C/C++ KV-cache optimization experiments for LLM inference. Live vLLM eviction and admission, H100/H200 benchmarks, and decode-aware llama.cpp transfers.
API documentation, OpenAPI specifications, and integration templates for Kilawatt Cloud.
Harness and processed results for measuring per-layer isolation on a dedicated H100 GPU Droplet: seven-day observation, induced contention, self-contention, MIG and MPS.
Reproducible H100 GPU-seconds/request study for 91 pinned Command A+ vLLM runs, with held-out evaluation, coverage diagnostics, and independent verification.
GPU checkpoint/restore for confidential LLM inference on NVIDIA H100 and Intel TDX. ACM SIGOPS ATC 2026.
Verifiable, hardware-attested Confidential Virtual Machine (CVM) runtime for private AI inference on NVIDIA Hopper GPUs
Interactive, scroll-driven visualization of NVIDIA H100 architecture, connecting GPU hardware with CUDA, Triton, and PyTorch programming concepts.
Real-Time LLM Request Scheduling Engine & GPU Telemetry Observability Platform. Simulates PagedAttention, Chunked Prefill, and Priority Preemption (vLLM / TGI) on virtualized NVIDIA H100 GPUs.
To associate your repository with the nvidia-h100 topic, visit your repo's landing page and select "manage topics."