bitsandbytes-rocm/bitsandbytes/cuda_setup/main.py

163 lines
6.1 KiB
Python
Raw Normal View History

"""
extract factors the build is dependent on:
2022-08-02 14:42:27 +00:00
[X] compute capability
[ ] TODO: Q - What if we have multiple GPUs of different makes?
- CUDA version
- Software:
- CPU-only: only CPU quantization functions (no optimizer, no matrix multipl)
- CuBLAS-LT: full-build 8-bit optimizer
- no CuBLAS-LT: no 8-bit matrix multiplication (`nomatmul`)
evaluation:
- if paths faulty, return meaningful error
- else:
- determine CUDA version
- determine capabilities
- based on that set the default path
"""
import ctypes
from .paths import determine_cuda_runtime_lib_path
from bitsandbytes.cextension import CUDASetup
def check_cuda_result(cuda, result_val):
2022-08-02 02:43:09 +00:00
# 3. Check for CUDA errors
if result_val != 0:
2022-08-02 02:43:09 +00:00
error_str = ctypes.c_char_p()
cuda.cuGetErrorString(result_val, ctypes.byref(error_str))
CUDASetup.get_instance.add_log_entry(f"CUDA exception! Error code: {error_str.value.decode()}")
def get_cuda_version(cuda, cudart_path):
# https://docs.nvidia.com/cuda/cuda-runtime-api/group__CUDART____VERSION.html#group__CUDART____VERSION
try:
cudart = ctypes.CDLL(cudart_path)
except OSError:
CUDASetup.get_instance.add_log_entry(f'ERROR: libcudart.so could not be read from path: {cudart_path}!')
return None
version = ctypes.c_int()
check_cuda_result(cuda, cudart.cudaRuntimeGetVersion(ctypes.byref(version)))
version = int(version.value)
major = version//1000
minor = (version-(major*1000))//10
if major < 11:
CUDASetup.get_instance().add_log_entry('CUDA SETUP: CUDA version lower than 11 are currenlty not supported for LLM.int8(). You will be only to use 8-bit optimizers and quantization routines!!')
return f'{major}{minor}'
def get_cuda_lib_handle():
# 1. find libcuda.so library (GPU driver) (/usr/lib)
try:
cuda = ctypes.CDLL("libcuda.so")
except OSError:
CUDA_RUNTIME_LIB.get_instance().add_log_entry('CUDA SETUP: WARNING! libcuda.so not found! Do you have a CUDA driver installed? If you are on a cluster, make sure you are on a CUDA machine!')
return None
check_cuda_result(cuda, cuda.cuInit(0))
return cuda
def get_compute_capabilities(cuda):
"""
1. find libcuda.so library (GPU driver) (/usr/lib)
init_device -> init variables -> call function by reference
2. call extern C function to determine CC
(https://docs.nvidia.com/cuda/cuda-driver-api/group__CUDA__DEVICE__DEPRECATED.html)
3. Check for CUDA errors
https://stackoverflow.com/questions/14038589/what-is-the-canonical-way-to-check-for-errors-using-the-cuda-runtime-api
# bits taken from https://gist.github.com/f0k/63a664160d016a491b2cbea15913d549
"""
2022-08-02 02:43:09 +00:00
nGpus = ctypes.c_int()
cc_major = ctypes.c_int()
cc_minor = ctypes.c_int()
device = ctypes.c_int()
2022-08-02 02:43:09 +00:00
check_cuda_result(cuda, cuda.cuDeviceGetCount(ctypes.byref(nGpus)))
ccs = []
for i in range(nGpus.value):
2022-08-02 02:43:09 +00:00
check_cuda_result(cuda, cuda.cuDeviceGet(ctypes.byref(device), i))
ref_major = ctypes.byref(cc_major)
ref_minor = ctypes.byref(cc_minor)
# 2. call extern C function to determine CC
check_cuda_result(
cuda, cuda.cuDeviceComputeCapability(ref_major, ref_minor, device)
)
ccs.append(f"{cc_major.value}.{cc_minor.value}")
2022-08-04 16:16:00 +00:00
return ccs
# def get_compute_capability()-> Union[List[str, ...], None]: # FIXME: error
def get_compute_capability(cuda):
"""
Extracts the highest compute capbility from all available GPUs, as compute
capabilities are downwards compatible. If no GPUs are detected, it returns
None.
"""
ccs = get_compute_capabilities(cuda)
if ccs:
# TODO: handle different compute capabilities; for now, take the max
return ccs[-1]
return None
def evaluate_cuda_setup():
# we remove this for now and see how things go
#print('')
#print('='*35 + 'BUG REPORT' + '='*35)
#print('Welcome to bitsandbytes. For bug reports, please submit your error trace to: https://github.com/TimDettmers/bitsandbytes/issues')
#print('For effortless bug reporting copy-paste your error into this form: https://docs.google.com/forms/d/e/1FAIpQLScPB8emS3Thkp66nvqwmjTEgxp8Y9ufuWTzFyr9kJ5AoI47dQ/viewform?usp=sf_link')
#print('='*80)
2022-08-23 20:59:34 +00:00
#if not torch.cuda.is_available():
#print('No GPU detected. Loading CPU library...')
#return binary_name
binary_name = "libbitsandbytes_cpu.so"
cuda_setup = CUDASetup.get_instance()
cudart_path = determine_cuda_runtime_lib_path()
if cudart_path is None:
cuda_setup.add_log_entry("WARNING: No libcudart.so found! Install CUDA or the cudatoolkit package (anaconda)!", is_warning=True)
return binary_name
cuda_setup.add_log_entry((f"CUDA SETUP: CUDA runtime path found: {cudart_path}"))
cuda = get_cuda_lib_handle()
cc = get_compute_capability(cuda)
cuda_setup.add_log_entry(f"CUDA SETUP: Highest compute capability among GPUs detected: {cc}")
cuda_version_string = get_cuda_version(cuda, cudart_path)
2022-08-04 16:16:00 +00:00
if cc == '':
cuda_setup.add_log_entry("WARNING: No GPU detected! Check your CUDA paths. Processing to load CPU-only library...", is_warning=True)
return binary_name
# 7.5 is the minimum CC vor cublaslt
has_cublaslt = cc in ["7.5", "8.0", "8.6"]
# TODO:
2022-08-02 02:43:09 +00:00
# (1) CUDA missing cases (no CUDA installed by CUDA driver (nvidia-smi accessible)
# (2) Multiple CUDA versions installed
2022-08-02 02:43:09 +00:00
# we use ls -l instead of nvcc to determine the cuda version
# since most installations will have the libcudart.so installed, but not the compiler
cuda_setup.add_log_entry(f'CUDA SETUP: Detected CUDA version {cuda_version_string}')
def get_binary_name():
"if not has_cublaslt (CC < 7.5), then we have to choose _nocublaslt.so"
bin_base_name = "libbitsandbytes_cuda"
if has_cublaslt:
return f"{bin_base_name}{cuda_version_string}.so"
else:
2022-08-04 16:16:00 +00:00
return f"{bin_base_name}{cuda_version_string}_nocublaslt.so"
binary_name = get_binary_name()
return binary_name, cudart_path, cuda, cc, cuda_version_string