0
$\begingroup$

Often, similar ideas or works have vastly different namings or covers. This is a hassle for searching in distant fields, especially in a non-native language.

Idea. We mathematically model (Petri nets, graphs, automatons, any formal way) architectures and algorithms. One of the models can be substituted for various behaviour data or even its implementation or mix of the model with a real data in general case. Then we simulate their interactions on paper or on an instrumental computer and obtain some useful predictions with aim on parallel computations without actually write the code and measuring its characteristics (full finished code at least, this isn't a simple profiling of a real implementation). For example, this way could trying to predict scaling, algorithm and architecture matching quality, performance, comparative measures of variants, etc.

Could you, please, name not obvious keywords or other leads (authors, titles, surveys are the most useful) to the articles containing similar ideas and their implementations?

"Obvious" keywords are "model, architecture, interaction, parallel computations, imitation, simulation" and their combinations. Synonyms could be missed too, if they aren't found by goggle with "~."

$\endgroup$
  • 1
    $\begingroup$ This seems quite broad to me, without much of a restriction on what kind of models and properties you want to look at. Community votes, please: too broad? $\endgroup$ – Raphael Jul 13 '16 at 10:59
  • $\begingroup$ I can narrow it as you said, but even in this form, there aren't plenty of actual (this is a modern practical IT field, so 2012+) works in English. In addition, I can narrow it by enlisting what I'd found. Subjective remark about the "broadness" - no fast answers where there should be plenty of them because of the "broadness." $\endgroup$ – Les Jul 13 '16 at 11:33
  • $\begingroup$ This seems super-broad to me. There are probably thousands of papers on modelling the behavior of algorithms. It seems the only filter you have on that is that you are interested in parallel algorithms. (The "etc." implies you aren't limiting this by what kinds of predictions the model is used for, which leaves this especially broad.) This could fit a large fraction of the work in model checking, formal verification, theoretical analysis of (parallel) algorithms, (parallel) algorithm design, compiler work on loop vectorization, program analysis of parallel algorithms, and more. $\endgroup$ – D.W. Jul 13 '16 at 17:31
  • $\begingroup$ Please, read the whole idea paragraph. It's a consequent thing, not a set of randomly ordered sentences. If they're thousands of papers, just toss some to me. Why write the posts? Prove your claims in the same time if you possess a superior understanding? $\endgroup$ – Les Jul 14 '16 at 9:06
  • $\begingroup$ One example going next. Other points are corrupted as well. "Compiler work on loop vectorization" - (sarcasm) compiler builds a general algorithm model, simulates interactions with a target device and uses it results in loops unrolling and vectorization. This is how it works in most of them. (sarcasm off) How it could be even in proximity?! You'd written many wide sub fields where the idea could be. I asked - where is it? Here's a vaaast difference in a search problem. $\endgroup$ – Les Jul 14 '16 at 9:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.