WSL installation of CUDA discrete graphics PyTorch deep learning environment configuration, why use WSL for speed testing?

Preface#

In order to preview the junior year courses, I wanted to learn PyTorch in advance.
So I came across the amazing tutorial “Dive into Deep Learning”, and used it as a reference to complete the environment setup. Thanks to the contributors for their selfless dedication! Thanks♪(･ω･)ﾉ

This tutorial demonstrates the environment setup for running PyTorch deep learning on a Windows computer with a dedicated GPU using the WSL Ubuntu subsystem. As for why to use WSL, I found through testing that the performance is much stronger than Windows. Due to space limitations, the WSL installation process is omitted.

Installation#

Device:#

Windows 11: WSL-Ubuntu-22.04
1660ti-6g dedicated GPU

Miniconda Installation#

Download the corresponding Linux version using wget from the Miniconda official website

wget https://repo.anaconda.com/miniconda/Miniconda3-py310_23.3.1-0-Linux-x86_64.sh

Use the sh command for default installation

sh Miniconda3-py310_23.3.1-0-Linux-x86_64.sh -b

Initialize the environment

~/miniconda3/bin/conda init

You will be prompted to close this Terminal and open a new one.
Create a new environment, the name d2l can be changed.

conda create --name d2l python=3.9 -y

Activate the environment

conda activate d2l

Note that you need to run this command every time to switch to the d2l environment.

To exit the current environment: conda deactivate.
To completely delete the environment named d2l: conda remove -n d2l --all

CUDA Installation#

CUDA (official website) is NVIDIA's official deep learning toolkit, as shown in the WSL options. Run the download commands.
cuda_download

wget https://developer.download.nvidia.com/compute/cuda/repos/wsl-ubuntu/x86_64/cuda-wsl-ubuntu.pin
sudo mv cuda-wsl-ubuntu.pin /etc/apt/preferences.d/cuda-repository-pin-600
wget https://developer.download.nvidia.com/compute/cuda/12.1.1/local_installers/cuda-repo-wsl-ubuntu-12-1-local_12.1.1-1_amd64.deb
sudo dpkg -i cuda-repo-wsl-ubuntu-12-1-local_12.1.1-1_amd64.deb
sudo cp /var/cuda-repo-wsl-ubuntu-12-1-local/cuda-*-keyring.gpg /usr/share/keyrings/
sudo apt-get update
sudo apt-get -y install cuda

After adding, you need to update the ~/.bashrc file

sudo vi ~/.bashrc

Press i to enter insert mode, add the following code to the end of the file, and be sure to modify it to the corresponding version.

export PATH=/usr/local/cuda-12.1/bin${PATH:+:${PATH}}
export LD_LIBRARY_PATH=/usr/local/cuda-12.1/lib64\
                         ${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}

Press Esc, then type :wq and hit enter to save.

source ~/.bashrc

Run the following command, if the output is as shown in the image, then CUDA installation is successful.

nvcc -V

nvcc_v_printout

Install PyTorch Framework#

For the GPU version, you need to select it from the PyTorch official website. I downloaded the Preview version which happens to be compatible with CUDA 12.1 as shown in the image, but it also has good backward compatibility.
pytorch_download

conda install pytorch torchvision torchaudio pytorch-cuda=12.1 -c pytorch-nightly -c nvidia

Download the d2l package

pip install d2l==0.17.6

GPU Testing#

Clone the test files

git clone https://github.com/pytorch/examples.git
cd examples/mnist/

Replace the content of the main.py file with the following:

from __future__ import print_function
import argparse
import torch
import time
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
from torchvision import datasets, transforms
from torch.optim.lr_scheduler import StepLR


class Net(nn.Module):
    def __init__(self):
        super(Net, self).__init__()
        self.conv1 = nn.Conv2d(1, 32, 3, 1)
        self.conv2 = nn.Conv2d(32, 64, 3, 1)
        self.dropout1 = nn.Dropout(0.25)
        self.dropout2 = nn.Dropout(0.5)
        self.fc1 = nn.Linear(9216, 128)
        self.fc2 = nn.Linear(128, 10)

    def forward(self, x):
        x = self.conv1(x)
        x = F.relu(x)
        x = self.conv2(x)
        x = F.relu(x)
        x = F.max_pool2d(x, 2)
        x = self.dropout1(x)
        x = torch.flatten(x, 1)
        x = self.fc1(x)
        x = F.relu(x)
        x = self.dropout2(x)
        x = self.fc2(x)
        output = F.log_softmax(x, dim=1)
        return output


def train(args, model, device, train_loader, optimizer, epoch):
    model.train()
    for batch_idx, (data, target) in enumerate(train_loader):
        data, target = data.to(device), target.to(device)
        optimizer.zero_grad()
        output = model(data)
        loss = F.nll_loss(output, target)
        loss.backward()
        optimizer.step()
        if batch_idx % args.log_interval == 0:
            print('Train Epoch: {} [{}/{} ({:.0f}%)]\tLoss: {:.6f}'.format(
                epoch, batch_idx * len(data), len(train_loader.dataset),
                100. * batch_idx / len(train_loader), loss.item()))
            if args.dry_run:
                break


