Read the code carefully

# Each function reports whether moving from 1 thread to 4 threads
# in CPython would yield a real speedup. The verdict is decided
# by the GIL rules, NOT by wall-clock measurement.

def cpu_bound():
    # Pure Python tight loop — GIL serializes every bytecode op.
    return "no_speedup"

def io_bound():
    # time.sleep / socket waits release the GIL.
    return "speedup"

def numpy_bound():
    # NumPy's C code releases the GIL during heavy array ops.
    return "speedup"

print(cpu_bound())
print(io_bound())
print(numpy_bound())

# What does this print? Type your prediction.

What will the program print? Write here: