Update best_practices.md

This commit is contained in:
Yunsong Wang
2025-09-19 12:55:40 -07:00
committed by GitHub
parent 0b25b91223
commit df27dcb5df

View File

@@ -1,6 +1,6 @@
# NVBench Best Practices
NVBench is a **small yet actively developed benchmarking library** for CUDA GPU workloads. The [README](https://github.com/NVIDIA/cuCollections/blob/dev/README.md) serves as the ideal starting point, providing detailed guidance for users to get hands-on experience—from installation to framework usage. It includes links to the [benchmark documentation](https://github.com/NVIDIA/nvbench/blob/main/docs/benchmarks.md), which covers all essential features and usage instructions, as well as links to [code examples](https://github.com/NVIDIA/nvbench/tree/main/examples) that demonstrate how to integrate NVBench into a users codebase.
NVBench is a **small yet actively developed benchmarking library** for CUDA GPU workloads. The [README](https://github.com/NVIDIA/cuCollections/blob/dev/README.md) serves as the ideal starting point, providing detailed guidance for users to get hands-on experience—from installation to framework usage. It includes links to the [benchmark documentation](https://github.com/NVIDIA/nvbench/blob/main/docs/benchmarks.md), which covers all essential features and usage instructions, as well as links to [code examples](https://github.com/NVIDIA/nvbench/tree/main/examples) that demonstrate how to integrate and apply various NVBench features within a users codebase.
This document is **not intended to replace** the detailed benchmark documentation ([here](https://github.com/NVIDIA/nvbench/blob/main/docs/benchmarks.md)) or the CLI help guides ([CLI help](https://github.com/NVIDIA/nvbench/blob/main/docs/cli_help.md) and [CLI axis help](https://github.com/NVIDIA/nvbench/blob/main/docs/cli_help_axis.md)). All examples provided are for demonstration purposes and are **not intended as recommendations for best practices in real-world scenarios**.