SwapLess reduces mean inference latency by up to 63.8% single-tenant and 77.4% multi-tenant on Edge TPUs by online adjustment of model partition points and CPU allocation using a queueing model that accounts for swapping costs.
Enabling Real-time AI Inference on Mobile Devices via GPU-CPU Collaborative Execution
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.DC 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Collaborative Processing for Multi-Tenant Inference on Memory-Constrained Edge TPUs
SwapLess reduces mean inference latency by up to 63.8% single-tenant and 77.4% multi-tenant on Edge TPUs by online adjustment of model partition points and CPU allocation using a queueing model that accounts for swapping costs.