Statements (51)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
gptkb:technology
|
| gptkbp:developedBy |
gptkb:AMD
|
| gptkbp:enables |
energy efficiency
high-speed data transfer data integrity cache coherency clock synchronization power management multi-GPU communication low latency communication communication between CPU and GPU communication between CPU cores communication between chiplets system scalability dynamic frequency scaling I/O die communication I/O scalability chip-to-chip communication chiplet interconnect core-to-core communication die-to-die communication dynamic voltage scaling heterogeneous integration high bandwidth communication memory controller communication memory scalability multi-socket communication |
| gptkbp:hasVersion |
gptkb:Infinity_Fabric_2.0
gptkb:Infinity_Fabric_3.0 |
| https://www.w3.org/2000/01/rdf-schema#label |
AMD Infinity Fabric
|
| gptkbp:introducedIn |
2017
|
| gptkbp:memoryBusWidth |
up to hundreds of GB/s
|
| gptkbp:replacedBy |
HyperTransport (in AMD CPUs)
|
| gptkbp:supports |
modular design
scalability |
| gptkbp:usedIn |
gptkb:AMD_EPYC_processors
gptkb:HPC_(High_Performance_Computing) gptkb:Zen_architecture gptkb:AMD_Ryzen_processors gptkb:AMD_Radeon_graphics_cards gptkb:Zen_2_architecture gptkb:Zen_3_architecture gptkb:Zen_4_architecture data center solutions heterogeneous computing server processors desktop processors laptop processors multi-die architectures |
| gptkbp:bfsParent |
gptkb:Nvidia_NVLink
|
| gptkbp:bfsLayer |
6
|