negative performance speedup with -fno-plt

[Farid Zakaria]

Hi!

I built Python with `-fno-plt` expecting to always see speedups (even
minor) since it's avoiding 1 jmp to the PLT.

I ran pyperformance benchmark[1] and for some benchmarks see a
negative speedup which is confusing. (The negative performance is
significant with a confidence interval 95%)

I'm comparing it against a Python built with BIND_NOW (-z,now) to
account for no-lazy binding in both. I'm a bit stumped at what could
be causing some negative speedups. Anyone got leads that might cause a
difference?

The benchmarks causing negative speedup for instance are: richards,
regex, coroutines

Appreciate any tips! Thank you.

[1] https://pyperformance.readthedocs.io/usage.html#installation