aurarouter-cuda13 0.5.8189__tar.gz
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- aurarouter_cuda13-0.5.8189/LICENSE +21 -0
- aurarouter_cuda13-0.5.8189/PKG-INFO +34 -0
- aurarouter_cuda13-0.5.8189/README.md +24 -0
- aurarouter_cuda13-0.5.8189/pyproject.toml +17 -0
- aurarouter_cuda13-0.5.8189/setup.cfg +4 -0
- aurarouter_cuda13-0.5.8189/src/aurarouter_cuda13/__init__.py +39 -0
- aurarouter_cuda13-0.5.8189/src/aurarouter_cuda13/diagnostics.py +50 -0
- aurarouter_cuda13-0.5.8189/src/aurarouter_cuda13/logging.py +8 -0
- aurarouter_cuda13-0.5.8189/src/aurarouter_cuda13/metadata.py +12 -0
- aurarouter_cuda13-0.5.8189/src/aurarouter_cuda13.egg-info/PKG-INFO +34 -0
- aurarouter_cuda13-0.5.8189/src/aurarouter_cuda13.egg-info/SOURCES.txt +11 -0
- aurarouter_cuda13-0.5.8189/src/aurarouter_cuda13.egg-info/dependency_links.txt +1 -0
- aurarouter_cuda13-0.5.8189/src/aurarouter_cuda13.egg-info/top_level.txt +1 -0
|
@@ -0,0 +1,21 @@
|
|
|
1
|
+
MIT License
|
|
2
|
+
|
|
3
|
+
Copyright (c) 2026 AuraCoreDynamics
|
|
4
|
+
|
|
5
|
+
Permission is hereby granted, free of charge, to any person obtaining a copy
|
|
6
|
+
of this software and associated documentation files (the "Software"), to deal
|
|
7
|
+
in the Software without restriction, including without limitation the rights
|
|
8
|
+
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
|
9
|
+
copies of the Software, and to permit persons to whom the Software is
|
|
10
|
+
furnished to do so, subject to the following conditions:
|
|
11
|
+
|
|
12
|
+
The above copyright notice and this permission notice shall be included in all
|
|
13
|
+
copies or substantial portions of the Software.
|
|
14
|
+
|
|
15
|
+
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
|
16
|
+
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
|
17
|
+
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
|
18
|
+
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
|
19
|
+
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
|
20
|
+
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
|
21
|
+
SOFTWARE.
|
|
@@ -0,0 +1,34 @@
|
|
|
1
|
+
Metadata-Version: 2.4
|
|
2
|
+
Name: aurarouter-cuda13
|
|
3
|
+
Version: 0.5.8189
|
|
4
|
+
Summary: CUDA 13.1 backend for AuraRouter
|
|
5
|
+
Author-email: Steven Siebert <info@auracoredynamics.com>
|
|
6
|
+
Requires-Python: >=3.12
|
|
7
|
+
Description-Content-Type: text/markdown
|
|
8
|
+
License-File: LICENSE
|
|
9
|
+
Dynamic: license-file
|
|
10
|
+
|
|
11
|
+
# aurarouter-cuda13
|
|
12
|
+
|
|
13
|
+
**AuraRouter High-Performance GPU Backend (CUDA 13.1)**
|
|
14
|
+
|
|
15
|
+
This package is a "sidecar" for **AuraRouter**. It contains the pre-compiled binary payload required to run local LLMs on modern NVIDIA GPUs with maximum performance.
|
|
16
|
+
|
|
17
|
+
## What's inside?
|
|
18
|
+
- `llama-server.exe`: Compiled with CUDA 13.1 support.
|
|
19
|
+
- `ggml-cuda.dll`: Optimized kernels for NVIDIA Ada (RTX 40), Blackwell (RTX 50), and Ampere (RTX 30) architectures.
|
|
20
|
+
- `cudart64_13.dll` & `cublas64_13.dll`: Required NVIDIA runtime libraries.
|
|
21
|
+
|
|
22
|
+
## Requirements
|
|
23
|
+
- **Hardware**: NVIDIA GPU (RTX 20-series or newer recommended).
|
|
24
|
+
- **Drivers**: NVIDIA Driver version **570.86** or higher.
|
|
25
|
+
- **AuraRouter**: Must be installed in the same environment.
|
|
26
|
+
|
|
27
|
+
## Usage
|
|
28
|
+
Simply install this package into the same virtual environment as `aurarouter`:
|
|
29
|
+
|
|
30
|
+
```bash
|
|
31
|
+
pip install aurarouter-cuda13
|
|
32
|
+
```
|
|
33
|
+
|
|
34
|
+
AuraRouter's `BinaryManager` will automatically detect this package and prioritize it over CPU-based backends. No configuration change is required.
