1.1.2 Flashcards
Types of processor
What is CISC?
This is a type of architecture that aims to
complete a task in as few lines of
assembly language as possible by having
specific assembly language instructions for
a variety of tasks.
A single assembly language instruction such as MULT A, B could be used to multiply A by B and store the result back in A
What is RISC?
Reduced Instruction Set Computers (RISC) take an opposite approach
A minimum number of very simple instructions, each taking one clock cycle, are used to accomplish all the required operations in multiple general purpose registers
Benefits of CISC
-Compatible with decades of software.
-Because the length of the assembly code is short compared to a RISC program, much less RAM is required (when RAM was more expensive this was a significant consideration).
-More powerful instructions make the assembly language (simpler) programmer’s job much easier but this means the processor architecture is more complex
-Many addressing modes
Downsides of CISC
-Power hungry - physically larger and much more complicated chips.
-More expensive to design and manufacture
-Get hotter so need heatsink/fan
-Only about 20% of the instructions are used most of the time
Benefits of RISC
-Because all of the instructions execute in a uniform amount of time (i.e. one clock cycle), pipelining is possible.
-Processor architecture is simpler
-Less complex circuitry is needed as instructions are basic making it easier to design and manufacture.
-Costs less to design and manufacture.
-Chip can be smaller and therefore less power hungry (good for mobiles/batteries).
-Fewer addressing modes.
Downsides of RISC
-Only software developed relatively recently runs on modern ARM processors.
-Some more complex tasks will be harder to code and programs written with this instruction set will likely be longer than with CISC (thus needing more RAM).
-RISC CPUs run at a lower clock speed
GPU
Graphics Processing Unit is a processor chip that performs rapid mathematical calculations, primarily for the purpose of rendering images.
GPUs are able to perform an instruction on multiple pieces of data at one time which is often what we want to do when processing graphics e.g. shading pixels so they can perform graphics operations faster than a CPU which is a general purpose processor
GPUs consist of a massively parallel architecture with thousands of small cores
handling each of the instructions simultaneously. They are best suited for
manipulating computer graphics and also where an algorithm involves
processing large blocks of data in compute-intensive calculations using vectors
and matrices.
What are GPUs used for?
-Video Editing and Content Creation.
-Image recognition used in self-driving cars, military intelligence.
-Physics simulation eg machines, fluid dynamics
-Engineering simulations eg air flow
-Data mining
-Bitcoin mining
-Distributed relational database queries using distributed set processing.
-Brute-force searches in cryptography.
-Computer simulations comparing many independent scenarios, such as climate models.
-Event simulation and reconstruction in particle physics.
-Supercomputers
What is parallel processing?
Parallel processing (progress is being made on more than one task simultaneously) can be achieved using GPUs (graphics processing units).
In a parallel processor system, several CPUs work together on a task.
There are several types of parallel architectures possible, all depending
on how the data and instructions are processed. An example is where
different cores execute on different threads (Multiple Instructions), which
are operating on different parts of memory (Multiple Data).
Benefits of parallel processing
-Huge performance increase in processing graphics or other embarrassingly parallel problems eg graphics rendering.
-Problems are often easily done in parallel eg graphics, data mining, weather modelling, traffic modelling etc.
-Multi-core could mean that if a single core stalls or is stuck waiting others can absorb the load.
-Faster processing meaning more tasks can be completed in a time frame. In graphics for example this can mean higher resolution and higher frame rate.
-Potential cost savings are computations are done quicker.
-Faster response for time critical systems like autopilot systems for planes
Limitations of parallel processing
-Some problems can not be done in parallel. Eg binary search
-There is an overhead in coordinating threads between CPUs or cores. Potential slow down if multithreading as a single core or processor is completing many threads rather than just one
-Some cores/processors work while others wait due to insufficient parallelism or unequal size tasks.
-Many programs are hard-coded to use a certain number of cores. This is done for a variety of reasons ranging from the nature of what the program is doing making it non-conducive to using multiple CPU cores to it simply being easier to program for a fixed number of cores.
-More costly as more sophisticated or additional hardware required compared to sequential processing.
-Deadlock can occur
-Race conditions leading to errors
What is a multicore system?
In a multicore system, the CPU has two or more independently processing units.
Multicore systems can cause problems for CPU designers because of the possibility of excessive overheating of each core.
