup provides total control and a valuable learning experience. Read more...