|
|
@@ -0,0 +1,24 @@
|
|
|
1
|
+
# aurarouter-cuda13
|
|
2
|
+
|
|
3
|
+
**AuraRouter High-Performance GPU Backend (CUDA 13.1)**
|
|
4
|
+
|
|
5
|
+
This package is a "sidecar" for **AuraRouter**. It contains the pre-compiled binary payload required to run local LLMs on modern NVIDIA GPUs with maximum performance.
|
|
6
|
+
|
|
7
|
+
## What's inside?
|
|
8
|
+
- `llama-server.exe`: Compiled with CUDA 13.1 support.
|
|
9
|
+
- `ggml-cuda.dll`: Optimized kernels for NVIDIA Ada (RTX 40), Blackwell (RTX 50), and Ampere (RTX 30) architectures.
|
|
10
|
+
- `cudart64_13.dll` & `cublas64_13.dll`: Required NVIDIA runtime libraries.
|
|
11
|
+
|
|
12
|
+
## Requirements
|
|
13
|
+
- **Hardware**: NVIDIA GPU (RTX 20-series or newer recommended).
|
|
14
|
+
- **Drivers**: NVIDIA Driver version **570.86** or higher.
|
|
15
|
+
- **AuraRouter**: Must be installed in the same environment.
|
|
16
|
+
|
|
17
|
+
## Usage
|
|
18
|
+
Simply install this package into the same virtual environment as `aurarouter`:
|
|
19
|
+
|
|
20
|
+
```bash
|
|
21
|
+
pip install aurarouter-cuda13
|
|
22
|
+
```
|
|
23
|
+
|
|
24
|
+
AuraRouter's `BinaryManager` will automatically detect this package and prioritize it over CPU-based backends. No configuration change is required.
|
|
@@ -0,0 +1,17 @@
|
|
|
1
|
+
[build-system]
|
|
2
|
+
requires = ["setuptools>=61.0"]
|
|
3
|
+
build-backend = "setuptools.build_meta"
|
|
4
|
+
|
|
5
|
+
[project]
|
|
6
|
+
name = "aurarouter-cuda13"
|
|
7
|
+
version = "0.5.8189"
|
|
8
|
+
authors = [{ name = "Steven Siebert", email = "info@auracoredynamics.com" }]
|
|
9
|
+
description = "CUDA 13.1 backend for AuraRouter"
|
|
10
|
+
readme = "README.md"
|
|
11
|
+
requires-python = ">=3.12"
|
|
12
|
+
|
|
13
|
+
[tool.setuptools.packages.find]
|
|
14
|
+
where = ["src"]
|
|
15
|
+
|
|
16
|
+
[tool.setuptools.package-data]
|
|
17
|
+
aurarouter_cuda13 = ["bin/**/*"]
|
|
@@ -0,0 +1,39 @@
|
|
|
1
|
+
\"\"\"AuraRouter CUDA 13.1 Backend Package\"\"\"
|
|
2
|
+
|
|
3
|
+
import os
|
|
4
|
+
import sys
|
|
5
|
+
import platform
|
|
6
|
+
from pathlib import Path
|
|
7
|
+
from .logging import get_logger
|
|
8
|
+
|
|
9
|
+
logger = get_logger()
|
|
10
|
+
|
|
11
|
+
def get_bin_dir() -> Path:
|
|
12
|
+
\"\"\"Return the absolute path to the bundled binaries directory for the current platform.\"\"\"
|
|
13
|
+
machine = platform.machine().lower()
|
|
14
|
+
|
|
15
|
+
if sys.platform == \"win32\" and machine in (\"amd64\", \"x86_64\"):
|
|
16
|
+
plat = \"win-x64\"
|
|
17
|
+
elif sys.platform == \"linux\" and machine == \"x86_64\":
|
|
18
|
+
plat = \"linux-x64\"
|
|
19
|
+
elif sys.platform == \"darwin\" and machine in (\"x86_64\", \"arm64\"):
|
|
20
|
+
plat = \"macos-x64\"
|
|
21
|
+
else:
|
|
22
|
+
raise RuntimeError(f\"Unsupported platform for this backend package: {sys.platform}/{machine}\")
|
|
23
|
+
|
|
24
|
+
return Path(__file__).parent / \"bin\" / plat
|
|
25
|
+
|
|
26
|
+
def setup_runtime_environment():
|
