[PR #15702] install: check for AMD GPU on hybrid systems before exiting #61966

Open
opened 2026-04-29 16:56:04 -05:00 by GiteaMirror · 0 comments
Owner

📋 Pull Request Information

Original PR: https://github.com/ollama/ollama/pull/15702
Author: @mverrilli
Created: 4/19/2026
Status: 🔄 Open

Base: mainHead: fix/issue-15665-hybrid-gpu-install


📝 Commits (1)

  • 5a88958 install: check for AMD GPU on hybrid systems before exiting

📊 Changes

1 file changed (+3 additions, -1 deletions)

View changed files

📝 scripts/install.sh (+3 -1)

📄 Description

Problem

On hybrid systems with both an NVIDIA GPU (`nvidia-smi` present) and an AMD GPU, `scripts/install.sh` exits immediately after detecting `nvidia-smi` — before ever reaching the AMD GPU check. The ROCm bundle is never downloaded, leaving the AMD GPU unsupported.

Reported on Framework Laptop 16 (NVIDIA dGPU + AMD Radeon 890M iGPU) on Fedora 43.

Fix

Rather than duplicating the ROCm download logic, only exit early from the `nvidia-smi` block when no AMD GPU is present. On hybrid systems it falls through to the existing AMD detection and install path.

Test scenarios

System Before After
Hybrid NVIDIA+AMD Exits, skips ROCm Falls through to existing AMD block, downloads ROCm
NVIDIA only Exits cleanly Exits cleanly (no change)
AMD only Downloads ROCm Downloads ROCm (no change)
NVIDIA via lspci (no nvidia-smi) Falls through to CUDA install Falls through to CUDA install (no change)
No GPU CPU-only warning CPU-only warning (no change)

Fixes #15665


🔄 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/ollama/ollama/pull/15702 **Author:** [@mverrilli](https://github.com/mverrilli) **Created:** 4/19/2026 **Status:** 🔄 Open **Base:** `main` ← **Head:** `fix/issue-15665-hybrid-gpu-install` --- ### 📝 Commits (1) - [`5a88958`](https://github.com/ollama/ollama/commit/5a88958f1727b5ef4a7f8c5c6ec0c3239439335a) install: check for AMD GPU on hybrid systems before exiting ### 📊 Changes **1 file changed** (+3 additions, -1 deletions) <details> <summary>View changed files</summary> 📝 `scripts/install.sh` (+3 -1) </details> ### 📄 Description ## Problem On hybrid systems with both an NVIDIA GPU (\`nvidia-smi\` present) and an AMD GPU, \`scripts/install.sh\` exits immediately after detecting \`nvidia-smi\` — before ever reaching the AMD GPU check. The ROCm bundle is never downloaded, leaving the AMD GPU unsupported. Reported on Framework Laptop 16 (NVIDIA dGPU + AMD Radeon 890M iGPU) on Fedora 43. ## Fix Rather than duplicating the ROCm download logic, only exit early from the \`nvidia-smi\` block when no AMD GPU is present. On hybrid systems it falls through to the existing AMD detection and install path. ## Test scenarios | System | Before | After | |--------|--------|-------| | Hybrid NVIDIA+AMD | Exits, skips ROCm | Falls through to existing AMD block, downloads ROCm | | NVIDIA only | Exits cleanly | Exits cleanly (no change) | | AMD only | Downloads ROCm | Downloads ROCm (no change) | | NVIDIA via lspci (no nvidia-smi) | Falls through to CUDA install | Falls through to CUDA install (no change) | | No GPU | CPU-only warning | CPU-only warning (no change) | Fixes #15665 --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>
GiteaMirror added the pull-request label 2026-04-29 16:56:04 -05:00
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#61966