RSS Anyway logoRSS Anyway

ingero.io

ingero.io
RSSAtomJSON
  1. 01.
    Agent + MCP + eBPF: 10,869 CUDA Kernel Events, Now Queryable
    ▲ 0 · ingero.io · 1d ago
  2. 02.
    What Happens When an AI Agent Gets Kernel-Level GPU Traces
    ▲ 0 · ingero.io · 2d ago
  3. 03.
    MCP as Observability Interface: Connecting AI Agents to Kernel Tracepoints
    ▲ 34 · ingero.io · 4d ago · 13 comments on HN
  4. 04.
    One Query, Four GPUs: Tracing a Distributed Training Stall Across Nodes
    ▲ 0 · ingero.io · 9d ago
  5. 05.
    CUDA Graphs: The 8-Year Overnight Success and the Observability Gap
    ▲ 0 · ingero.io · 11d ago
  6. 06.
    124x Slower: What PyTorch DataLoader Actually Does at the Kernel Level
    ▲ 0 · ingero.io · 17d ago
  7. 07.
    Tracing a 13x PyTorch Slowdown to a Hidden NumPy Synchronization
    ▲ 0 · ingero.io · 18d ago
  8. 08.
    Tracing torch.cuda.empty_cache() on an RTX 4090 – Where Do the 53 MB Go?
    ▲ 0 · ingero.io · 24d ago
  9. 09.
    Catching a vLLM Latency Spike with eBPF and an Open-Weight LLM
    ▲ 0 · ingero.io · 25d ago
  10. 10.
    GPU Incident at 3am: eBPF Tracing from Page to Root Cause in 60 Seconds
    ▲ 0 · ingero.io · 1mo ago
  11. 11.
    nvidia-smi Reports 97% Utilization While the GPU Sits Idle
    ▲ 0 · ingero.io · 1mo ago