GPU Computing Accelerated
GPU computing accelerated GPU cloud servers use the GPU pass-through technology and are mainly suitable for AI deep learning & training, inference, scientific computing, video transcoding, and graphics processing.
Commercially available: PI7,PI2
Features of computing accelerated GPU cloud server
Specification Name | Graphics Card Specification | Quantity of Graphics Cards | Single-GPU Performance | Disk Type |
PI7 | Nvidia Tesla A10 | 1, 2, 4 | 31.2 TFLOPS single-precision floating point computing | Common I/O |
PI2 | Nvidia Tesla T4 | 1, 2, 4 | 8.1 TFLOPS single-precision floating point computing | Common I/O |
GPU image acceleration basic type
The GPU cloud server of GPU image acceleration basic type is based on the NVIDIA GRID virtual GPU technology. A public image is integrated with the GRID drive and contains the license for NVIDIA GRID vWS to effectively reduce the cost of use in small-scale demands. This kind of cloud server is suitable for image rendering and small-scale AI inference scenarios.
Commercially available: G7,
Features of GPU cloud server of image acceleration basic type
Specification Name | Graphics Card Specification | Quantity of Graphics Cards | Single-GPU Performance | Disk Type |
G7 | Nvidia Tesla A10 | 1/4, 1/2, 1 | 31.2 TFLOPS single-precision floating point computing | Common I/O |