L3 - Types of processor

characteristics of von neumann architecture

characteristics of von neumann architecture

• shared memory for instructions and data in the same format (word length)

• single CU follows linear FDE cycle, one instruction at a time

• registers offer fast data access

• one bus is simpler for control unit design but forms a bottleneck

• used in PCs, laptops, servers and high performance computers

characteristics of harvard architecture

• instructions and data stored in separate memory units with separate buses with separate word lengths

• CU for two buses is more complicated and expensive

• parallel access to data and instructions

• used in digital signal processing and embedded systems

• reading and writing is done at the same time as fetching instructions

characteristics of contemporary architectures

• SIMD (single instruction multiple data)

  • parallel processing, processor carries out single instruction on multiple data like GPU stuff

• MIMD (multiple instruction multiple data)

  • multiple instructions carried out on multiple data across several cores

• distributed computing

  • Multiple computers on a shared network each take on part of a bigger problem, eg SETI@home in NASA 

   

characteristics of CISC (complex instruction set compiler)

• More complex hardware with many transistors 

• Shorter code requiring less RAM 

• Closer to high level languages 

• Less work for compiler to do 

• Some instructions require multiple clock cycles to complete 

• Fewer registers 

• Lots of transistors which use up power and generate heat 

• Larger in size than RISC CPUs due to heat sink/fan 

 

characteristics of risc (reduced instruction set compiler)

• Simpler hardware 

• Longer code requiring more RAM 

• Further away from high level languages 

• Compiler has to do more work 

• Every operation takes place in one clock cycle (allows for pipelining; more efficiency) 

• More registers to hold many smaller, simpler instructions 

• Less transistors so less power consumption, less heat and lower cost 

• Smaller in size than CISC CPUs 

differences between CPU and GPU

• GPU is more specialised towards parallel computing as opposed to general processing functions

• GPU runs processes in parallel, but CPU runs them serially

• GPU has more cores but they’re each less powerful than CPU cores

• GPU has high throughput but CPU has low latency

• GPU is used in high performance computing devices

what are co-processors

• they supplement the functions of the primary processor

• they carry out a limited range of functions like floating point arithmetic, graphics processing, digital signal processing

• it may not be a general purpose processor with the ability to fetch its own instructions, input/output etc