What is CUDA?
Why D is the answer
NVIDIA's parallel computing platform and programming model
CUDA lets software use GPUs for general computing.
In detail
- CUDA lets software use GPUs for general computing.
- Many AI libraries build on it.
Use case
An illustrative scenario, not a real incident.
A developer uses CUDA to run general computing on a GPU, not only graphics.
See it in different ways
flowchart TD A["Applications"] --> B["NVIDIA AI Enterprise and frameworks"] B --> C["CUDA libraries"] C --> D["GPU driver"] D --> E["GPU"]
stateDiagram-v2 [*] --> Applications Applications --> AiEnterprise AiEnterprise --> Cuda Cuda --> Driver Driver --> Gpu Gpu --> [*]
flowchart LR S0["Application"] --> S1["CUDA libraries"] --> S2["Driver"] --> S3["GPU"]
Key point: CUDA is NVIDIA's parallel platform.
Why the other options are wrong
- A. A networking protocol for connecting data center switches
- It is not a network protocol.
- B. A cooling technique used in high-density data centers
- It is not cooling.
- C. A storage format for labeled training datasets
- It is not a storage format.