Wavecrafters

Our Services

GPU Computing

We help bring the power of supercomputing to your business by helping to parallelize your code. In our philosophy, its important that you maintain your autonomy.
Thats why your team will always have the knowledge of how the code works

Visualization

Understanding often comes only from correctly visualizing your data. We provide the tools and assistance to help render your data in a comprehensive way. We have experts in charts, graph, 2D, 3D and Virtual Reality visualization

Machine Learning

We are experts in Deep Neural Networks, Genetic Algorithms, Support Vector Machines and other techniques that, when combined with the tremendous computation capabilities of the GPUs, can be used to learn from your data, revealing vital information that would otherwise remain hidden

Big Data

We can help bring the power of the Big Data to your organization. Using the power of the GPUs and our Machine Learning knowledge, we can harvest, analyze and store huge amounts of data that would otherwise be untreatable

Why Choose Us

Co-development

It is important for us that you maintain control of your code and application. Thats why we work together with your team, usually explaining and programming with them every line of code.

Closed Price

As a company we understand the importance of a controlled budget. Thats why we always agree on a closed price at the beginning of the project. If it takes more hours, we wont bill you for them. That way you always remain in control.

Platform Independent

We always adapt to your technology and your existing framework. We have experience working under both Linux and Windows environments. We have worked with all sorts of languages, such us CUDA, Java, C, C++, OpenGL, .NET, Javascript, NodeJS, Octave / Matlab, etc.

Customized training

We always taylor the courses and adapt the material to your specific needs. You will never receive a general course that will be of little use for your organization.

Our Blog

Exclusive prefix sum scan

Today Im going to try to explain a really useful algorithm for any parallel developer, but that took me a while to understand and to program myself. In this example we’ll be doing the exclusive version of the prefix sum scan, as it is a little bit more difficult than the inclusive version. It is possible to do this by computing the inclusive scan and then substracting the original vector from the result, but in this post we are going to do it a little bit different.

KNN normalizing the data

In a previous post we saw the knn classifier but said little about the data itsef, just that we use hash with two keys: {features: [f1, f2, ..., fn], label: lbl}{:lang="ruby"} The new point can use the same format, {features:[f1, f2, ..., fn]}{:lang="ruby"} but without the label, because is what we want to find.

Playing with classifiers KNN

In a recent project we needed a classifier, so we’ve been playing with clasiffiers last months. At the end we choosed SVM for the project, but I want to show you one that I like and it’s easy to implement, the k-nearest neighbors (k-NN) classifier.

Vicente Cuéllar

CEO

Civil Engineer with more than 15 years of experience working in software development and high performance computing. Has ample experience in such fields as Machine Learning, Big Data and Numerical Modelization. Has been working for many years in parallelization (CUDA and MPI) and a has passion for data visualization (OpenGL, HTML5, Canvas, etc.)

Stay Connected

Anything you want to ask? Feel free to contact us!

Wavecrafters

GPU Computing Experts

Our Services

GPU Computing

Visualization

Machine Learning

Big Data

Why Choose Us

Co-development

Closed Price

Platform Independent

Customized training

Our Blog

Exclusive prefix sum scan

KNN normalizing the data

Playing with classifiers KNN

Our Team

Vicente Cuéllar

Contact Us

Stay Connected

Anything you want to ask? Feel free to contact us!

Call Us

+34 91 433 62 39

Contact

contact@wavecrafters.com

Twitter

@wavecrafters