All skills
parallelism-planning
Compares data, tensor, sequence, context, pipeline and expert parallelism by capacity and communication cost.
- Part of bojieli/ai-infra-book
- Machine Learning · GPU acceleration
More GPU acceleration skills
See all- cudaq-guide2.5kGuides CUDA-Q setup, simulation targets, QPU access, and writing quantum kernels.
- compute-sanitizer-diagnose-cuda-bugs105Finds CUDA memory errors, data races, uninitialized reads, and sync bugs and explains how to read the report.
- flash-attention7Speeds up training and inference for long-sequence transformer models.
Also in this pack
- accelerator-selectionCompares accelerators on capacity, effective compute, memory bandwidth, power and cost for a given model workload.
- agent-runtime-capacityPlans how agent model calls and tool environments are created, kept, released and restored for capacity and latency.
- cluster-network-analysisEstimates cross-server traffic, network bottlenecks and sync waits for multi-node training or inference.
- distributed-inferenceCompares cross-instance inference options like replicas, prefill/decode split, expert placement and shared KV cache.