stornic56/openvino-runtime-simple-installer
1
0
Fork 0
mirror of https://github.com/stornic56/openvino-runtime-simple-installer.git synced 2026-10-31 13:08:06 +00:00
Code Issues Packages Projects Releases Wiki Activity

Update: OpenVINO 2026.4.0 + NEO 26.35.39758.10 + fixes de robustez y Mesa

Browse Source 
- OpenVINO Runtime 2026.4.0.22959.99c81491cc3 (BASE_URL packages/2026.4)
- NEO 26.35.39758.10 + IGC 2.41.5+22716 (6 paquetes de produccion, sin dbgsym/-devel)
- Mesa >= 25.2 en Debian 13: mesa-libgallium sustituye mesa-va-drivers/mesa-vdpau-drivers (Rompe < 25.2.8-3)
- INSTALL_DIR derivado de la version (/opt/intel/openvino_2026.4.0) con politica replace y symlink estable
- rm -rf cubre symlink O directorio real en la limpieza previa
- deps/: set -euo pipefail + guard ${1:-} (sin abortos esporios)
- DRY flags/GPU: read -r con guard EOF, deteccion GPU sin SIGPIPE, trap EXIT en modulos, GPU_PRESENT como comparacion
- .gitignore: artefactos de instalacion fuera del repo
This commit is contained in:
stornic56
2026-09-28 22:41:41 -05:00
parent 3b0c2051f1
commit 9fa67d432c
8 changed files with 288 additions and 258 deletions
Download Patch File Download Diff File Expand all files Collapse all files
README.md
+25 -16
View File
@@ -4,10 +4,11 @@ This project repository provides a comprehensive automated installation and conf
## Supported Operating Systems
* Debian 13
* Ubuntu 24.04
* Ubuntu 26.04
* Fedora 43
* Debian 13
* Ubuntu 22.04
* Ubuntu 24.04
* Ubuntu 26.04
* Fedora 43
---
@@ -16,7 +17,9 @@ This project repository provides a comprehensive automated installation and conf
The installer scripts are designed to be executed as a root user (`sudo`). Ensure all necessary system dependencies (such as `git`, `wget`, and the respective package managers like `dpkg`/`dnf`) exist before execution.
### Setup & Clone Repository
First, clone the repository from GitHub:
```bash
# 1. Clone the installer repository
git clone https://github.com/stornic56/openvino-runtime-simple-installer
@@ -27,7 +30,9 @@ chmod +x main.sh
```
### Execution
Execute the primary installation script using elevated privileges (root access is mandatory):
```bash
sudo ./main.sh
```
@@ -37,19 +42,21 @@ sudo ./main.sh
The execution of `main.sh` automatically handles OS detection and conditionally installs components based on hardware presence, ensuring a fully configured inference pipeline:
**Core Runtime & Libraries:**
* **OpenVINO Toolkit 2026.0:** The complete runtime package is downloaded from the official OpenVINO repository (`https://storage.openvinotoolkit.org/repositories/openvino/packages/2026.0/linux`) and installed under `/opt/intel/openvino_2026`. It is the same version used in llama.cpp in its builds.
* **System Dependencies:** Required libraries, including `cmake`, Python 3 (with version-specific dependencies like `libpython3.13` for Debian 13 or `libpython3.13` for Ubuntu 26.04), GCC compilers (`g++`), NumPy bindings, and standard development tools are installed via system package managers (`apt`/`dnf`).
* **OpenVINO Toolkit 2026.4:** The complete runtime package is downloaded from the official OpenVINO repository (`https://storage.openvinotoolkit.org/repositories/openvino/packages/2026.4/linux`) and installed under `/opt/intel/openvino_2026`. It is the same version used in llama.cpp in its builds.
* **System Dependencies:** Required libraries, including `cmake`, Python 3 (with version-specific dependencies like `libpython3.13` for Debian 13 or `libpython3.13` for Ubuntu 26.04), GCC compilers (`g++`), NumPy bindings, and standard development tools are installed via system package managers (`apt`/`dnf`).
**Hardware Acceleration & Drivers (NEO):**
* **Intel Compute Runtime NEO:** If an Intel GPU is detected, the respective module script installs critical proprietary drivers:
* `intel-igc-core`: Graphics Compiler core components.
* `intel-igc-opencl`: OpenCL interface libraries for graphics compilation.
* `intel-ocloc`/`libze-intel-gpu1`: Compute Runtime interfaces and GPU acceleration layers (OpenCL/oneAPI).
* **Multimedia Support:** Installation of `mesa-va-drivers`, `vainfo`, and related packages to ensure Video Acceleration API support, crucial for media inference tasks.
* **Intel Compute Runtime NEO (26.35.39758.10):** If an Intel GPU is detected, the respective module script installs critical proprietary drivers:
* `intel-igc-core` (2.41.5): Graphics Compiler core components.
* `intel-igc-opencl` (2.41.5): OpenCL interface libraries for graphics compilation.
* `intel-ocloc`/`libze-intel-gpu1`: Compute Runtime interfaces and GPU acceleration layers (OpenCL/oneAPI).
* **Multimedia Support:** Installation of `mesa-va-drivers`, `vainfo`, and related packages to ensure Video Acceleration API support, crucial for media inference tasks.
**Environment Configuration:**
* User permissions are configured by adding the executing user (`$SUDO_USER`) to the system groups `render` and `video` to allow proper GPU access post-installation.
* User permissions are configured by adding the executing user (`$SUDO_USER`) to the system groups `render` and `video` to allow proper GPU access post-installation.
## After Installation
@@ -66,8 +73,10 @@ clinfo | grep -i "device name"
4. Check video acceleration
vainfo
```
## Note:
## Note
Users should reference the following official projects:
- Llamacpp: [Official Repository Link](https://github.com/ggml-org/llama.cpp) and [OpenvinoBackend](https://github.com/ggml-org/llama.cpp/blob/master/docs/backend/OPENVINO.md)
- OpenVino Toolkit: [Docs](https://docs.openvino.ai/2026/get-started/install-openvino/install-openvino-archive-linux.html) and [github](https://github.com/openvinotoolkit/openvino)
- Intel NEO Drivers: [Intel Compute Runtime](https://github.com/intel/compute-runtime)
* Llamacpp: [Official Repository Link](https://github.com/ggml-org/llama.cpp) and [OpenvinoBackend](https://github.com/ggml-org/llama.cpp/blob/master/docs/backend/OPENVINO.md)
* OpenVino Toolkit: [Docs](https://docs.openvino.ai/2026/get-started/install-openvino/install-openvino-archive-linux.html) and [github](https://github.com/openvinotoolkit/openvino)
* Intel NEO Drivers: [Intel Compute Runtime](https://github.com/intel/compute-runtime)