NVIDIA None Interview Question
NonesTeam: power
Country: United States
Interview Type: In-Person
Single core:
pros: +less complexity of HW design
+less complexity of SW design (here inside a lot of factors)
cons: -potentially higher power consumption
-in some cases higher context switching overheads
Dual core:
pros: +Allows more flexible system solutions
+In some cases can seriously increase perofrmance with high latency memory: twice less cycles to wait for memory. I had projects in which it played decisive role. It also allows to decrease memory frequency and save some power with same performance
cons: -HW design higher complexity
-requires mechanisms of resource sharing
1- Single 1Gig processor
- massy8448 March 15, 2014Pros: High CPU frequency means shorter execution cycle, faster interrupt handler (better real-timeness, higher internal timer resolution) and less power consumption compared to a dual core due to the fact that there are less transistors on he chip to use power which results in higher battery life. Usually CPU clock rate also can be reduced by software to get further reduction in power consumption when Mobile is idle.
The silicon chip also would be smaller in size and would weight lighter due to reduction in number of populated transistors which makes it more suitable for small sized mobile phones and more portable.
cons: High frequency means higher radio interference in PCB level so it make PCB design more complicated to ensure level of the inducted noise remains in an acceptable level.
If the application is heavily multi-tasked then overhead of the CPU context switching can be also a problem by means of reduction in overall CPU performance.
1- Dual core 500Mhz processor
Pros: lower frequency, lower noise generation, lower chance for interfere with other devices in critical applications such as Airplane , simpler PCB design which means reduced cost of manufacturing, providing the real parallel processing which increases processing power in heavily muti-threaded applications. In critical applications having two CPUs may alos be considered as 1oo2 CPU redundancy so as to make sure if one CPU failed there might be still another CPU to take control and keep software running.
Cons:
Bigger silicon, heavier, more power consumption, slower interrupt handler ( increased interrupt response time)