NVIDIA Grace Hopper Superchip
GPTKB entity
Statements (30)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
gptkb:microprocessor
|
| gptkbp:announced |
2022
|
| gptkbp:architecture |
gptkb:ARM_(CPU)
gptkb:Hopper_(GPU) |
| gptkbp:connects |
gptkb:NVLink-C2C
|
| gptkbp:contains |
gptkb:NVIDIA_Hopper_GPU
gptkb:NVIDIA_Grace_CPU |
| gptkbp:CPU_core_count |
up to 72
|
| gptkbp:CPU_memory_bandwidth |
up to 1 TB/s
|
| gptkbp:deployment |
2023
|
| gptkbp:formFactor |
gptkb:model
|
| gptkbp:GPU_memory_bandwidth |
up to 3 TB/s
|
| https://www.w3.org/2000/01/rdf-schema#label |
NVIDIA Grace Hopper Superchip
|
| gptkbp:manufacturer |
gptkb:NVIDIA
|
| gptkbp:memoryType |
gptkb:HBM3
gptkb:LPDDR5x |
| gptkbp:namedAfter |
gptkb:Grace_Hopper
|
| gptkbp:supports |
gptkb:NVIDIA_AI_Enterprise
gptkb:NVIDIA_NVLink_Switch_System gptkb:CCIX gptkb:CXL gptkb:PCIe_Gen5 gptkb:NVIDIA_CUDA |
| gptkbp:targetMarket |
high-performance computing
AI workloads |
| gptkbp:used_in |
gptkb:NVIDIA_DGX_GH200_system
|
| gptkbp:uses |
data centers
supercomputers |
| gptkbp:bfsParent |
gptkb:RISC-V_instruction_set_architecture
|
| gptkbp:bfsLayer |
5
|