浮点32数据和浮点64数据使用相同的内存？

Question

我曾尝试使用tracemalloc跟踪编码的内存使用情况，发现对于浮点32数据和浮点64数据，它们的内存使用峰值相同。为什么?

输出屏幕截图：

浮动64：

import tracemalloc
import numpy as np
tracemalloc.start()
arr = np.zeros((100000000,), dtype=np.float64)
for i in range(100000000):
    arr[i] = i
current, peak = tracemalloc.get_traced_memory()
#current memory is the memory the code is currently using and 
#peak memory is the maximum space the program used while executing.
tracemalloc.stop()
print(f"Current memory usage is {current / 10**6}MB; Peak was {peak / 10**6}MB")

输出：

Current memory usage is 800.14981MB; Peak was 800.160145MB

第32条：

import tracemalloc
import numpy as np
tracemalloc.start()
arr = np.zeros((100000000,), dtype=np.float32)
arr = arr.astype(np.float32)
for i in range(100000000):
    arr[i] = i
current, peak = tracemalloc.get_traced_memory()
tracemalloc.stop()
print(f"Current memory usage is {current / 10**6}MB; Peak was {peak / 10**6}MB")

输出：

Current memory usage is 400.149626MB; Peak was 800.149468MB

Answer 1

您可以看到，这两个代码段并不相同。在第二个代码段中有一个附加语句

arr = arr.astype(np.float32)

该语句（即使arr已经具有类型np.float32）：

• 复制阵列（占用的内存量与旧阵列相同）
• 然后将新数组分配给arr
• 然后丢弃旧数组

因此，峰值内存加倍