In this post I will test several compiler options and switches that could make the particle system run faster.
Read more to see how I've reached around 20% of performance improvement!
I am very curious about the D language and its community. Although, I do not have lot's of experience with this language, I try to track news and important updates. Recently, I've noticed that there is another book released regarding the language: D Cookbook, by Adam D. Ruppe
Let's see what's inside this book
It's time to start improving the particle code and push more pixels to the screen! So far, the system is capable to animate and do some basic rendering with OpenGL. I've shown you even some nice pictures and movies... but how many particles can it hold? What is the performance? Is it that bad? What are the hot spots and bottlenecks? We need to figure that out in the first place!
Recently, I've written an article about using a .NET third party library to generate reports from apps. You can find it on this in my previous post. In my opinion, the whole idea might be useful, for instance, for performance tests. I often try to make such in my blog. Basically you do some tests and then output results to the console or a txt file...
But, wouldn't it be great to write results directly to a spreadsheet and draw some charts automatically? Can it be done in C++ as well?
On my programming blog, I often try to compare performance characteristics of different algorithms or concepts. I usually log performance output (like elapsed time) to the console or a txt file and then copy this to a spreadsheet and analyse. But recently, I've found another way of doing this: I've used Spire.XSL library to generate the final spreadsheet file - with all the tables and charts! Read further to learn how one can leverage this library for various automation tasks.
In the previous particle post the particle generation system was introduced. But after a new particle is created we need to have a way to update its parameters. This time we will take a look at updaters - those are the classes that, actually, makes things moving and living.
Recently ended DConf 2014 conference was, as usually, a great event filled with interesting topics about the D language. I still need to update my little knowledge about the language and see more presentations, but one keynote especially drew my attention. This was a talk from Scott Meyers called The Last Thing D Needs.
But... hmmm... why C++ guy on a D conference? And why he had slides without D code... only C++? Strange...
In one of my previous post I wrote about performance differences between using
vector<shared_ptr<Obj>>. Somehow I knew that the post is a bit unfinished and need some more investigation.
This time I will try to explain memory access patterns used in my code and why performance is lost in some parts.
Please see my notes from the second day at Digital Dragons 2014 Conference from Cracow.
Presentations started a bit earlier than the day before - at 9:00. Although a lot of people had been at the official DD afterparty and Gala awards, at the morning sessions there were some decent number of survivors :)