# Guides

Task guides for working on Eugo: installing Python packages, managing your Ray clusters, and more.

Source: https://docs.eugo.io/docs/guides/

---

- [Installing Python packages](https://docs.eugo.io/docs/guides/installing-python-packages.md): What Eugo preinstalls, how to check it, and how to add your own Python packages: in your own session with pip, a venv or uv, and on your Ray cluster with a runtime environment.
- [EugoHPC Manager](https://docs.eugo.io/docs/guides/hpc-cluster-manager.md): How your workspaces and Ray clusters are organized, what the Allow Tasks on Head Node setting does, and how to connect to a cluster by hand with your own runtime environment.
- [Cluster settings](https://docs.eugo.io/docs/guides/cluster-settings.md): Every setting in the Cluster Launcher when you create a Ray cluster in the EugoHPC Manager: what it does, its default and its limits.
- [Connecting a notebook to your cluster](https://docs.eugo.io/docs/guides/connecting-a-notebook.md): Connect a notebook to an Eugo cluster with the EugoHPC Manager, check that it is connected, disconnect, and run your own Python scripts on the cluster.
- [The eugo.hpc API](https://docs.eugo.io/docs/guides/eugo-hpc-api.md): The functions eugo.hpc gives you for running work on a Ray cluster: connecting, distributed functions and actors, objects, the token file, and how each name maps to Ray.
- [Memory for distributed jobs](https://docs.eugo.io/docs/guides/memory-for-distributed-jobs.md): How memory works when your code runs on Ray through eugo.hpc: what the object store is, how big it is and how to change it, what spilling to disk means, and what to do when Ray kills a task for running out of memory.
- [Lowering peak memory with one glibc setting on Graviton2](https://docs.eugo.io/docs/guides/lowering-peak-memory.md): An opt-in glibc setting that can lower the peak memory of a memory-bound program on Graviton2 (AWS Lambda, AWS Fargate): what it saves, what it costs, and how to check that it took effect.
- [Compiling code in notebook cells](https://docs.eugo.io/docs/guides/compile-cells.md): Compile and run C, C++, CUDA, Fortran and Rust code, or run JavaScript and TypeScript, straight from a notebook cell with Eugo's cell magics: which sessions have them, how a cell runs, every option, and an example of each.
- [Compiling your own C, C++ and Fortran code in Eugo](https://docs.eugo.io/docs/guides/compiling-your-own-code.md): Building your own native code with the compilers in the Eugo image: the errors you may meet and how to fix them, GMP and MPFR, the sanitizers and the static analyzers.
- [Running PyTorch on the GPU in Eugo](https://docs.eugo.io/docs/guides/pytorch-on-the-gpu.md): Run attention under mixed precision so Eugo's PyTorch uses its fused GPU attention kernels, plus attention dropout, FlashAttention in Hugging Face Transformers, the T4G and other GPU notes.
- [Writing video with FFmpeg: the GPU encoder and small pictures](https://docs.eugo.io/docs/guides/video-with-ffmpeg-on-the-gpu.md): Which H.264 encoder Eugo's FFmpeg picks on a machine with an NVIDIA GPU, the GPU encoder's minimum picture size, and what to do for small videos in ffmpeg, PyAV, mediapy and matplotlib.
