Skill UI

This skill provides comprehensive monitoring capabilities for GPU instances running on Vast.ai. It collects key metrics including GPU utilization, instance uptime, temperature, and accumulated costs. Use this when setting up operational monitoring dashboards, configuring critical alerts (e.g., idle GPUs, high temperature, budget overruns), or automating resource usage tracking in a cloud computing environment.

This assistant provides comprehensive performance profiling for complex computing systems. It analyzes various targets, including Python scripts, running processes, GPU utilization, and memory bottlenecks. Utilizing both external tools (like nvidia-smi, cProfile) and advanced code instrumentation, it measures CPU, memory, and interconnect performance, generating a structured report with actionable insights and a mandatory changelog of all modifications.

This skill provides comprehensive guidance for deploying and managing GPU-accelerated AI inference services on CoreWeave Kubernetes (CKS). It covers best practices including containerization using NVIDIA CUDA base images, configuring specific GPU resource limits (A100/H100), setting up robust health checks, and executing controlled rolling updates. Ideal for managing multi-model inference and scaling demanding AI workloads in a cloud environment.

This runbook provides structured steps for responding to critical production incidents on the CoreWeave platform. Use it when dealing with GPU workload failures, inference service outages, or general Kubernetes resource issues. It guides users through checking pod status, node health, and diagnosing common model loading errors to ensure rapid service recovery.

A comprehensive solution for monitoring CoreWeave cluster events and GPU workload status using webhooks. It tracks critical lifecycle events such as pod readiness, job completion, volume attachment, and node health. Essential for building automated scaling, alerting, and recovery pipelines for large-scale GPU-accelerated inference and training workloads.

A powerful, kubectl/docker-style CLI tool designed to manage and submit distributed computing jobs on the Qizhi (启智) platform. Use it for tasks like logging in, checking resource availability, discovering workspaces, and submitting complex GPU, CPU, or batch jobs efficiently.

A comprehensive guide for positioning research manuscripts for the ACM SIGGRAPH I3D Symposium. It covers core topics like real-time rendering, interactive 3D systems, and GPU algorithms, advising authors on how to frame their work, structure evidence, and compare it against related academic venues to maximize submission success.

A comprehensive project management guide for the MLSys conference submission cycle. It covers the entire research lifecycle, from initial venue selection and track choice through paper submission, review responses, artifact evaluation, and final conference presentation. The guide emphasizes backward planning, managing critical resources like shared GPU time, and mitigating common failure modes.

This guide outlines the necessary steps to ensure that parallel computing performance papers are scientifically reproducible. Since parallel results depend heavily on machine architecture, topology, and runtime conditions, proper documentation must cover everything from CPU/GPU specifications and interconnects to thread pinning policies, random seeds, and compiler flags. The focus is on proving the scaling trend, not just the absolute numbers, allowing results to hold true across different hardware platforms.

Provision, configure, and monitor NVIDIA GPU servers for AI inference and training workloads, including driver/CUDA installation, Docker GPU support, MIG partitioning, and health monitoring.