Processor architecture Flashcards
What is the Von Neumann bottleneck
Fetch and data operations can not occur at the same time with a shared bus
What is the Harvard architecture?
Data and instructions are stored in separate memory units with Separate buses
Advantages of Harvard over Von Neumann
While data is being written to and read from memory the next instruction can be read from memory therefore making it much faster
What is parallel processing?
When a computer carriers out multiple computations simultaneously to solve a given problem
What are the two ways parallel processing can be achieved?
Single instruction multiple data - A single instruction is carried out on multiple bits of data simultaneously
Multiple instruction multiple data - Multiple instructions are carried out on multiple bits of data simultaneously
What are super computers?
Massive parallel machines
+/- of super computers
+ Can tackle huge problems
- energy consumption
- cost a LOOOOTTTTTTTTT
What is distributed processing?
Developed as an alternative to parallel processing, it splits up parts of a problem across multiple machines on a network
What process does HLL go through when being converted into machine code?
HLL is complied by complier into assembly language. Assembly language is then converted into machine code with the use of an assembler
What is the Von Neumann architecture?
Memory CPU input/output all use same bus which means instructions are executed linearly.
+/- of CISC processors (Use)
+ Larger more feature rich instruction set
+ takes up less memory
+ Smaller code requiring less RAM
+ Less work for complier
- Instructions take multiple cycles to complete
- Slower: They may only need one instruction to be processed however they are slower as they take multiple cycles to complete the instruction.
- fewer general purpose registers
+/- of RISC processors (Use)
\+ Simpler instruction set and simpler hardware \+ Faster clock speeds \+ More general purpose registers \+ consume less power \+ cheaper \+ Instructions carried out in one cycle - smaller instruction set - Limited instructions means the code has to be complied more - Longer code requiring more RAM
+/- of multicore systems
\+Faster with large data \+Makes max use of CPU \+Not limited by Von Neumann bottleneck -Not all data is suitable (all data must be independent from previous results) -Costly
What advantages does RISC have in terms of manufacturing?
- Simpler physical construction
- Less complicated circuitry due to performing fewer instruction types
- They are made smaller, use less power and generate less heat
