在之前配置各种深度学习环境的时候经常需要先检测一下电脑的软硬件环境，其实整个过程比较重复和固定，所以趁着空闲时间就简单写了个用于电脑环境检测的脚本。检测内容主要包括Python版本、PIP版本、Conda版本、CUDA版本、电脑系统、CPU核数、CPU频率、内存、硬盘等内容。
1.实现原理与工具
整个实现的原理和工具都比较简单，对于软件环境检测部分，主要是利用不同库查询版本的命令实现的。具体来说就是利用Python执行版本查询命令并接收返回的文本结果，在返回的文本中提取版本信息。主要使用getstatusoutput()函数，需要注意的是在Python2中这个函数在commands包中，在Python3中移到了subprocess包中。对于一些硬件信息的检测主要通过psutil包实现，如果没有安装的话直接pip install psutil即可安装。
2.实现代码
完整代码如下：
import sys
import platform
def checkPythonVersion():
return float(sys.version[:3])
def checkOS():
os_info = platform.system()
return os_info
def checkPipVersionPy3():
import subprocess
code, pip_version = subprocess.getstatusoutput("pip --version")
if code != 0:
pip_version = 0
else:
pip_version = pip_version.split("pip")[1].split("from")[0].strip()
parts = pip_version.split(".")
pip_version = float(parts[0] + "." + parts[1])
return pip_version
def checkPipVersionPy2():
import commands
code, pip_version = commands.getstatusoutput("pip --version")
if code != 0:
pip_version = 0
else:
pip_version = pip_version.split("pip")[1].split("from")[0].strip()
parts = pip_version.split(".")
pip_version = float(parts[0] + "." + parts[1])
return pip_version
def checkCUDAPy3():
import subprocess
code, nvcc_info = subprocess.getstatusoutput("nvcc --version")
if code != 0:
nvcc_info = 0
else:
nvcc_info = float(nvcc_info.split("release")[1].split(",")[0].strip())
return nvcc_info
def checkCUDAPy2():
import commands
code, nvcc_info = commands.getstatusoutput("nvcc --version")
if code != 0:
nvcc_info = 0
else:
nvcc_info = float(nvcc_info.split("release")[1].split(",")[0].strip())
return nvcc_info
def checkCondaPy3():
import subprocess
code, conda_info = subprocess.getstatusoutput("conda --version")
if code != 0:
conda_info = 0
else:
parts = conda_info.split("conda")[1].split(".")
conda_info = float(parts[0] + "." + parts[1])
return conda_info
def checkCondaPy2():
import commands
code, conda_info = commands.getstatusoutput("conda --version")
if code != 0:
conda_info = 0
else:
parts = conda_info.split("conda")[1].split(".")
conda_info = float(parts[0] + "." + parts[1])
return conda_info
def cpuCount():
try:
import psutil
number = psutil.cpu_count()
return str(number)
except:
return "No 'psutil' on your computer,install it with 'pip install psutil'"
def cpuFreq():
try:
import psutil
info = psutil.cpu_freq()
info = str(round(float(info[2]) / 1000, 2))
return str(info)
except:
return "No 'psutil' on your computer,install it with 'pip install psutil'"
def cpuUsage():
try:
import psutil
info = psutil.cpu_percent(interval=1, percpu=True)
return info
except:
return "No 'psutil' on your computer,install it with 'pip install psutil'"
def memSize():
try:
import psutil
info = psutil.virtual_memory()
info = str(round(float(info[0]) / pow(1024, 3), 2))
return str(info)
except:
return "No 'psutil' on your computer,install it with 'pip install psutil'"
def memInfo():
try:
import psutil
info = psutil.virtual_memory()
usage = str(info[2])
avail = str(round(float(info[1]) / pow(1024, 3), 2))
return usage, avail
except:
return "No 'psutil' on your computer,install it with 'pip install psutil'"
def diskSize():
try:
import psutil
disk_size = []
for item in psutil.disk_partitions():
mountpoint = item[1]
opts = item[3]
if opts == 'cdrom':
continue
size = psutil.disk_usage(mountpoint)
size = round(float(size[0]) / pow(1024, 3), 2)
disk_size.append(size)
total_size = round(sum(disk_size), 2)
return str(total_size)
except:
return "No 'psutil' on your computer,install it with 'pip install psutil'"
def diskInfo():
try:
import psutil
part_size = []
part_usage = []
part_name = []
part_fstype = []
for item in psutil.disk_partitions():
name = item[0]
mountpoint = item[1]
fstype = item[2]
opts = item[3]
if opts == 'cdrom':
continue
info = psutil.disk_usage(mountpoint)
size = round(float(info[0]) / pow(1024, 3), 2)
part_size.append(size)
part_usage.append(str(info[3]) + "%")
part_name.append(name)
part_fstype.append(fstype)
return part_name, part_size, part_usage, part_fstype
except:
return "No 'psutil' on your computer,install it with 'pip install psutil'"
if __name__ == '__main__':
python_v = checkPythonVersion()
os = checkOS()
if python_v > 3:
pip_v = checkPipVersionPy3()
cuda_v = checkCUDAPy3()
conda_v = checkCondaPy3()
else:
pip_v = checkPipVersionPy2()
cuda_v = checkCUDAPy2()
conda_v = checkCondaPy2()
print("=======================System Information=======================")
print("* This computer runs " + os + " system.")
print("* Python version(running this script):" + python_v.__str__())
if pip_v == 0:
print("* Pip version:It seems no pip on your computer.")
else:
print("* Pip version:" + pip_v.__str__())
if cuda_v == 0:
print("* CUDA version:It seems no CUDA on your computer.")
else:
print("* CUDA version:" + cuda_v.__str__())
if conda_v == 0:
print("* Conda version:It seems no Conda on your computer.")
else:
print("* Conda version:" + conda_v.__str__())
print("=======================System Information=======================")
cpu_count = cpuCount()
cpu_freq = cpuFreq()
cpu_usage = cpuUsage()
mem_size = memSize()
mem_usage = memInfo()
disk_size = diskSize()
disk_info = diskInfo()
print("======================Hardware Information======================")
print("* CPU kernel:" + cpu_count)
print("* CPU base frequency:" + cpu_freq + " GHz")
if type(cpu_usage) is str:
print("* CPU use percentage(current,every kernel):\n " + cpu_usage)
else:
print("* CPU use percentage(current,every kernel):\n " + str(cpu_usage))
print("* Memory total size:" + mem_size + " GB")
if type(mem_usage) is str:
print("* Memory use percentage(current):" + mem_usage)
else:
print("* Memory use percentage(current):" + mem_usage[0] + "%, free:" + mem_usage[1] + " GB")
print("* Disk total size:" + disk_size + " GB")
print("* Disk partion info:")
if type(disk_info) is str:
print(" " + disk_info)
else:
print(" Identifier\tTotal size(GB)\tUsage(percentage)\tFile format")
for i in range(len(disk_info)):
print(" " + disk_info[0][i] + "\t\t\t" + str(disk_info[1][i]) + "\t\t\t" + str(
disk_info[2][i]) + "\t\t\t\t" +
disk_info[3][i])
print("======================Hardware Information======================")
在不同电脑环境的测试结果如下。
这是在我台式机的Windows系统上的测试结果。
这是在我笔记本的Ubuntu上的测试结果。
最后，按照惯例，代码放到了Github上，点击查看。
本文作者原创,未经许可不得转载,谢谢配合
