python多线程模型问题

写了一个简单的多线程程序
(python 2.4)
（4核server）

#!/usr/bin/env python
# -*- coding:gb2312 -*-


import sys
import time
import threading
import os
def run(arg):
        pid = os.getpid()
        print "This thread id:%d" % pid
        print "msg:",arg
        ret = 0
        while True:
                #time.sleep(1)
                ret += 1


print "Parent pid: %d" % os.getpid()
for i in range(0,4):
        msg = "index %d" % i
        ps = threading.Thread(target = run,args =(msg,))
        ps.start()

   
输出：
Parent pid: 7749
This thread id:7751
msg: index 0
This thread id:7752
msg: index 1
This thread id:7753
msg: index 2
This thread id:7754
msg: index 3

pstree -p 输出：
        |-sshd2(18660)-+-sshd2(31624)---bash(31626)---python(7549)---python(7550)-+-python(7551)
        |              |                                                          |-python(7552)
        |              |                                                          |-python(7553)
        |              |                                                          `-python(7554)
        |              `-sshd2(3634)---bash(3636)---pstree(7555)

top,每个线程占用30左右的cpu

问题：
1.python线程模型是哪种？Linux 线程模型的比较：LinuxThreads 和NPTL
2.为什么每个线程系统占用率不是很高（增加到100线程，每个线程占用 cpu比例为2%）