def test(model, device, test_loader):
    model.eval()
    test_loss = 0
    correct = 0
    with torch.no_grad():
        for data, target in test_loader:
            data, target = data.to(device), target.to(device)
            output = model(data)
            test_loss += F.nll_loss(output, target, reduction='sum').item()  # sum up batch loss
            pred = output.argmax(dim=1, keepdim=True)  # get the index of the max log-probability
            correct += pred.eq(target.view_as(pred)).sum().item()

    test_loss /= len(test_loader.dataset)

    print('\nTest set: Average loss: {:.4f}, Accuracy: {}/{} ({:.0f}%)\n'.format(
        test_loss, correct, len(test_loader.dataset),
        100. * correct / len(test_loader.dataset)))


def main():
    # Training settings
    parser = argparse.ArgumentParser(description='PyTorch MNIST Example')
    parser.add_argument('--batch-size', type=int, default=64, metavar='N',
                        help='input batch size for training (default: 64)')
    parser.add_argument('--test-batch-size', type=int, default=1000, metavar='N',
                        help='input batch size for testing (default: 1000)')
    parser.add_argument('--epochs', type=int, default=14, metavar='N',
                        help='number of epochs to train (default: 14)')
    parser.add_argument('--lr', type=float, default=1.0, metavar='LR',
                        help='learning rate (default: 1.0)')
    parser.add_argument('--gamma', type=float, default=0.7, metavar='M',
                        help='Learning rate step gamma (default: 0.7)')
    parser.add_argument('--no-cuda', action='store_true', default=False,
                        help='disables CUDA training')
    parser.add_argument('--no-mps', action='store_true', default=False,
                        help='disables macOS GPU training')
    parser.add_argument('--dry-run', action='store_true', default=False,
                        help='quickly check a single pass')
    parser.add_argument('--seed', type=int, default=1, metavar='S',
                        help='random seed (default: 1)')
    parser.add_argument('--log-interval', type=int, default=10, metavar='N',
                        help='how many batches to wait before logging training status')
    parser.add_argument('--save-model', action='store_true', default=False,
                        help='For Saving the current Model')
    args = parser.parse_args()
    use_cuda = not args.no_cuda and torch.cuda.is_available()
    use_mps = not args.no_mps and torch.backends.mps.is_available()

    torch.manual_seed(args.seed)

    if use_cuda:
        device = torch.device("cuda")
    elif use_mps:
        device = torch.device("mps")
    else:
        device = torch.device("cpu")

    train_kwargs = {'batch_size': args.batch_size}
    test_kwargs = {'batch_size': args.test_batch_size}
    if use_cuda:
        cuda_kwargs = {'num_workers': 5,	# Number of threads
                       'pin_memory': True,
                       'shuffle': True}
        train_kwargs.update(cuda_kwargs)
        test_kwargs.update(cuda_kwargs)

    transform=transforms.Compose([
        transforms.ToTensor(),
        transforms.Normalize((0.1307,), (0.3081,))
        ])
    dataset1 = datasets.MNIST('../data', train=True, download=True,
                       transform=transform)
    dataset2 = datasets.MNIST('../data', train=False,
                       transform=transform)
    train_loader = torch.utils.data.DataLoader(dataset1,**train_kwargs)
    test_loader = torch.utils.data.DataLoader(dataset2, **test_kwargs)

    model = Net().to(device)
    optimizer = optim.Adadelta(model.parameters(), lr=args.lr)

    scheduler = StepLR(optimizer, step_size=1, gamma=args.gamma)
    start_time = time.time()
    for epoch in range(1, args.epochs + 1):
        train(args, model, device, train_loader, optimizer, epoch)
        test(model, device, test_loader)
        scheduler.step()
    end_time = time.time()
    total_time = end_time - start_time
    print(f"Total time: {total_time:.2f} seconds.")
    if args.save_model:
        torch.save(model.state_dict(), "mnist_cnn.pt")


if __name__ == '__main__':
    main()

This uses 5 threads, which can be modified as needed.
Run the corresponding mode for speed testing; depending on the configuration, the CPU and GPU running codes may be exactly opposite.

$ python main.py	# CPU mode
$ CUDA_VISIBLE_DEVICES=2 python main.py  # GPU mode

Source and reference for the test code:
pytorch/examples - github
Deep Learning: Performance Comparison of Windows 11 VS WSL2 VS Ubuntu, PyTorch 2.0 Performance Testing!

Completion!#

The environment setup is complete, and I will continue to follow “Dive into Deep Learning”.

mkdir d2l-zh && cd d2l-zh
curl https://zh-v2.d2l.ai/d2l-zh-2.0.0.zip -o d2l-zh.zip
unzip d2l-zh.zip && rm d2l-zh.zip
cd pytorch

。。。。。