preview

Software Optimization Methods Of Changing A Software System

Decent Essays
Open Document

HW 4
Software Optimization Techniques

Software optimization is process of changing a software system to enable some aspect of the process to work more efficiently using less memory storage and less power.

Profiling and timing code execution:
We need to identify portions of code that run frequently which are called hotspots and make these identified hotspots run faster.
In Profiling the first step is to understand the code in terms of its computations and requests.
Next we need to identify any bottlenecks that may disrupt the performance.
Third we need to set objectives.
Finally we need to improve the performance cycles.

Using platform specific features:
We need to keep in mind the cache coherency, pipelining and branch guessing and side …show more content…

More the requirements implies slower network solution.
In order to do this we need to consolidate the CSS and the JS files, Embedding CSS images into CSS which reduces extra space, load images only as they scroll into our view.
Next we need to reduce the bytes which also leads to slower network solutions.
For this we need to enable zip compressions, compress images fully and resizing the images as per the screen size.

Slower javascript leads to weaker CPU:
For this we need to make sure that javaScript is async to the page load.
Asynch avoids delay.
We need to also remove the unused codes.

Numerical methods:
Linear programming:
The objective function is a real-valued function which is defined on a polyhedron. Main function is finding a point in the polyhedron that has the smallest or the largest values if and when a point like that exists.

Quadratic programming:
QP is minimizing/maximizing a quadratic function which consists of variables that are subject to LC on these very variables.

Stochastic programming:
This is mainly used for modeling problems that have uncertainity. Almost all of the Real world problems include unknown parameters(atlleast few).

Advanced optimization:
Hill climbing:
Hill climbing is used for reaching the end state from the first (start) node. Decision of the next nodes from the starting nodes, can be done by various algorithms.
There are two variants of hill climbing which are simple hill climbing

Get Access
Related
Get Access