Prerequisites

• Agent client versions:

  • Claude Code version 2.1 and later

  • OpenClaw version 2026.3.2 and later

  • VSCode with Github Copilot activated

• Environment preparations:

  • Preinstall the torch, triton, and pytest packages to enable Kernel testing across different hardware platforms. Import torch through the following python code to check whether CUDA is available:

    import torch
    
    # Check if CUDA is available
    print("CUDA available:", torch.cuda.is_available())
    
    # If available, perform a simple GPU computation
    if torch.cuda.is_available():
        x = torch.tensor([1.0, 2.0, 3.0]).cuda()
        y = torch.tensor([4.0, 5.0, 6.0]).cuda()
        z = x + y  # Triggers CUDA kernel
        print("Result:", z)
        print("Device:", z.device)
    else:
        print("CUDA is not available")
    

    For Huawei Ascend, install and import torch_npu in addition to installing and importing standard torch package.

  • Preinstall FlagTree.