fix(renderer): make --gl egl/angle actually use the GPU; add GPU userspace to the Docker image by thecodacus · Pull Request #13 · thecodacus/rendiv

thecodacus · 2026-09-28T22:25:20Z

Problem

rendiv render --gl egl / --gl angle still rendered on the CPU. Headless Chromium picks SwiftShader for plain --use-gl=egl and --use-gl=angle, even with a GPU passed into the container. The Studio image also has no Vulkan loader, no GLVND EGL and no NVIDIA vendor files, so Chromium can't reach the driver the NVIDIA runtime injects.

Measured inside the Studio container (RTX 3060, --gpus all) by printing the WebGL renderer string:

flags	renderer
`--use-gl=egl` (old `--gl egl`)	ANGLE (Google, Vulkan 1.3.0 (SwiftShader …))
`--use-gl=angle` (old `--gl angle`)	ANGLE (Google, Vulkan 1.3.0 (SwiftShader …))
`--use-gl=angle --use-angle=gl-egl`	ANGLE (NVIDIA Corporation, NVIDIA GeForce RTX 3060/PCIe/SSE2, OpenGL ES 3.2)
`--use-gl=angle --use-angle=vulkan`	ANGLE (NVIDIA, Vulkan 1.4.329 (NVIDIA GeForce RTX 3060))

(the last two only after adding the loaders + vendor files below)

Changes

renderer: egl → --use-gl=angle --use-angle=gl-egl, angle → --use-gl=angle --use-angle=vulkan. swiftshader unchanged. Same option names, no API change.
Dockerfile (runtime stage): libvulkan1 libegl1 libgl1 libglvnd0, the NVIDIA Vulkan ICD + GLVND EGL vendor JSONs, ENV NVIDIA_DRIVER_CAPABILITIES=all, and a usage note for --gpus all. No GPU → still renders on the CPU.
Docs: CLI help, the rendiv-video skill table and docs/rendering-performance.md describe what each mode uses.

Result

Full render of an ~18-min, GL-heavy video (three.js/WebGL plates + motion blur) in the patched Studio container: ~3 h → ~13 min.

Summary by CodeRabbit

New Features
- Linux rendering can use GPU acceleration through ANGLE on Vulkan or EGL/GLES. SwiftShader remains available for CPU rendering.
- Docker rendering supports NVIDIA GPUs when launched with GPU access enabled; the image includes the required graphics runtime components.
Documentation
- Updated rendering guidance with GPU setup instructions and a performance comparison between GPU rendering and SwiftShader.

…space to the Docker image Plain --use-gl=egl / --use-gl=angle let headless Chromium fall back to SwiftShader even with a GPU and drivers present, so "GPU" renders silently ran on the CPU. egl now passes --use-angle=gl-egl and angle --use-angle=vulkan (verified on an NVIDIA RTX 3060 in Docker via the WebGL renderer string). The Studio image gains the Vulkan loader and GLVND EGL/GL (libvulkan1, libegl1, libgl1, libglvnd0), the NVIDIA Vulkan ICD and EGL vendor JSONs, and NVIDIA_DRIVER_CAPABILITIES=all, so a container run with --gpus all can render on the GPU. Without a GPU it still renders on the CPU. Measured: a GL-heavy 18-min video went from ~3 h to ~13 min. Co-Authored-By: Claude Opus 5.5 <[email protected]>

coderabbitai · 2026-09-28T22:25:36Z

Navigate logical layers of code changes, visualize relationships, and explore their blast radius.

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Advanced

Run ID: 1403f482-4d70-42a8-82e1-726e34671837

📥 Commits

Reviewing files that changed from the base of the PR and between 6d6aca2 and 4abf31c.

📒 Files selected for processing (5)

Dockerfile
docs/rendering-performance.md
packages/cli/src/commands/render.ts
packages/renderer/src/browser.ts
packages/skills/rendiv-video/rules/cli-and-studio.md

Included review availability: This review used your included allowance. Your plan provides up to 2 included reviews per hour; 1 remain after this review.

📝 Walkthrough

Walkthrough

The render options now select explicit ANGLE backends for EGL/GLES and Vulkan. The Docker image adds Vulkan and GLVND libraries and NVIDIA vendor configuration. The rendering guide and CLI descriptions document GPU options and Docker GPU access.

