Statements (20)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
GPU technology
|
| gptkbp:allows |
up to 32 simultaneous hardware work queues
|
| gptkbp:availableOn |
Tesla K20 GPU
|
| gptkbp:benefit |
high performance computing (HPC) applications
|
| gptkbp:developedBy |
gptkb:NVIDIA
|
| gptkbp:enables |
greater concurrency in GPU workloads
multiple CPU cores to launch work on a single GPU simultaneously |
| gptkbp:features |
gptkb:NVIDIA_Quadro_GPUs
gptkb:NVIDIA_Tesla_GPUs NVIDIA GeForce GPUs (Kepler and later) |
| https://www.w3.org/2000/01/rdf-schema#label |
Hyper-Q
|
| gptkbp:improves |
GPU utilization
|
| gptkbp:introducedIn |
gptkb:NVIDIA_Kepler_architecture
|
| gptkbp:reduces |
CPU-GPU bottlenecks
|
| gptkbp:replacedBy |
single hardware work queue in previous architectures
|
| gptkbp:supports |
gptkb:CUDA
|
| gptkbp:bfsParent |
gptkb:Nvidia_Kepler_architecture
gptkb:Kepler_architecture gptkb:Kepler_GPU_architecture |
| gptkbp:bfsLayer |
7
|