Python: How to reduce run time?

Question

I have a question how to reduce the run time.
The code I made is Python.
It takes a huge data set as input, process it, calculate and write output to an array.
Most calculations may be quite simple such as summation. In input file, there are about 100 millions of rows and 3 columns. The problem I faced is so large run time. How to reduce the run time?

Here is the code I wrote.
I need to write all new values (from GenePair to RM_pval with header) I calculated from to new file. Thank you so much in advance.

fi = open ('1.txt')
fo = open ('2.txt','w')

import math
def log(x):
    return math.log(x)

from math import sqrt

import sys
sys.path.append('/tools/lib/python2.7/site-packages')
import numpy
import scipy
import numpy as np
from scipy.stats.distributions import norm

for line in fi.xreadlines():
    tmp = line.split('\t')

    GenePair = tmp[0].strip()

    PCC_A = float(tmp[1].strip())
    PCC_B = float(tmp[2].strip())

    ZVAL_A = 0.5 * log((1+PCC_A)/(1-PCC_A))
    ZVAL_B = 0.5 * log((1+PCC_B)/(1-PCC_B))

    ABS_ZVAL_A = abs(ZVAL_A)
    ABS_ZVAL_B = abs(ZVAL_B)

    Var_A = float(1) / float(21-3) #SAMPLESIZE - 3
    Var_B = float(1) / float(18-3) #SAMPLESIZE - 3

    WT_A = 1/Var_A #float
    WT_B = 1/Var_B #float

    ZVAL_A_X_WT_A = ZVAL_A * WT_A #float
    ZVAL_B_X_WT_B = ZVAL_B * WT_B #float

    SumofWT = (WT_A + WT_B) #float
    SumofZVAL_X_WT = (ZVAL_A_X_WT_A + ZVAL_B_X_WT_B) #float

    #FIXED MODEL
    meanES = SumofZVAL_X_WT / SumofWT #float
    Var = float(1) / SumofWT #float
    SE = math.sqrt(float(Var)) #float
    LL = meanES - (1.96 * SE) #float
    UL = meanES - (1.96 * SE) #float
    z_score = meanES / SE #float
    p_val = scipy.stats.norm.sf(z_score)

    #CAL
    ES_POWER_X_WT_A = pow(ZVAL_A,2) * WT_A #float
    ES_POWER_X_WT_B = pow(ZVAL_B,2) * WT_B #float
    WT_POWER_A = pow(WT_A,2)
    WT_POWER_B = pow(WT_B,2)
    SumofES_POWER_X_WT = ES_POWER_X_WT_A + ES_POWER_X_WT_B
    SumofWT_POWER = WT_POWER_A + WT_POWER_B

    #COMPUTE TAU
    tmp_A = ZVAL_A - meanES
    tmp_B = ZVAL_B - meanES
    temp = pow(SumofZVAL_X_WT,2)

    Q = SumofES_POWER_X_WT - (temp /(SumofWT))      
    if PCC_A !=0 or PCC_B !=0:
        df = 0
    else:
        df = 1

    c = SumofWT - ((pow(SumofWT,2))/SumofWT)
    if c == 0:
        tau_square = 0
    else:
        tau_square = (Q - df) / c

    #calculation
    Var_total_A = Var_A + tau_square
    Var_total_B = Var_B + tau_square

    WT_total_A = float(1) / Var_total_A
    WT_total_B = float(1) / Var_total_B

    ZVAL_X_WT_total_A = ZVAL_A * WT_total_A
    ZVAL_X_WT_total_B = ZVAL_B * WT_total_B

    Sumoftotal_WT = WT_total_A + WT_total_B
    Sumoftotal_ZVAL_X_WT= ZVAL_X_WT_total_A + ZVAL_X_WT_total_B


    #RANDOM MODEL
    RM_meanES = Sumoftotal_ZVAL_X_WT / Sumoftotal_WT
    RM_Var = float(1) / Sumoftotal_WT
    RM_SE = math.sqrt(float(RM_Var))
    RM_LL = RM_meanES - (1.96 * RM_SE)
    RM_UL = RM_meanES + (1.96 * RM_SE)
    RM_z_score = RM_meanES / RM_Var
    RM_p_val = scipy.stats.norm.sf(RM_z_score)

Answer 1

Definitely do the profiler thing but... I think the only major speedup will happen due to parallelism. Taking advantage of multiple cores is of paramount importance if you are going to run cpu bound problems like this. Try putting each line through a different (thread/process). This raises more questions of course, for example does the data need to be in the same order from the input file? if so just enumerate that and stick a second variable on the big_hairy_func for which line it will be.

here is some boilerplate code to get started

notes:

xreadlines is deprecated even though it deals with large files for line in file: replaces it.

fi = open('1.txt')
fo = open('2.txt','w')

import math
def log(x):
    return math.log(x)

from math import sqrt

import multiprocessing as mp
import sys
sys.path.append('/tools/lib/python2.7/site-packages')
import scipy
import numpy as np
from scipy.stats.distributions import norm

def big_hairy_func(linefromfile):
    <majority of your post here>
    return <whatever data you were going to write to 'fo'>

if __name__ == '__main__':
    pool = mp.Pool(4) #rule of thumb.  Replace '4' with the number of cores on your system
    result = pool.map(big_hairy_func, (input for input in fi.readlines()))
    <write the result to fo that you haven't posted>

xreadlines was deprecated in python 2.3 so with that version I'm not sure if the generator function will work. Let me know if you have questions about compatibility with your version of python.