Test code error: no available pyatomic implementation for this platform/compiler #92
Description
I tried to follow the instructions for installing this package, but I run into an error when running the test code.
I have CUDA 12.4, vcredist 2015-2022, and it seems that the bundled TinyCC works when I call it directly with tcc.exe -v.
I see that pyatomic displays an error about not having an implementation for the platform or compiler, but I'm not sure if it is related to the CalledProcessError. Any ideas?
I get the following trace when running the test code:
E:\>py test.py
In file included from C:/Users/ssit5/AppData/Local/Temp/tmp7h3_h1dj/cuda_utils.c:12:
In file included from C:/Program Files/Python313/Include/Python.h:70:
In file included from C:/Program Files/Python313/Include/pyatomic.h:9:
C:/Program Files/Python313/Include/cpython/pyatomic.h:543: error: #error "no available pyatomic implementation for this platform/compiler"
Traceback (most recent call last):
File "E:\test.py", line 25, in <module>
b_compiled = add(a, a)
File "E:\test.py", line 20, in add
add_kernel[grid](x, y, output, n_elements, BLOCK_SIZE=1024)
~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Program Files\Python313\Lib\site-packages\triton\runtime\jit.py", line 330, in <lambda>
return lambda *args, **kwargs: self.run(grid=grid, warmup=False, *args, **kwargs)
~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Program Files\Python313\Lib\site-packages\triton\runtime\jit.py", line 568, in run
device = driver.active.get_current_device()
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Program Files\Python313\Lib\site-packages\triton\runtime\driver.py", line 23, in __getattr__
self._initialize_obj()
~~~~~~~~~~~~~~~~~~~~^^
File "C:\Program Files\Python313\Lib\site-packages\triton\runtime\driver.py", line 20, in _initialize_obj
self._obj = self._init_fn()
~~~~~~~~~~~~~^^
File "C:\Program Files\Python313\Lib\site-packages\triton\runtime\driver.py", line 9, in _create_driver
return actives[0]()
~~~~~~~~~~^^
File "C:\Program Files\Python313\Lib\site-packages\triton\backends\nvidia\driver.py", line 493, in __init__
self.utils = CudaUtils() # TODO: make static
~~~~~~~~~^^
File "C:\Program Files\Python313\Lib\site-packages\triton\backends\nvidia\driver.py", line 92, in __init__
mod = compile_module_from_src(Path(os.path.join(dirname, "driver.c")).read_text(), "cuda_utils")
File "C:\Program Files\Python313\Lib\site-packages\triton\backends\nvidia\driver.py", line 69, in compile_module_from_src
so = _build(name, src_path, tmpdir, library_dirs(), include_dir, libraries)
File "C:\Program Files\Python313\Lib\site-packages\triton\runtime\build.py", line 96, in _build
ret = subprocess.check_call(cc_cmd)
File "C:\Program Files\Python313\Lib\subprocess.py", line 421, in check_call
raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command '['C:\\Program Files\\Python313\\Lib\\site-packages\\triton\\runtime\\tcc\\tcc.exe', 'C:\\Users\\ssit5\\AppData\\Local\\Temp\\tmp7h3_h1dj\\cuda_utils.c', '-O3', '-shared', '-fPIC', '-Wno-psabi', '-o', 'C:\\Users\\ssit5\\AppData\\Local\\Temp\\tmp7h3_h1dj\\cuda_utils.cp313-win_amd64.pyd', '-lcuda', '-lpython3', '-LC:\\Program Files\\Python313\\Lib\\site-packages\\triton\\backends\\nvidia\\lib', '-LC:\\Program Files\\NVIDIA GPU Computing Toolkit\\CUDA\\v12.4\\lib\\x64', '-LC:\\Program Files\\Python313\\libs', '-IC:\\Program Files\\Python313\\Lib\\site-packages\\triton\\backends\\nvidia\\include', '-IC:\\Program Files\\NVIDIA GPU Computing Toolkit\\CUDA\\v12.4\\include', '-IC:\\Users\\ssit5\\AppData\\Local\\Temp\\tmp7h3_h1dj', '-IC:\\Program Files\\Python313\\Include']' returned non-zero exit status 1.