Computer Architecture describes the set of rules and systems that all work together to create a computer system. Parallel processing or parallel computing refers to the action of speeding up a computational task by dividing it into smaller jobs across multiple processors. Some applications for parallel processing include computational astrophysics, geoprocessing, financial risk management, video color correction and medical imaging.
source: 12 Parallel Processing Examples to Know | Built In
Although I do not see a question, let me offer some observations.
"Computer Architecture describes the set of rules and systems that all work together to create a computer system." is a true statement. Some additional discussion is here: https://en.wikipedia.org/wiki/Computer_architecture.
"Parallel processing or parallel computing refers to the action of speeding up a computational task by dividing it into smaller jobs across multiple processors." is one use of parallel processing, whether it be in a shared or non-shared memory context.
An expanded use of parallel processing comes from considering multiple computers. These can be geographically co-located or distributed. Independent processes run on those computers. Those processes can be tasked or queried as needed.
Lets continue this discussion. It is a very interesting topic.
I would add that there is the instruction level parallelism that is generated by the compiler with little intervention of programmers. In this case there are not multiple processors, but multiple functional units (ALUs) in the same processor.
Best regards.
Eduard Babulak
To me it also is not clear what the exact question is.
Nevertheless, from my perspective I would like to respond to "what is missing?", with: "Programming languages embracing concurrency as a first class citizen are underrepresented."
Most programming languages nowadays try to add concurrency/parallelism as an afterthought or addition to existing frameworks.
However, one counterexample would be occam as programming language for the Transputer architecture, although not so modern. ;-)
Not exactly sure of the angle of the question, however, the question appears to read as though a parallel computer system is self contained with the parallelism built inside the box. These days, modern computer systems are not single boxes, but often clouds of processes many of which work in parallel. Looking at very large supercomputers that have hundreds of thousands processors, the interprocess, communication and coordination is absolutely vital to their success.
In answer to what's missing I would say "a communication network" for interprocess communication and coordination.
Clive England
I fully agree with your remarks!
Personally, I consider "Distributed Systems" as a superset of "Parallel Systems".
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Programming Languages
- Python