CS Module 1 Lesson 2

What is the main problem that needs to be overcome to make the response time of memory compatible with the speed of the processor?

• Memory consistency

• Synchronization issues

• Cache coherency

• Memory cycle time

In a shared memory system, what is the main feature regarding memory access?

• Each processor has its own private memory

• Data is stored in cache memories only

• All processors have equal access to data and instructions

• Processors communicate through message-passing protocol

What is the main disadvantage of a distributed memory system?

• Limited number of processors

• Communication between processors is more difficult

• Complexity in managing synchronization

• Difficulty in implementing cache coherency

What is the main characteristic of Uniform Memory Access (UMA) systems?

• Memory is divided into high-speed access areas

• Memory access time is constant for each processor

• Memory is physically distributed among processors

• Communication between processors is through message-passing protocol

What distinguishes Cache-Only Memory Architecture (COMA) systems from other architectures?

• Data is stored as duplicates in main memory

• Processors communicate through message-passing protocol

• Only cached memories are used

• Memory is physically distributed among processors

What is the main advantage of a cluster architecture?

• No conflicts at the communication bus level

• Fast communication between processors

• High scalability in terms of number of processors

• Ease of implementing cache coherency

What is the purpose of a Fail-over cluster?

• Ensure high performance computing

• Balance load between nodes

• Provide extremely high performance

• Monitor node activity and ensure continuous service

What is the main characteristic of Heterogeneous architectures?

• CPU assigns tasks to GPU in a parallel manner

• CPU and GPU work together with CPU controlling the computation

• GPU is considered as an autonomous processing unit

• CPU and GPU work independently

What is the communication method between CPU and GPU in Heterogeneous architectures?

• High-speed bus and shared memory

• Message-passing protocol only

• Shared physical memory only

• High-speed bus only

What is the main advantage of Massively Parallel Processing (MPP) machines?

• High performance with hundreds of processors

• Limited number of processors

• Single computing unit in the cluster

• Low scalability