Among 8 subjects on Elective II for Bachelor in Computer Engineering and Bachelor in Electronics & Communication Engineering, Advanced Computer Architecture is focused in providing advanced knowledge of computer architecture including parallel architectures, instruction-level parallel architectures, superscalar architectures, thread and process-level parallel architecture. The course content of Advanced Computer Architecture includes 3 lecture, 1 tutorial, 1.5 practical and is assigned for Fourth Year – Second Part. The subject code is CT 765 04. Following syllabus, marking scheme and references are as per the latest update of IOE Syllabus.
- Computational models (6 hours)
- Computational model
- Von Neumann Computational model
- Evolution and interpretation of the concept of computer architecture
- Interpretation of the concept of the computer architectures at different levels of abstraction
- Multilevel hierarchical framework
- Parallel Processing (7 hours)
- Process, Thread, Processes and threads in languages
- Concurrent and parallel execution and programming languages
- Types of available parallelism
- Levels of available functional parallelism
- Utilization of functional parallelism
- Classification of parallel architectures
- Relationships between languages and parallel architectures
- Pipelined Processors (7 hours)
- Principle of pipelining
- Structure of pipelines
- Performance measures
- Application scenarios of pipelines
- Layout of a pipeline, Dependence resolution
- Design space
- Pipelined processing of loads and stores
- Superscalar Processors (8 hours)
- The emergence and widespread adaption of superscalar processors
- Specific tasks of superscalar processing
- Parallel decoding
- Superscalar instruction issue
- Scope of shelving
- Layout of shelving buffers
- Operand fetch policies
- Instruction dispatch schemes
- Scope of register renaming with example
- Processing of control transfer Instructions (7 hours)
- Types of branches, Performance measures of branch processing
- Branch handling
- Delayed branching
- Branch processing
- Multiday branching
- Thread and process-level parallel architectures (10 hours)
- MIMD architectures
- Distributed memory MIMD architectures
- Fine-gain and Medium-gain systems
- Coarse-grain multicomputer
- Cache coherence
- Uniform memory access(UMA) machines
- Cache-coherent non-uniform memory access(CC-NUMA) machines
- Cache only memory architecture(COMA)
References:
- Advanced Computer Architectures: a design space approach, Deszo Sima, Terence Fountain, Peter Kacsuk
- Computer Architecture and organization, John P. Hayes
- Computer Organization and Design, David A. Patterson, John L. Hennessy
Evaluation Scheme:
The questions will cover all the chapters of the syllabus. The evaluation scheme will be as indicated in the table below:
Chapters |
Hours |
Marks Distribution* |
1 |
6 |
10 |
2 |
7 |
13 |
3 |
7 |
13 |
4 |
8 |
14 |
5 |
7 |
13 |
6 |
10 |
17 |
Total |
45 |
80 |
*There could be a minor deviation in Marks distribution