Question

1. How can systems utilize the principle of “locality” to increase performance?

2. What is a “multi-banked” cache, and how can it be used to optimize the use of a cache to reduce memory access times?

3. What factors are responsible for the resurgence of the popularity of virtual machines?

4. What is “loop unrolling,” and how does it exploit parallelism?

Answer 1

1. Principle of locality is also called as locality of reference. It is a situation where the OS fetches the closest instruction and also finds the next closest one in handy when it is needed. Thus fetching out the mostly used and next coming instruction helps in faster execution without any delay which results in saving time and it eventually increases the performance.
2. Multi-bank cache divides the data into many different memory units. That means we can work or read the instructions in parallel. Since the parallel division and execution takes place this reduces the time for execution of the instruction. By spreading the memory units and executing them in parallel helps in no time wait and eventually it gives quick response time.
3. virtual machines are useful in running multiple operating systems simultaneously.

• During the past few years the usage of the virtual machines had grown more and gained popularity.
• The following were some of the factors that were responsible for the usage and popularity of virtual machines.
• Virtual machines allow us to run multiple OS on the same system without effecting the actual hard ware or physical behavior of the system.
• The maintenance cost is also low when we use virtual machine because it doesn't effect the physical hardware of the system.
• Continuous up time availability is observed in case of virtual machines.
• The server failures are also been rectified where we can switch to some other server if the actual one doesn't work.

4. Loop unrolling is a loop transformation where we can write the code present inside a loop for many times.

Loop unrolling helps in increasing the functionality of the hardware units which eventually exploits the parallelism.

In case of parallelism the multiple instructions execute at the same time where when loop unrolling happens it takes time to run by making the other system slow this eventually halts and interrupts the parallelism.