Experimental study of aggressive undervolting in FPGAs
Autor: | Salami, Behzad, Unsal, Osman Sabri, Cristal Kestelman, Adrián|||0000-0003-1277-9296 |
---|---|
Jazyk: | angličtina |
Rok vydání: | 2019 |
Předmět: |
Low voltage integrated circuits -- Reliability
Energy Reliability FPGAs Energia elèctrica -- Consum Voltage underscaling FPGAs Enginyeria electrònica::Circuits electrònics [Àrees temàtiques de la UPC] Hardware_PERFORMANCEANDRELIABILITY Circuits integrats de baix consum -- Fiabilitat Electric power consumption Circuits integrats de baix consum Hardware_INTEGRATEDCIRCUITS Informàtica::Impacte ambiental [Àrees temàtiques de la UPC] High performance computing Informàtica::Arquitectura de computadors [Àrees temàtiques de la UPC] Càlcul intensiu (Informàtica) Low voltage integrated circuits |
Zdroj: | UPCommons. Portal del coneixement obert de la UPC Universitat Politècnica de Catalunya (UPC) |
Popis: | In this work, we evaluate aggressive undervolting, i.e., voltage scaling below the nominal level to reduce the energy consumption of Field Programmable Gate Arrays (FPGAs). Usually, voltage guardbands are added by chip vendors to ensure the worst-case process and environmental scenarios. Through experimenting on several FPGA architectures, we measure this voltage guardband to be on average 39% of the nominal level, which in turn, delivers more than an order of magnitude power savings. However, further undervolting below the voltage guardband may cause reliability issues as the result of the circuit delay increase, i.e., start to appear faults. We extensively characterize the behavior of these faults in terms of the rate, location, type, as well as sensitivity to environmental temperature, with a concentration of on-chip memories, or Block RAMs (BRAMs). Finally, we evaluate a typical FPGA-based Neural Network (NN) accelerator under low-voltage BRAM operations. In consequence, the substantial NN energy savings come with the cost of NN accuracy loss. To attain power savings without NN accuracy loss, we propose a novel technique that relies on the deterministic behavior of undervolting faults and can limit the accuracy loss to 0.1% without any timing-slack overhead |
Databáze: | OpenAIRE |
Externí odkaz: |