Closed AndrewVSutherland closed 8 years ago
I will either fix or hide these links.
The problem occurs in line 325 of siegel_modular_form.py when it is attempting to compute the ring of integers of a number field with large discriminant in sage. For the page http://127.0.0.1:37777/ModularForm/GSp/Q/Sp4Z_2.26_C it is the degree 8 field defined by
x^8 - 2*x^7 - 357101291310*x^6 + 52840539355719536*x^5 + 22640564812749741524120*x^4 - 4204304986232338333203658272*x^3 - 233023237121289760198239377830176*x^2 + 41633688176726774757701479405996976640*x + 1568887010644434880435299480241036148266368
I think the easiest fix is to simple remove those samples from the database where the ring of integers cannot be readily computed (even precomputing them seems problematic). The discriminant of the polynomial above does not appear to be easy to factor (so sage isn't even getting past this necessary first step), and polredbest does not improve this polynomial (the same comments apply to several others which are worse).
Fixed in #1577.
If you go to the page http://www.lmfdb.org/ModularForm/GSp/Q/Sp4Z_2/ and click on any of the Galois orbit links for weights 26, 28, or 30 in the first table the server appears to go into an infinite loop (and eventually times out with a server error). If you run the lmfdb locally and request the page
http://127.0.0.1:37777/ModularForm/GSp/Q/Sp4Z_2.26_C
it appears to get stuck in an infinite loop (or possibly just a very long computation) and will eventually consume a lot of memory (67 GB when I left it running overnight). It appears to get stuck inside the _maximal_order number field method in Sage which makes a call to _pari_integral_basis that does not return in less than 8 hours.
Below is the traceback you get on ctrl-C: