I’ve published Inside oneTBB, a project exploring the software infrastructure behind multithreaded execution. My goal is to understand how the scheduler organizes and distributes work, then experiment with applying those mechanisms to CPU simulation.

I’ve published an interactive study of the IBM SP2, exploring the path from CPU microarchitecture to a complete supercomputer cluster. My goal is to connect these different scales: how a POWER2 processor works, how it fits into a compute node, and how many nodes communicate and work together.

I’ve published Coffee Lake, a project exploring how a CPU executes multithreaded workloads. My goal is to understand that execution through modeling and simulation, connecting what happens inside each core with how multiple cores interact through caches and shared memory.

I have finished a prototype of a beast application tool, but feel uncertain if I have the passion to keep improving that.

I have finalized the basic idea of project Xe-Diagram – aiming to show case a project that can visualize public architecture view. I am very satisfied on the current status. With sub pages reworked, reviewed and published, I think my project can sit better than the public Programmer’s Reference Manuals and can serve better as a knowledge base like MCP server for Intel post-si GPUs.

Glad to have published my vibe Xe-Diagram

Link: https://xwang186.com/xe-diagram/ (also can be accessed thru blog menu)

Architecture: Xe-HPG(A770)

Referred docs: Intel® Graphics for Linux* – Programmer’s Reference Manuals

(6/16/2026 update: seems the orginal link https://www.x.org/docs/intel/ACM/ — become unavailable now)

5/19/2026 – Finished! Glad to have included all my initial thoughts into the plugin!

I have been spent quite long time with ChatGPT, on different topics. I am the 1% top user , without using any token api but just chat on web/app.

I have found it less meaningful to summarize “domain knowledge” into articles. This is a waste of time because AI agent can do 20x better than me.

By Train from SAC

Introduction to NVDLA

The NVIDIA Deep Learning Accelerator (NVDLA) is an open-source hardware project designed to accelerate deep learning inference tasks. Its flexible, scalable architecture enables developers to implement AI acceleration on various platforms, from FPGAs to ASICs. By offering a complete stack of hardware and software, NVDLA has become a cornerstone for many projects in the AI hardware space, making it an invaluable tool for prototyping and development.