mrq/bitsandbytes-rocm
1
0
Fork 1
Code Issues 1 Pull Requests Packages Projects Releases Wiki Activity

tentative refactoring of the compute capabilities code

Browse Source
This commit is contained in:
Titus von Koeller 2022-08-02 21:27:36 -07:00
parent 59a615b386
commit 96bc209baf
1 changed files with 35 additions and 21 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

52
bitsandbytes/cuda_setup/compute_capability.py
View File

@ -2,27 +2,28 @@ import ctypes
from dataclasses import dataclass, field
CUDA_SUCCESS = 0
@dataclass
class CudaLibVals:
# code bits taken from
# https://gist.github.com/f0k/63a664160d016a491b2cbea15913d549
nGpus = ctypes.c_int()
cc_major = ctypes.c_int()
cc_minor = ctypes.c_int()
device = ctypes.c_int()
error_str = ctypes.c_char_p()
nGpus: ctypes.c_int = field(default=ctypes.c_int())
cc_major: ctypes.c_int = field(default=ctypes.c_int())
cc_minor: ctypes.c_int = field(default=ctypes.c_int())
device: ctypes.c_int = field(default=ctypes.c_int())
error_str: ctypes.c_char_p = field(default=ctypes.c_char_p())
cuda: ctypes.CDLL = field(init=False, repr=False)
ccs: List[str, ...] = field(init=False)
def load_cuda_lib(self):
def _initialize_driver_API(self):
self.check_cuda_result(self.cuda.cuInit(0))
def _load_cuda_lib(self):
"""
1. find libcuda.so library (GPU driver) (/usr/lib)
init_device -> init variables -> call function by reference
"""
libnames = ("libcuda.so")
libnames = "libcuda.so"
for libname in libnames:
try:
self.cuda = ctypes.CDLL(libname)
@ -33,32 +34,45 @@ class CudaLibVals:
else:
raise OSError("could not load any of: " + " ".join(libnames))
def check_cuda_result(self, result_val):
def call_cuda_func(self, function_obj, **kwargs):
CUDA_SUCCESS = 0 # constant taken from cuda.h
pass
# if (CUDA_SUCCESS := function_obj(
def _error_handle(cuda_lib_call_return_value):
"""
2. call extern C function to determine CC
(see https://docs.nvidia.com/cuda/cuda-driver-api/group__CUDA__DEVICE__DEPRECATED.html)
"""
cls_fields: Tuple[Field, ...] = fields(self.__class__)
CUDA_SUCCESS = 0 # constant taken from cuda.h
if result_val != 0:
self.cuda.cuGetErrorString(result_val, ctypes.byref(self.error_str))
if cuda_lib_call_return_value != CUDA_SUCCESS:
self.cuda.cuGetErrorString(
cuda_lib_call_return_value,
ctypes.byref(self.error_str),
)
print("Count not initialize CUDA - failure!")
raise Exception("CUDA exception!")
return result_val
return cuda_lib_call_return_value
def __post_init__(self):
self.load_cuda_lib()
self.check_cuda_result(self.cuda.cuInit(0))
self.check_cuda_result(self.cuda, self.cuda.cuDeviceGetCount(ctypes.byref(self.nGpus)))
self._load_cuda_lib()
self._initialize_driver_API()
self.check_cuda_result(
self.cuda, self.cuda.cuDeviceGetCount(ctypes.byref(self.nGpus))
)
tmp_ccs = []
for gpu_index in range(self.nGpus.value):
check_cuda_result(
self.cuda, self.cuda.cuDeviceGet(ctypes.byref(self.device), gpu_index)
self.cuda,
self.cuda.cuDeviceGet(ctypes.byref(self.device), gpu_index),
)
check_cuda_result(
self.cuda,
self.cuda.cuDeviceComputeCapability(
ctypes.byref(self.cc_major), ctypes.byref(self.cc_minor), self.device
ctypes.byref(self.cc_major),
ctypes.byref(self.cc_minor),
self.device,
),
)
tmp_ccs.append(f"{self.cc_major.value}.{self.cc_minor.value}")