|
27
|
+
\"\"\"Configure the environment (like PATH/DLL search paths) to ensure binaries can load their dependencies.\"\"\"
|
|
28
|
+
try:
|
|
29
|
+
bin_dir = get_bin_dir()
|
|
30
|
+
if sys.platform == \"win32\" and hasattr(os, \"add_dll_directory\"):
|
|
31
|
+
# Ensure sibling DLLs (like cudart, cublas) are discoverable on Windows
|
|
32
|
+
logger.debug(f\"Adding DLL directory: {bin_dir}\")
|
|
33
|
+
os.add_dll_directory(str(bin_dir))
|
|
34
|
+
|
|
35
|
+
# Add to PATH as a fallback for subprocesses
|
|
36
|
+
logger.debug(f\"Prepending to PATH: {bin_dir}\")
|
|
37
|
+
os.environ[\"PATH\"] = str(bin_dir) + os.pathsep + os.environ.get(\"PATH\", \"\")
|
|
38
|
+
except Exception as e:
|
|
39
|
+
logger.error(f\"Failed to setup runtime environment: {e}\")
|
|
@@ -0,0 +1,50 @@
|
|
|
1
|
+
import ctypes
|
|
2
|
+
import subprocess
|
|
3
|
+
import os
|
|
4
|
+
from . import get_bin_dir
|
|
5
|
+
from .logging import get_logger
|
|
6
|
+
|
|
7
|
+
logger = get_logger(\"diagnostics\")
|
|
8
|
+
|
|
9
|
+
def run_diagnostic():
|
|
10
|
+
\"\"\"Verify that the GPU and CUDA DLLs are working correctly.\"\"\"
|
|
11
|
+
logger.info(\"Starting GPU/CUDA diagnostics...\")
|
|
12
|
+
results = {
|
|
13
|
+
\"gpu_present\": False,
|
|
14
|
+
\"driver_version\": \"Unknown\",
|
|
15
|
+
\"dll_load_status\": \"Not Tested\",
|
|
16
|
+
\"vram_total_mb\": 0,
|
|
17
|
+
\"vram_free_mb\": 0
|
|
18
|
+
}
|
|
19
|
+
|
|
20
|
+
# 1. Check for nvidia-smi
|
|
21
|
+
try:
|
|
22
|
+
smi = subprocess.check_output([\"nvidia-smi\", \"--query-gpu=driver_version,memory.total,memory.free\", \"--format=csv,noheader,nounits\"], encoding='utf-8')
|
|
23
|
+
parts = smi.strip().split(',')
|
|
24
|
+
results[\"gpu_present\"] = True
|
|
25
|
+
results[\"driver_version\"] = parts[0].strip()
|
|
26
|
+
results[\"vram_total_mb\"] = int(parts[1].strip())
|
|
27
|
+
results[\"vram_free_mb\"] = int(parts[2].strip())
|
|
28
|
+
logger.debug(f\"NVIDIA GPU found. Driver: {results['driver_version']}\")
|
|
29
|
+
except Exception as e:
|
|
30
|
+
logger.warning(f\"nvidia-smi failed or not found: {e}\")
|
|
31
|
+
results[\"gpu_present\"] = False
|
|
32
|
+
|
|
33
|
+
# 2. Try loading the CUDA Runtime DLL
|
|
34
|
+
try:
|
|
35
|
+
bin_dir = get_bin_dir()
|
|
36
|
+
dll_name = f\"cudart64_13.dll\"
|
|
37
|
+
dll_path = str(bin_dir / dll_name)
|
|
38
|
+
if os.path.exists(dll_path):
|
|
39
|
+
logger.debug(f\"Attempting to load {dll_name} from {bin_dir}\")
|
|
40
|
+
ctypes.WinDLL(dll_path)
|
|
41
|
+
results[\"dll_load_status\"] = \"Success\"
|
|
42
|
+
logger.info(\"CUDA DLL load successful.\")
|
|
43
|
+
else:
|
|
44
|
+
results[\"dll_load_status\"] = f\"Missing: {dll_name}\"
|
|
45
|
+
logger.error(f\"CUDA DLL not found in backend bin directory: {dll_path}\")
|
|
46
|
+
except Exception as e:
|
|
47
|
+
results[\"dll_load_status\"] = f\"Error: {str(e)}\"
|
|
48
|
+
logger.error(f\"Failed to load CUDA DLL: {e}\")
|
|
49
|
+
|
|
50
|
+
return results
|
|
