Question

Your answer

1. Explain why overhead is less when creating a new thread than the overhead to create a process.

   

Your answer

2. Discuss the types of parallelism and give examples of cases where each would be appropriate.

Your answer

3. Provide an example where a multicore-processor-based system could have neither concurrency nor parallelism.

4. Using Amdahl's Law, calculate the speedup gain for the following situation:

    75 percent parallel with (a) eight processing cores and (b) sixteen processing cores

    57 percent parallel with (a) two processing cores and (b) four processing cores

Your answer

    89 percent parallel with (a) four processing cores and (b) eight processing cores

Answer 1

Ans(2) :

• Bit-level parallelism: It is the frame of parallel computing which is based on the expanding processor’s measure. It decreases the number of enlightening that the framework must execute in arrange to perform a errand on large-sized data. Example: Consider a situation where an 8-bit processor must compute the whole of two 16-bit integrability. It must to begin with whole up the 8 lower-order bits, at that point include the 8 higher-order bits, hence requiring two enlightening to perform the operation. A 16-bit processor can perform the operation with fair one instruction.
• Data-level parallelism : Informational from a single stream work concurrently on a few data .Restricted by non-regular information control designs and by memory bandwidth.
• Instruction-level parallelism: A processor can as it were address less than one instruction for each clock cycle phase. Multiple informational from the same instruction stream can be executed concurrently These enlightening can be re-ordered and gathered which are afterward on executed concurrently without influencing the result of the program. Usually called instruction-level parallelism.
• Task Parallelism: Errand parallelism utilizes the deterioration of a errand into subtasks and after that distributing each of the subtasks for execution. The processors perform execution of sub assignments concurrently.

=======================================================================================

Ans(4) :

=======================================================================================

Ans(1) :

• Less overhead to set up and end vs a process: since exceptionally small memory replicating is required (fair the string stack), strings are speedier to begin than processes. To begin a prepare, the entire handle range must be copied for the modern prepare duplicate to begin. While some operating frameworks as it were duplicate memory once it is adjusted (copy-on-write), this can be not all around ensured.
• A multithreading may be a specialized frame of multitasking. Multitasking strings require less overhead than multitasking forms. I ought to characterize another term related to strings handle: A handle comprises of the memory space apportioned by the working framework that can contain one or more strings. A string cannot exist on its claim; it must be a portion of a prepare. A handle remains running until all of the non-daemon strings are done executing. Multithreading empowers you to type in exceptionally productive programs that make most extreme utilize of the CPU, since sit out of gear time can be kept to a least.

=======================================================================================