CMOS technology improvement allows to increase the number of cores integrated on a single chip and makes Network-on-Chips (NoCs) a key component from the performance and reliability standpoints. Unfortunately, continuous scaling of CMOS technology poses severe concerns regarding failure mechanisms such as NBTI and stressmigration, that are crucial in achieving acceptable component lifetime. Process variation complicates the scenario, decreasing device lifetime and performance predictability during chip fabrication. This paper presents a novel sensor-less methodology to reduce the NBTI degradation in the on-chip network virtual channel buffers, considering process variation effects as well. Experimental validation is obtained using a cycle accurate simulator considering both real and synthetic traffic patterns. We compare our methodology to the best sensor-wise approach used as reference golden model. The proposed sensor-less strategy achieves results within 25% to the optimal sensor-wise methodology while this gap is reduced around 10% decreasing the number of virtual channels per input port. Moreover, our proposal can mitigate NBTI impact both in short and long run, since we recover both the most degraded VC (short run) as well as all the other VCs (long term).
A sensor-less NBTI mitigation methodology for NoC architectures
ZONI, DAVIDE;FORNACIARI, WILLIAM
2012-01-01
Abstract
CMOS technology improvement allows to increase the number of cores integrated on a single chip and makes Network-on-Chips (NoCs) a key component from the performance and reliability standpoints. Unfortunately, continuous scaling of CMOS technology poses severe concerns regarding failure mechanisms such as NBTI and stressmigration, that are crucial in achieving acceptable component lifetime. Process variation complicates the scenario, decreasing device lifetime and performance predictability during chip fabrication. This paper presents a novel sensor-less methodology to reduce the NBTI degradation in the on-chip network virtual channel buffers, considering process variation effects as well. Experimental validation is obtained using a cycle accurate simulator considering both real and synthetic traffic patterns. We compare our methodology to the best sensor-wise approach used as reference golden model. The proposed sensor-less strategy achieves results within 25% to the optimal sensor-wise methodology while this gap is reduced around 10% decreasing the number of virtual channels per input port. Moreover, our proposal can mitigate NBTI impact both in short and long run, since we recover both the most degraded VC (short run) as well as all the other VCs (long term).File | Dimensione | Formato | |
---|---|---|---|
socc2012.pdf
accesso aperto
Descrizione: main document
:
Publisher’s version
Dimensione
187.15 kB
Formato
Adobe PDF
|
187.15 kB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.