The
interrupt latency and context switching represent characteristics typical to the nMPRA processor, these coefficients having acceptable values for a deterministic architecture [15].
4 shows that
interrupt latency takes a bit more time because of the hardware response and the ISR execution that happens before the context switch to the new task that will serve the external request.
It offers advanced features that are optimised for MCU and real-time embedded applications, including reduced
interrupt latency, flash acceleration, debug features including iFlowTrace and support for AHB Lite as the interconnect interface.
Actually, there are four elements that contribute to the overall delay between an interrupt and the time the associated application program starts to run:
interrupt latency, interrupt handler duration, scheduler latency, and scheduling duration.
Events that incur more than six cycles of latency can mask the
interrupt latency. For example, instruction cache misses usually take long enough that the interrupt is delivered to the processor before the instruction that incurred the IMISS has issued.
The latency figures include
interrupt latency, that is the receiver does not poll the network device.
We have also considered the
interrupt latency which is 12 clock cycles for the fast interrupt (FIQ) and 25 clock cycles for ordinary interrupt (IRQ) [33].