pygraphblas
pygraphblas copied to clipboard
exception GrB_OUT_OF_MEMORY
Hello! I ran into a problem: when I want to multiply matrices of a large enough size, OutOfMemory is thrown with the following traceback:
Traceback (most recent call last):
File "/home/arseniy.terekhov/Repo/CFPQ_PyAlgo/example.py", line 21, in <module>
m = matrix_base_algo(g, gr)
File "/home/arseniy.terekhov/Repo/CFPQ_PyAlgo/src/algo/matrix_base.py", line 14, in matrix_base_algo
m[l] += m[r1] @ m[r2]
File "/home/arseniy.terekhov/.local/lib/python3.8/site-packages/pygraphblas-3.2.0-py3.8-linux-x86_64.egg/pygraphblas/matrix.py", line 786, in __matmul__
return self.mxm(other)
File "/home/arseniy.terekhov/.local/lib/python3.8/site-packages/pygraphblas-3.2.0-py3.8-linux-x86_64.egg/pygraphblas/matrix.py", line 755, in mxm
_check(lib.GrB_mxm(
File "/home/arseniy.terekhov/.local/lib/python3.8/site-packages/pygraphblas-3.2.0-py3.8-linux-x86_64.egg/pygraphblas/base.py", line 119, in _check
raise _error_codes[res](ffi.string(lib.GrB_error()))
pygraphblas.base.OutOfMemory: b'GraphBLAS error: GrB_OUT_OF_MEMORY\nfunction: GrB_mxm (C, M, accum, semiring, A, B, desc)\nout of memory\n'
At the same time, neither the swap nor even the RAM does not overflow. Or I don't understand what this exception means. I am using 3.2.0 version of pygraphblas, 3.3.0 version of suite sparse and OS ubuntu 20.04.1 LTS. Thank you in advance!
Hi @simpletonDL, this isn't really an issue with pygraphblas, the error is being thrown by the underlying suite sparse library. The only situation I think this happens is when the system malloc (or whatever malloc is passed to GxB_xinit which for pygraphblas is the system malloc) returns NULL, meaning the system is out of memory. Are you sure you're not running with any operating system constraints? Process limits?