Benchmark of Error-Free Transforms for summation

This benchmark compares accuracy and efficiency of several summation algorithms in floating point arithmetics

def naive_sum(x):
    return choreo.scipy_plus.cython.eft_lib.FastSumK(x,0)

def builtin_sum(x):
    return sum(x)

def np_sum(x):
    return np.sum(x)

def m_fsum(x):
    return m.fsum(x)

def SumK_1(x):
    return choreo.scipy_plus.cython.eft_lib.SumK(x,1)

def SumK_2(x):
    return choreo.scipy_plus.cython.eft_lib.SumK(x,2)

def SumK_3(x):
    return choreo.scipy_plus.cython.eft_lib.SumK(x,3)

def FastSumK_1(x):
    return choreo.scipy_plus.cython.eft_lib.FastSumK(x,1)

def FastSumK_2(x):
    return choreo.scipy_plus.cython.eft_lib.FastSumK(x,2)

def FastSumK_3(x):
    return choreo.scipy_plus.cython.eft_lib.FastSumK(x,3)

def prepare_x(n):
    x = np.random.random(n)
    return [(x, 'x')]