Hello,
I compile a simple kernel: vec_add.cu
__global__ void vecAdd(float *a, float *b, float *c)
{
c[threadIdx.x] = a[threadIdx.x] + b[threadIdx.x];
}
nvcc -cubin vec_add.cu
I load it, and this works fine
cuda_mod = cuda.module_from_file("vec_add.cubin")
This blows up however:
>>> vec_add = cuda_mod.get_function("vecAdd")
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
pycuda._driver.LogicError: cuModuleGetFunction failed: not found
Any idea? Same problem also when compiling the kernel to a ptx and
using module_from_buffer
here are my imports:
import pycuda.driver as cuda
import pycuda.autoinit
DATA
VERSION = (0, 93)
VERSION_STATUS = ''
VERSION_TEXT = '0.93'
And this is on fedora 11 x86_64, cuda version 2.3,
Any ideas?
_______________________________________________
PyCUDA mailing list
PyCUDA@tiker.net
http://tiker.net/mailman/listinfo/pycuda_tiker.net
