After integrating Linux into the Windows 10 OS ecosystem earlier this year, NVIDIA has now showcased how its CUDA Cores will help the Windows Subsystem for Linux (WSL). NVIDIA’s digital keynote demonstrated the company’s GPU-accelerated support via CUDA on WSL. Additionally, NVIDIA also demonstrated how AI frameworks run as Linux executables on Microsoft Windows platforms.
NVIDIA and Microsoft have opened up Public Preview accessibility for those registered in the NVIDIA Developer Program and the Microsoft Windows Insider Program. Members can access the CUDA Driver/Container package via Downloads and Documentation from NVIDIA’s CUDA on our WSL page.
/dev/dri 経由じゃなくて WSL 専用ドライバなのねhttps://t.co/Qy0UJZ6xtB
— :t6l: (@_1_k) June 17, 2020
How Microsoft’s Windows Subsystem for Linux (WSL) Enables Windows Users To Efficiently Run And Access Linux:
Microsoft’s Windows Subsystem for Linux (WSL) enables Windows users to run native, unmodified Linux command-line tools directly on Windows. WSL was first introduced at Microsoft Build 2016. The company claims more than 3.5 million monthly active devices now run WSL. Now Microsoft and NVIDIA have imparted the ability to benefit from AI Framework and GPU-accelerated compute capabilities.
The Windows Subsystem for Linux 2 or WSL2 can now perform GPU compute functions, including using NVIDIA’s CUDA technology. The new features deliver on the promises Microsoft made at this year’s May Build 2020 conference. Incidentally, the company had previewed Graphical User Interface (GUI) support for the Windows Subsystem for Linux. But the feature is still missing.
🔥GPU accelerated machine learning (#ML) training inside the Windows Subsystem for Linux (#WSL)👉https://t.co/exHGGE6nGy
✔️Enabling professionals through NVIDIA CUDA support
✔️Empowering students and beginners through DirectML
📌GPU accelerated ML training https://t.co/OPRF4zN9XS pic.twitter.com/q5JmYBxepS
— 🔮WZor👁️ #StopKillingPeople! #AllLivesMatter✌️ (@WZorNET) June 17, 2020
Specifically speaking, the initial preview of NVIDIA’s CUDA GPU Compute for WSL2 includes machine-learning support for ML tools, libraries, and popular frameworks, including PyTorch and TensorFlow. Microsoft is also providing a preview package of TensorFlow with a DirectML backend.
Interestingly, AMD is eagerly supporting WSL as well. The company released a new graphics driver preview unlocking DirectX 12 compatible GPU-acceleration within the WSL. The driver is compatible with a wide range of AMD Radeon Graphics hardware.
Microsoft has now implemented GPU compute support, which was present in natively installed Linux distributions for close to a decade. Nonetheless, WSL2 is rapidly and eagerly attempting to offer all of the capabilities that a native Linux installation does. Speaking about the capabilities, Microsoft wrote, “This preview will initially support artificial intelligence (AI) and machine learning (ML) workflows, enabling professionals and students alike to run ML training workloads across the breadth of GPUs in the Windows ecosystem.”
Microsoft Simplifies The WSL2 Installation And Usage:
Windows 10 Build 20150 contains new features that attempt to simplify the process even further. Interested users need to merely type wsl.exe -install as a command, and Windows 10 will complete all the procedures. Similarly, users can type in wsl.exe -update to update the Linux kernel, wsl.exe -update -status to check the status of the kernel, and wsl.exe -update -rollback to roll back the kernel to a previous version.
NVIDIA and @Microsoft have released a public preview for CUDA on Windows Subsystem for Linux (WSL) – enabling users to benefit from #GPU-accelerated infrastructure, software, and container support on Windows.
— NVIDIA HPC Developer (@NVIDIAHPCDev) June 17, 2020
WSL’s improvements are part of Windows 10 Build 20150, which is a part of the Dev Channel of Insider builds. The Dev Channel is essentially or formerly referred to as the Fast Ring channel which is dedicated to testing new features that aren’t necessarily tied to any upcoming Windows 10 feature release. In other words, features introduced in the Dev Channel may or may not be introduced in the immediate stable release of Windows 10.
While WSL2 essentially allows users to run a Linux kernel from within Windows, the platform doesn’t perform quite as well as a full-fledged Linux distribution running natively on the hardware. However, Microsoft’s efforts are geared towards improving the experience and performance. Incidentally, WSL2 isn’t natively integrated into Windows. It is available as an optional feature. Users can choose to enable the feature using the “Turn Windows features on and off” Control Panel within Windows 10. In the current iteration, Windows 10 will download the necessary files and ask you to reboot the PC to complete the installation.