WSL2
Windows Subsystem for Linux 2, a Microsoft technology that runs a complete Linux kernel within the Windows operating system without a graphical interface. It is commonly used by developers to run development toolchains.
Core Facts
Docker Desktop chose to run on top of WSL2.
Unlike WSL1, WSL2 runs a full Linux kernel and is powered at the underlying level by a Hyper-V lightweight virtual machine.
WSL2 installation requires Windows 10 version 2004 or above, or Windows 11.
Timeline (last 90 days)
Cloud APIs typically have a first-token latency of 200-500 ms, and BitRun's roughly 130 ms on WSL2 is already usable.
On Linux x86_64 (WSL2/Ubuntu, AVX2), BitRun has a first-token latency of about 130 ms and a generation speed of 40-45 t/s.
Running tools in WSL2 can borrow the Linux side's sandboxing capabilities, but the cost is that the file system is split into two, and path mapping, permission bits and symlink behavior may be inconsistent.
The experimental machine configuration was a Lenovo IdeaPad gaming laptop with an Intel i7-10750H, 16GB of RAM, and a GTX 1650 with 4GB of VRAM, running PyTorch and CUDA under WSL2
Mainstream implementation methods for bundled runtimes include embedding a lightweight container engine (such as Podman in daemonless mode), using a WebAssembly sandbox, or relying on platform-native virtualization capabilities (such as macOS's Virtualization.framework or Windows's WSL2).
ODS supports NVIDIA and AMD on Windows via WSL2, and natively supports Apple Silicon on macOS (using Metal).
The WSL2 DXG bridge solution accesses the GPU through the DirectX Graphics kernel bridge layer without relying on the traditional /dev/kfd device node.
Windows users who need Claude Code's full sandbox should preferably run it inside WSL2.
The new version adds an AMD ROCm Docker image containing Unsloth Studio and JupyterLab, and the ROCm image can access the GPU via WSL2's DXG bridge.
WSL2's DXG bridge allows the Windows Subsystem for Linux to access the host GPU via DirectX drivers, bypassing the traditional ROCm path that relies on the /dev/kfd character device.
13 more timeline events
All Facts (20)
Docker Desktop chose to run on top of WSL2.
80%VerifiedUnlike WSL1, WSL2 runs a full Linux kernel and is powered at the underlying level by a Hyper-V lightweight virtual machine.
80%Partially VerifiedWSL2 installation requires Windows 10 version 2004 or above, or Windows 11.
100%UnverifiedMagentic-UI部署需要Python 3.10+（推荐3.12）和Docker Desktop，Windows用户需配置WSL2
95%UnverifiedWSL2的每个Linux发行版以ext4格式的虚拟硬盘文件（.vhdx）存储在Windows文件系统中，默认路径位于C盘的%LOCALAPPDATA%\Packages目录下
90%UnverifiedClaude Code要求Node.js 18及以上版本，支持macOS和Linux操作系统，Windows用户需使用WSL2
90%Unverified微软与英伟达合作开发了WDDM GPU虚拟化技术，允许WSL2内的Linux进程直接访问宿主机的GPU资源
85%UnverifiedThe new version adds an AMD ROCm Docker image containing Unsloth Studio and JupyterLab, and the ROCm image can access the GPU via WSL2's DXG bridge.
70%UnverifiedWSL2的文件系统性能相比WSL1提升了20倍以上
70%UnverifiedWSL2基于轻量级Hyper-V虚拟机实现，GPU虚拟化性能损耗通常低于5%
70%UnverifiedCloud APIs typically have a first-token latency of 200-500 ms, and BitRun's roughly 130 ms on WSL2 is already usable.
50%UnverifiedOn Linux x86_64 (WSL2/Ubuntu, AVX2), BitRun has a first-token latency of about 130 ms and a generation speed of 40-45 t/s.
50%UnverifiedRunning tools in WSL2 can borrow the Linux side's sandboxing capabilities, but the cost is that the file system is split into two, and path mapping, permission bits and symlink behavior may be inconsistent.
50%UnverifiedThe experimental machine configuration was a Lenovo IdeaPad gaming laptop with an Intel i7-10750H, 16GB of RAM, and a GTX 1650 with 4GB of VRAM, running PyTorch and CUDA under WSL2
50%UnverifiedMainstream implementation methods for bundled runtimes include embedding a lightweight container engine (such as Podman in daemonless mode), using a WebAssembly sandbox, or relying on platform-native virtualization capabilities (such as macOS's Virtualization.framework or Windows's WSL2).
50%UnverifiedODS supports NVIDIA and AMD on Windows via WSL2, and natively supports Apple Silicon on macOS (using Metal).
50%UnverifiedThe WSL2 DXG bridge solution accesses the GPU through the DirectX Graphics kernel bridge layer without relying on the traditional /dev/kfd device node.
50%UnverifiedWindows users who need Claude Code's full sandbox should preferably run it inside WSL2.
50%UnverifiedWSL2's DXG bridge allows the Windows Subsystem for Linux to access the host GPU via DirectX drivers, bypassing the traditional ROCm path that relies on the /dev/kfd character device.
50%UnverifiedOn Windows, Docker runs Linux containers through WSL2 or Hyper-V.
50%