A simple script to visualize the memory usage on Windows by accumulation.

memview.py

import subprocess as subp
import pandas as pd
from matplotlib import pyplot as plt

p = subp.Popen(['tasklist', '/fo', 'csv'], stdout=subp.PIPE)
df = pd.read_csv(p.stdout, encoding='gbk')
df['mem'] = df['内存使用 '].transform(lambda x: int(x[:-2].replace(',', '')))
df['proc'] = df['映像名称']
df = df.groupby('proc')['mem'] \
        .agg(['sum', 'count']) \
        .rename(columns={'sum': 'mem'}) \
        .sort_values('mem') \
        .reset_index()
df['tot'] = df['mem'].cumsum()
ax = df.plot('mem', 'tot', 'scatter')
for it in df.tail(20).iterrows():
    idx, it = it
    label = '{}, {}'.format(it['proc'], it['count']) \
        if it['count'] > 1 else it['proc']
    ax.annotate(label, (it['mem'], it['tot']))
plt.show()