[PR #47] [MERGED] Handle Linux integrated GPU memory budgeting (CORE-128) #82

New issue

Closed

opened 2026-10-26 22:24:24 +00:00 by bsakaguchi · 0 comments

bsakaguchi commented

2026-10-26 22:24:24 +00:00

Owner

📋 Pull Request Information

Original PR: https://github.com/Comfy-Org/comfy-aimdo/pull/47
Author: @rattus128
Created: 5/7/2026
Status: ✅ Merged
Merged: 5/12/2026
Merged by: @alexisrolland

Base: master ← Head: dev/unified-headroom

📝 Commits (1)

159abbe Handle Linux integrated GPU memory budgeting

📊 Changes

6 files changed (+91 additions, -0 deletions)

View changed files

📝 src-cuda/dispatch.c (+1 -0)
📝 src/control.c (+78 -0)
📝 src/control.h (+4 -0)
📝 src/gpu_abi.h (+4 -0)
📝 src/gpu_dispatch.h (+3 -0)
📝 src/plat.h (+1 -0)

📄 Description

Detect integrated CUDA devices and reserve RAM headroom based on total shared memory. For Linux integrated GPUs, poll /proc/meminfo MemAvailable instead of cuMemGetInfo so eviction decisions respect system RAM pressure.

Contribution Agreement

I agree that my contributions are licensed under the GPLv3.
I grant Comfy Org the rights to relicense these contributions as outlined in CONTRIBUTING.md.

This adds proper support for GDX spark.

_{🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.}

## 📋 Pull Request Information **Original PR:** https://github.com/Comfy-Org/comfy-aimdo/pull/47 **Author:** [@rattus128](https://github.com/rattus128) **Created:** 5/7/2026 **Status:** ✅ Merged **Merged:** 5/12/2026 **Merged by:** [@alexisrolland](https://github.com/alexisrolland) **Base:** `master` ← **Head:** `dev/unified-headroom` --- ### 📝 Commits (1) - [`159abbe`](https://github.com/Comfy-Org/comfy-aimdo/commit/159abbebff1a6ab19d7d1be51198c23d1080c1f7) Handle Linux integrated GPU memory budgeting ### 📊 Changes **6 files changed** (+91 additions, -0 deletions) <details> <summary>View changed files</summary> 📝 `src-cuda/dispatch.c` (+1 -0) 📝 `src/control.c` (+78 -0) 📝 `src/control.h` (+4 -0) 📝 `src/gpu_abi.h` (+4 -0) 📝 `src/gpu_dispatch.h` (+3 -0) 📝 `src/plat.h` (+1 -0) </details> ### 📄 Description Detect integrated CUDA devices and reserve RAM headroom based on total shared memory. For Linux integrated GPUs, poll /proc/meminfo MemAvailable instead of cuMemGetInfo so eviction decisions respect system RAM pressure. ### Contribution Agreement - [X] I agree that my contributions are licensed under the GPLv3. - [X] I grant **Comfy Org** the rights to relicense these contributions as outlined in [CONTRIBUTING.md](./CONTRIBUTING.md). This adds proper support for GDX spark. --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>

bsakaguchi

2026-10-26 22:24:24 +00:00

closed this issue
added the
pull-request
label

No labels

pull-request

No milestone

No project

No assignees

1 participant

Notifications

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference

Comfy-Org/comfy-aimdo#82

No description provided.

Rows
Columns