Key Architecture Highlights
- FP32 Compute performance: 31.2 teraFLOPS
- Tensor Compute performance: TF32 Tensor Core: common:62.5 teraFLOPS; sparse:125 teraFLOPS BFLOAT16 Tensor Core: common:125 teraFLOPS; sparse:250 teraFLOPS FP16 Tensor Core: common:125 teraFLOPS; sparse:250 teraFLOPS
- Integer Compute performance: INT8 Tensor Core: Normal: 250 TOPS; Sparse: 500 TOPS; INT4 Tensor Core: Normal: 500 TOPS; Sparse: 1, 000 TOPS
Applications
- AI training and inference
- High-performance computing
- Data center
Engineering Parameters
| FP32 Compute performance | 31.2 teraFLOPS |
|---|---|
| Tensor Compute performance | TF32 Tensor Core: common:62.5 teraFLOPS; sparse:125 teraFLOPS BFLOAT16 Tensor Core: common:125 teraFLOPS; sparse:250 teraFLOPS FP16 Tensor Core: common:125 teraFLOPS; sparse:250 teraFLOPS |
| Integer Compute performance | INT8 Tensor Core: Normal: 250 TOPS; Sparse: 500 TOPS INT4 Tensor Core: Normal: 500 TOPS; Sparse: 1, 000 TOPS |
| RT Cores | 72 RT Cores |
| Video decoding | 1 encoder 2 decoder (+AV1 decode) |
| GPU memory ／ Memory | 24GB GDDR6 |
| GPU memory bandwidth | 600GB/s |
| Interconnect Interface | PCIe Gen4 64GB/s |
| Form factor and installation | Single-slot, full-height, full-length (FHFL) |
| Power consumption | 150W |
Warranty
Public manufacturer information does not specify the warranty term applicable to our supply channel. The supply region, channel and actual warranty service are confirmed in the supply contract.
Manufacturer information
Manufacturer informationVerified Documentation & Datasheets
Manufacturer product information
Download Document