paraframework/device.py at master · ethcocoder/paraframework · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
import os

class Device:
    _instance = None

    def __new__(cls):
        if cls._instance is None:
            cls._instance = super(Device, cls).__new__(cls)
            cls._instance._initialize()
        return cls._instance

    def _initialize(self):
        import sys
        import os
        # Ensure paradma is in path
        paradma_path = os.path.join(os.getcwd(), "paradma")
        if paradma_path not in sys.path:
            sys.path.append(paradma_path)

        self._current_device = 'cpu'
        try:
            self._cpu_backend = __import__('numpy')
            # Paradma uses numpy as its base execution engine
            pass
        except ImportError:
            self._cpu_backend = __import__('numpy')
            print("Paradma not found. Falling back to Numpy.")

        self._gpu_backend = None

        # Check for CuPy availability
        try:
            import cupy
            self._gpu_backend = cupy
            print("CuPy found. GPU acceleration available.")
        except ImportError:
            print("CuPy not found. Running on CPU only.")

    @property
    def backend(self):
        if self._current_device == 'gpu' and self._gpu_backend:
            return self._gpu_backend
        return self._cpu_backend

    @property
    def current_device(self):
        return self._current_device

    def set_device(self, device_type):
        if device_type not in ['cpu', 'gpu']:
            raise ValueError("Device type must be 'cpu' or 'gpu'.")
        if device_type == 'gpu' and not self._gpu_backend:
            raise RuntimeError("CuPy not available. Cannot set device to GPU.")
        self._current_device = device_type
        print(f"Device set to {self._current_device.upper()}")

    def to_cpu(self, array):
        if self._gpu_backend and isinstance(array, self._gpu_backend.ndarray):
            return self._gpu_backend.asnumpy(array)
        return array

    def to_gpu(self, array):
        if self._gpu_backend:
            if isinstance(array, self._cpu_backend.ndarray):
                return self._gpu_backend.asarray(array)
            return array
        raise RuntimeError("CuPy not available. Cannot move array to GPU.")

# Global device instance
device = Device()