MGMT 1050 Chapter Notes - Chapter 9: Simultaneous Multithreading, Very Long Instruction Word, Multiprocessing
![](https://new-preview-html.oneclass.com/Gy340B26qOkoN9bLwAvGQAdXzp1xr5w7/bg1.png)
MGMT 1050 Chapter 9 Notes – Summary
Introduction
• The first section investigates different CPU architectures, with particular focus on the
modern manifestation and organization of traditional architectures
• The section also briefly considers two interesting recent architectures, the Trans meta
VLIW and Intel EPIC architectures.
• We consider various CPU features and enhancements, with an emphasis on alternatives
to the traditional control unit/ALU CPU organization.
• We explain how these alternative organizations address major bottlenecks that limit
CPU execution speed, with a number of innovative techniques for improving CPU
performance.
• Looks at memory enhancements.
• The most significant improvement in memory access speed is cache memory. Cache
memory is discussed in considerable depth.
• We present a general model that includes the features, enhancements, and techniques
described.
• This model represents the organization of most current CPUs.
• Considers the concept of multiprocessing: a computer organization consisting of
multiple CPUs directly connected together, sharing memory, major buses, and I/O.
• This organization adds both performance enhancement and additional design
challenges.
• We also briefly introduce a complementary feature, simultaneous multithreading.
• Two types of multiprocessors are presented.
• The symmetrical multiprocessor is more common.
• It is well-suited for general purpose computing.
• An alternative, the master-slave multiprocessor is useful for computer applications
characterized by computationally intense, repetitive operations, such as graphics
processing.
find more resources at oneclass.com
find more resources at oneclass.com