[PyCUDA] error about index type when running sparse cg example

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

[PyCUDA] error about index type when running sparse cg example

brennan
Hello,

I'm trying to run the cg example posted at
https://andreask.cs.illinois.edu/PyCuda/Examples/SparseSolve

where I have copied the file into one called: py_cuda_cg_test.py

I'm testing using a 5 by 5 sparse symmetric .mm file I found here
https://people.sc.fsu.edu/~jburkardt/data/mm/m_05_05_crs.mm

though after some messing around I don't think that the file matters much

when I run the command
$ python py_cuda_cg_test.py m_05_05_crs.mm --is-symmetric

I get the error:

Traceback (most recent call last):
  File "py_cuda_cg_test.py", line 88, in <module>
    main_cg()
  File "py_cuda_cg_test.py", line 35, in main_cg
    spmv = PacketedSpMV(csr_mat, options.is_symmetric, csr_mat.dtype)
  File "/home/bs162/.local/lib/python2.7/site-packages/pycuda/sparse/packeted.py", line 185, in __init__
    local_row_costs)
  File "pycuda/sparse/pkt_build_cython.pyx", line 22, in pycuda.sparse.pkt_build_cython.build_pkt_data_structure (/home/bs162/.pyxbld/temp.linux-x86_64-2.7/pyrex/pycuda/sparse/pkt_build_cython.c:1219)
TypeError: 'numpy.float64' object cannot be interpreted as an index

which seems to indicate that the index information is somehow being cast from an int to a float during the call to packeted.py

I get this error with both python 2.7 and 3.6 and I am using numpy 1.13.3. Any help would be greatly appreciated!

Thank you!
_______________________________________________
PyCUDA mailing list -- [hidden email]
To unsubscribe send an email to [hidden email]
Reply | Threaded
Open this post in threaded view
|

[PyCUDA] Re: error about index type when running sparse cg example

Andreas Kloeckner
[hidden email] writes:

> I'm trying to run the cg example posted at
> https://andreask.cs.illinois.edu/PyCuda/Examples/SparseSolve
>
> where I have copied the file into one called: py_cuda_cg_test.py
>
> I'm testing using a 5 by 5 sparse symmetric .mm file I found here
> https://people.sc.fsu.edu/~jburkardt/data/mm/m_05_05_crs.mm
>
> though after some messing around I don't think that the file matters much
>
> when I run the command
> $ python py_cuda_cg_test.py m_05_05_crs.mm --is-symmetric
>
> I get the error:
>
> Traceback (most recent call last):
>   File "py_cuda_cg_test.py", line 88, in <module>
>     main_cg()
>   File "py_cuda_cg_test.py", line 35, in main_cg
>     spmv = PacketedSpMV(csr_mat, options.is_symmetric, csr_mat.dtype)
>   File "/home/bs162/.local/lib/python2.7/site-packages/pycuda/sparse/packeted.py", line 185, in __init__
>     local_row_costs)
>   File "pycuda/sparse/pkt_build_cython.pyx", line 22, in pycuda.sparse.pkt_build_cython.build_pkt_data_structure (/home/bs162/.pyxbld/temp.linux-x86_64-2.7/pyrex/pycuda/sparse/pkt_build_cython.c:1219)
> TypeError: 'numpy.float64' object cannot be interpreted as an index
>
> which seems to indicate that the index information is somehow being cast from an int to a float during the call to packeted.py
>
> I get this error with both python 2.7 and 3.6 and I am using numpy 1.13.3. Any help would be greatly appreciated!
While I'd be happy to merge patches, I do not currently have time to
support the (undocumented/experimental) sparse matrix/linear solver
functionality in PyCUDA.

Sorry,
Andreas

_______________________________________________
PyCUDA mailing list -- [hidden email]
To unsubscribe send an email to [hidden email]

signature.asc (847 bytes) Download Attachment