Changes

GPU Rendering

Layer / File(s)	Summary
GL backend selection `packages/cli/src/commands/render.ts`, `packages/renderer/src/browser.ts`, `packages/skills/rendiv-video/rules/cli-and-studio.md`	The `--gl` descriptions identify the EGL/GLES and Vulkan backends. The renderer maps `egl` to ANGLE’s `gl-egl` backend and `angle` to Vulkan.
Docker GPU support `Dockerfile`, `docs/rendering-performance.md`	The Docker image adds Vulkan and GLVND libraries, NVIDIA vendor configuration files, and `NVIDIA_DRIVER_CAPABILITIES=all`. Docker comments and the rendering guide document GPU access and rendering options.

Priority: ➖ Normal

Estimated code review effort: 2 (Simple) | ~10 minutes

Change: Bug fix

Merge Risk: ⚪ Minimal · up to 4abf3

This change makes the egl and angle GPU render modes select explicit ANGLE backends and adds NVIDIA GPU support to the Docker image. No actionable merge-blocking risk remains. Default CPU rendering is unchanged.

Architecture Summary

Architecture risk: 🔵 Low · up to 4abf3

The change affects 5 systems.

Changed systems: packages/cli, packages/renderer, Dockerfile, docs, packages/skills

Architecture concerns
No architecture-level concerns identified.

Review details

Systems and components

observed — packages/cli (library) was modified; 1 changed file maps to changed impact.
observed — packages/renderer (library) was modified; 1 changed file maps to changed impact.
observed — Dockerfile (service) was modified; 1 changed file maps to changed impact.
observed — docs (service) was modified; 1 changed file maps to changed impact.

Before / after behavior

observed — Modified behavior in Dockerfile: The usage comments add NVIDIA GPU setup instructions, identify angle/Vulkan and egl rendering options, and state that rendering falls back to CPU SwiftShader without a GPU.
observed — Modified behavior in Dockerfile: The runtime image adds Vulkan and GLVND libraries, creates NVIDIA ICD and EGL vendor JSON files for the injected driver libraries, and sets NVIDIA_DRIVER_CAPABILITIES=all. Previously, this install block only removed apt package lists after installing the existing dependencies.
observed — Modified behavior in docs/rendering-performance.md: The software GL fix now recommends --gl angle on Vulkan or --gl egl on Linux GPU, and documents Docker’s --gpus all setup and a measured RTX 3060 rendering comparison. It replaces the prior --use-gl=egl or native-GPU guidance; the macOS headless-GPU caveat remains.
observed — Modified behavior in packages/cli/src/commands/render.ts: The --gl option description now specifies the EGL/GLES and Vulkan backends for egl and angle; the available values and swiftshader default are unchanged.

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly describes the main renderer fix and the Docker image GPU userspace changes.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 1 functions across 2 files. (3 skipped: 3 …
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

📝 Generate docstrings

Commit to this branch
Create a new PR

🧪 Generate unit tests (beta)

Commit to this branch
Create a new PR

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

_{Comment @coderabbitai help to get the list of available commands.}

Navigation Menu

Repository navigation

fix(renderer): make --gl egl/angle actually use the GPU; add GPU userspace to the Docker image - #13

fix(renderer): make --gl egl/angle actually use the GPU; add GPU userspace to the Docker image#13
thecodacus wants to merge 1 commit into
masterfrom
fix/gpu-rendering

thecodacus commented Sep 28, 2026 •

edited by coderabbitai Bot

Loading

Uh oh!

coderabbitai Bot commented Sep 28, 2026 •

edited

Loading

Walkthrough

Changes

Uh oh!

This branch has not been deployed

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Conversation

thecodacus commented Sep 28, 2026 • edited by coderabbitai Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Problem

Changes

Result

Summary by CodeRabbit

Uh oh!

coderabbitai Bot commented Sep 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Walkthrough

Changes

Architecture Summary

Uh oh!

This branch has not been deployed

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

thecodacus commented Sep 28, 2026 •

edited by coderabbitai Bot

Loading

coderabbitai Bot commented Sep 28, 2026 •

edited

Loading