Using time_fourier_transform on a single-core machine

[salmanarif]

When using time_fourier_transform on a single core machine (in this case the free AWS instance), the program ends with the following error:

std::bad_alloc

test_fourier_transform runs fine, and both time_fourier_transform and test_fourier_transform run fine on multi-core machines. Is this error due to the limitations of the processing power of the machine?

[m8pple]

Ah, yes. You're hitting the memory limit, so when it tries to allocate the buffer for a big transform
it throws an exception. IIRC the free instances are limited to 1GB, so you'll hit it quite quickly. Does
it at least get some of the smaller transforms done?