Functional Architecture

MindStudio provides a one-stop environment for AI development and supports jobs for model, application, and operator development. With functions such as model visualization, computing power testing, and local simulation and debugging for the integrated development environment (IDE), MindStudio can help you complete AI application development efficiently.

Figure 1 shows the MindStudio functional framework.

Figure 1 Toolchain architecture

In a heterogeneous computing architecture, the Ascend AI Processor and CPU work together through a PCIe bus.

  • The host is an x86_64 server, an AArch64 server, or a Windows PC that is connected to the hardware device powered by the Ascend AI Processor. The host leverages the neural-network (NN) compute capability provided by the Ascend AI Processor to implement services. See the development platform layer in Figure 1.
  • A device is a hardware device powered by the Ascend AI Processor. It connects to a server over the PCIe interface and provides the NN compute capability for the server. See the processor layer in Figure 1.

For the Ascend RC form, since the hardware device powered by the Ascend AI Processor is co-deployed with the interconnected ARM server, the host and device sides are not distinguished and are referred to as the host.