def cula_gpu_solve(A_, b_, trans='T'):
A_shape = A_.shape
b_shape = b_.shape
assert (len(A_shape) == 2)
assert (len(b_shape) == 2)
if trans in ['T', 'C']:
l, n = A_shape
k, m = b_shape
if n != k:
raise ValueError('A and b must be aligned.')
elif trans in ['N']:
n, l = A_shape
k, m = b_shape
if l != m:
raise ValueError('A and b must be aligned.')
else:
raise ValueError('Invalid value for trans')
lda = max(1, n)
ldb = max(1, n, l)
# construct pointer arrays needed for culaDeviceSgels
# Cula requires you to pass a pointer for A and b.
A_ptr = A_.gpudata
b_ptr = b_.gpudata
cula.culaDeviceSgels(trans, n, l, m, A_ptr, lda, b_ptr, ldb)
return A_, b_
def cula_gpu_solve(A_, b_, trans='T'):
A_shape = A_.shape
b_shape = b_.shape
assert(len(A_shape) == 2)
assert(len(b_shape) == 2)
if trans in ['T', 'C']:
l, n = A_shape
k, m = b_shape
if n != k:
raise ValueError('A and b must be aligned.')
elif trans in ['N']:
n, l = A_shape
k, m = b_shape
if l != m:
raise ValueError('A and b must be aligned.')
else:
raise ValueError('Invalid value for trans')
lda = max(1, n)
ldb = max(1, n, l)
# construct pointer arrays needed for culaDeviceSgels
# Cula requires you to pass a pointer for A and b.
A_ptr = A_.gpudata
b_ptr = b_.gpudata
cula.culaDeviceSgels(trans, n, l, m, A_ptr, lda, b_ptr, ldb)
return A_, b_