HAMi

Term 3 · Sep-Nov · 4 projects

• Fix GPU Memory Isolation for Child and SSH Processes

  LFX Mentorship · Listed in the official cohort

  @archlitchiSource IssueLFX portal

  Investigate and fix cases where processes started later inside a HAMi-managed container, including child processes and processes reached through SSH or a new login shell, may not retain the expected GPU memory limit. The first milestone is to reproduce the behavior on a current supported HAMi release. The implementation approach must be agreed with HAMi maintainers rather than assuming a specific configuration mechanism in advance.

  • Linux debugging
  • container runtime analysis
  • security reasoning
  • C/C++
  • testing

• HAMi GPU Observability: Metrics and Dashboards

  LFX Mentorship · Listed in the official cohort

  @mesutoezdilSource IssueLFX portal

  Improve the usefulness, consistency, and documentation of HAMi GPU observability for operators running shared GPU workloads. The program will focus on practical metrics and dashboards, while keeping tracing work bounded to a design and proof of concept for one mentor-approved control-plane path.

  • Metrics instrumentation
  • PromQL
  • dashboard design
  • observability
  • Kubernetes debugging
  • HAMi internals

• HAMi GPU Sharing Workshop and Documentation

  LFX Mentorship · Listed in the official cohort

  @fishmanSource IssueLFX portal

  Create practical, reproducible learning materials that help users and new contributors deploy HAMi, understand GPU sharing, and troubleshoot common installation, scheduling, and allocation problems. The work should contribute to active HAMi documentation repositories and use a documented, supported HAMi environment.

  • Technical writing
  • documentation testing
  • troubleshooting
  • GitHub workflow

• Reduce HAMi-core Initialization Lock Contention

  LFX Mentorship · Listed in the official cohort

  @archlitchiSource IssueLFX portal

  Improve HAMi-core startup performance when hundreds of processes initialize CUDA concurrently and compete for the shared unified lock. The work should follow the maintainer discussion in the upstream issue and preserve correctness while reducing initialization latency and contention.

  • Linux concurrency
  • performance profiling
  • benchmarking
  • synchronization design
  • testing