We use cookies to measure traffic and show relevant ads. By clicking Accept you agree to our use of analytics and advertising cookies.
AMD 将 ROCm 统一为 OneROCm，用于所有加速器。Triton 为 AMD/Nvidia 编写内核。vLLM 推理与 CUDA 相当。六周发布周期。在 Strix Halo 上测试堆栈。