@@ -0,0 +1,8 @@
|
|
|
1
|
+
import logging
|
|
2
|
+
|
|
3
|
+
def get_logger(name: str = None) -> logging.Logger:
|
|
4
|
+
\"\"\"Return a logger prefixed with the AuraRouter backend namespace.\"\"\"
|
|
5
|
+
base_name = \"AuraRouter.Backend.CUDA13\"
|
|
6
|
+
if name:
|
|
7
|
+
return logging.getLogger(f\"{base_name}.{name}\")
|
|
8
|
+
return logging.getLogger(base_name)
|
|
@@ -0,0 +1,12 @@
|
|
|
1
|
+
\"\"\"Backend metadata for AuraRouter\"\"\"
|
|
2
|
+
|
|
3
|
+
METADATA = {
|
|
4
|
+
\"package_name\": \"aurarouter-cuda13\",
|
|
5
|
+
\"flavor\": \"CUDA 13.1\",
|
|
6
|
+
\"compute_type\": \"GPU (NVIDIA)\",
|
|
7
|
+
\"llama_cpp_build\": \"b8189\",
|
|
8
|
+
\"is_gpu\": True
|
|
9
|
+
}
|
|
10
|
+
|
|
11
|
+
def get_info():
|
|
12
|
+
return METADATA
|
|
@@ -0,0 +1,34 @@
|
|
|
1
|
+
Metadata-Version: 2.4
|
|
2
|
+
Name: aurarouter-cuda13
|
|
3
|
+
Version: 0.5.8189
|
|
4
|
+
Summary: CUDA 13.1 backend for AuraRouter
|
|
5
|
+
Author-email: Steven Siebert <info@auracoredynamics.com>
|
|
6
|
+
Requires-Python: >=3.12
|
|
7
|
+
Description-Content-Type: text/markdown
|
|
8
|
+
License-File: LICENSE
|
|
9
|
+
Dynamic: license-file
|
|
10
|
+
|
|
11
|
+
# aurarouter-cuda13
|
|
12
|
+
|
|
13
|
+
**AuraRouter High-Performance GPU Backend (CUDA 13.1)**
|
|
14
|
+
|
|
15
|
+
This package is a "sidecar" for **AuraRouter**. It contains the pre-compiled binary payload required to run local LLMs on modern NVIDIA GPUs with maximum performance.
|
|
16
|
+
|
|
17
|
+
## What's inside?
|
|
18
|
+
- `llama-server.exe`: Compiled with CUDA 13.1 support.
|
|
19
|
+
- `ggml-cuda.dll`: Optimized kernels for NVIDIA Ada (RTX 40), Blackwell (RTX 50), and Ampere (RTX 30) architectures.
|
|
20
|
+
- `cudart64_13.dll` & `cublas64_13.dll`: Required NVIDIA runtime libraries.
|
|
21
|
+
|
|
22
|
+
## Requirements
|
|
23
|
+
- **Hardware**: NVIDIA GPU (RTX 20-series or newer recommended).
|
|
24
|
+
- **Drivers**: NVIDIA Driver version **570.86** or higher.
|
|
25
|
+
- **AuraRouter**: Must be installed in the same environment.
|
|
26
|
+
|
|
27
|
+
## Usage
|
|
28
|
+
Simply install this package into the same virtual environment as `aurarouter`:
|
|
29
|
+
|
|
30
|
+
```bash
|
|
31
|
+
pip install aurarouter-cuda13
|
|
32
|
+
```
|
|
33
|
+
|
|
34
|
+
AuraRouter's `BinaryManager` will automatically detect this package and prioritize it over CPU-based backends. No configuration change is required.
|
|
@@ -0,0 +1,11 @@
|
|
|
1
|
+
LICENSE
|
|
2
|
+
README.md
|
|
3
|
+
pyproject.toml
|
|
4
|
+
src/aurarouter_cuda13/__init__.py
|
|
5
|
+
src/aurarouter_cuda13/diagnostics.py
|
|
6
|
+
src/aurarouter_cuda13/logging.py
|
|
7
|
+
src/aurarouter_cuda13/metadata.py
|
|
8
|
+
src/aurarouter_cuda13.egg-info/PKG-INFO
|
|
9
|
+
src/aurarouter_cuda13.egg-info/SOURCES.txt
|
|
10
|
+
src/aurarouter_cuda13.egg-info/dependency_links.txt
|
|
11
|
+
src/aurarouter_cuda13.egg-info/top_level.txt
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
aurarouter_cuda13
|