mabshoff wrote: >> #20 0x0805a213 in PyIter_Next (iter=0xa70a66c) at Objects/abstract.c:2375 >> #21 0x0121c5bd in __pyx_f_py_7matrix2_6Matrix_permanent >> (__pyx_v_self=0x9c37194, unused=0x0) at sage/matrix/matrix2.c:1633 > > The above corresponds to the following lines in matrix2.pyx:279-281: > > tmp = [] > for cols in lst: > tmp.append(self.prod_of_row_sums(cols)) > > The crash itself happens when the tmp.append() fails. That indicates a > problem with python's memory management and not an issue in the Sage > code. I would expect python to throw some kind of allocation error, > but since python uses pymalloc it usually asks for a "large" amount of > memory to parcel out smaller chunks for the requests it receives. That > could explain why despite having consumed roughly 80% of memory > available the allocation fails. The other 20% might also be taken up > to a large extent by the kernel and user space, so that makes a > situation where you encounter a "no more memory" situation very > likely. The matrix code in Sage does throw a RuntimeError when it > fails to allocate memory, so I think it is off the hook for now. > > Because you run at about 80% memory consumption and have a large > uptime I would suggest to do a reboot and see if the computation still > fails; hopefully your heap fragmentation problem is gone due to the > reboot and the computation will succeed. Depending on the version of > Sage you last tried this with the memory footprint of Sage might have > grown in relative terms so that the computation no longer succeeds > with the amount of RAM you have, so another possibility would be to > increase the amount of swap you have in that box. Adding more physical > RAM will probably makes the problem go away, too. >
I rebooted with no succes. Same error using 50-60% of 1 GB memory. Swap space is 5 GB. Jaap > If anything pops up from the still running debug sessions I will let > you know. > --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~----------~----~----~----~------~----~------~--~---
