01 / FIELD PROFILE

I work where software meets the machine.

I'm Hans. I work on GPU systems at a fabless semiconductor company, following the handoffs between hardware, firmware, the Linux kernel, and the runtime.

Field
GPU systems
Method
Trace boundaries
Output
Public field notes

A system in layers

REFERENCE PLATE / 01
Layered GPU execution hierarchy showing streaming multiprocessors, register files, CUDA cores, instruction units, shared memory, and device memory.
FIG. 01 / GPU EXECUTION HIERARCHYORIGINAL REDRAW AFTER SEMIANALYSIS ↗

Areas of focus

WORKING LAYERS / 04
  1. 01

    Driver and kernel integration

    Module boundaries, kernel interfaces, and resource lifecycles.

  2. 02

    Bring-up and failure analysis

    Boot stages, firmware handoffs, and locating the layer that actually failed.

  3. 03

    Memory systems

    Address translation, residency, migration, coherency, and fault handling.

  4. 04

    Systems tooling

    Tracing, crash evidence, and reproducible debugging workflows.

About this site

PUBLIC NOTEBOOK / 02

This site is a public field notebook: compact notes for mechanisms worth remembering, and longer explanations when a boundary needs to be followed end to end.

It is intentionally static, lightweight, and kept close to the source files. The views here are my own; confidential work stays private.

Elsewhere

PUBLIC LINKS