考虑以下网络:
%%time
import torch
from torch.autograd import grad
import torch.nn as nn
import torch.optim as optim
class net_x(nn.Module):
def __init__(self):
super(net_x, self).__init__()
self.fc1=nn.Linear(1, 20)
self.fc2=nn.Linear(20, 20)
self.out=nn.Linear(20, 400) #a,b,c,d
def forward(self, x):
x=torch.tanh(self.fc1(x))
x=torch.tanh(self.fc2(x))
x=self.out(x)
return x
nx = net_x()
#input
val = 100
t = torch.rand(val, requires_grad = True) #input vector
t = torch.reshape(t, (val,1)) #reshape for batch
#method
dx = torch.autograd.functional.jacobian(lambda t_: nx(t_), t)
这个输出
CPU times: user 11.1 s, sys: 3.52 ms, total: 11.1 s
Wall time: 11.1 s
但是,当我改为使用带有.to(device)
的GPU时:
%%time
import torch
from torch.autograd import grad
import torch.nn as nn
import torch.optim as optim
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
class net_x(nn.Module):
def __init__(self):
super(net_x, self).__init__()
self.fc1=nn.Linear(1, 20)
self.fc2=nn.Linear(20, 20)
self.out=nn.Linear(20, 400) #a,b,c,d
def forward(self, x):
x=torch.tanh(self.fc1(x))
x=torch.tanh(self.fc2(x))
x=self.out(x)
return x
nx = net_x()
nx.to(device)
#input
val = 100
t = torch.rand(val, requires_grad = True) #input vector
t = torch.reshape(t, (val,1)).to(device) #reshape for batch
#method
dx = torch.autograd.functional.jacobian(lambda t_: nx(t_), t)
这将产生:
CPU times: user 18.6 s, sys: 1.5 s, total: 20.1 s
Wall time: 19.5 s
更新1: 检查将输入和模型移动到设备的过程的计时:
%%time
nx.to(device)
t.to(device)
这将产生:
CPU times: user 2.05 ms, sys: 0 ns, total: 2.05 ms
Wall time: 2.13 ms
更新2:
看来@Gulzar是对的。我将批处理大小更改为1000(val=1000
),CPU输出:
Wall time: 8min 44s
而GPU输出:
Wall time: 3min 12s
摇摆不定的回答
GPU是“较弱”的计算机,其计算核心比CPU多得多。
数据必须每隔一段时间以“昂贵”的方式从RAM内存传递到GRAM,这样他们才能处理数据
如果数据是“大的”,并且可以对该数据进行并行处理,那么那里的计算速度可能会更快
如果数据不够大,那么传输数据的成本,或者使用较弱的内核并同步它们的成本可能会超过并行化的好处
GPU什么时候有用
相关问题 更多 >
编程相关推荐