Blog
From Google Summer of Code and the Ruby Grant: NMatrix on JRuby, ArrayFire and CUDA bindings for Ruby, and calling D from Ruby.
RBCuda is a software library designed to integrate with Ruby. It provides an interface for leveraging NVIDIA CUDA technology to accelerate computations on GPUs (Graphics Processing Units). This can significantly speed up tasks like matrix operations, simulations, and data processing by harnessing...
I have been working on a new project called RbCUDA.
GSoC 2017 is about to end. This post summarises my work during the course of summer.
ArrayFire-rb has been interfaced to NMatrix so that they can work together. Sometimes, copying data from GPU memory to CPU memory and vice versa can be slower than the computation time. Hence, a lot of developers prefer to use GPU computing for the most time consuming task. The interface between ...
ArrayFire can run not just on GPU devices but also on CPU devices, which is one of its really cool features. With version 3.2 the unified-backend helps in changing the ArrayFire backend on the fly. When using the unified-backend, the preference order for the default backend is CUDA > OpenCL &g...
ArrayFire has Device and Util class that contains of methods which can be used to manage the CPU and GPU devices and various utility functions like printing and storing the arrays. This blog will explain my work regarding the bindings I have created and how to use them.
This post explains about the Statistics and Random class of arrayfire gem. ArrayFire::Statistics class consists of methods that can be used to operate on Af_Array for statistical analysis. ArrayFire::Random class contains of methods to generate Random numbers and array of Random numbers on the GPU.
This post covers how to use BLAS and LAPACK routines with ArrayFire-rb, and how these functions have been implemented by creating Ruby bindings to ArrayFire C API.
Language bindings are interfaces that allow code written in one programming language to interact with or use functionalities provided by libraries or code written in another programming language. They serve as a bridge between different programming environments, making it possible for developers ...
This post covers how I implemented the test-suite and Algorithm class for ArrayFire-rb using ArrayFire C API, and how to use Algorithm class methods with Af_Array.
This post covers how to create arrays using arrayfire-rb, perform elementwise operations like addition, multiplication, etc. and how these functions have been implemented by creating Ruby bindings to ArrayFire C API.
I have been working on creating ArrayFire bindings for Ruby. ArrayFire is an opensource library that is very useful and highly popular for GPGPU computings. It has strong abstractions that makes it very easy for a programmer to benefit from GPU without being bothered about the configurations.
In this blog and a few others that will follow, I would share my expereiences creating Ruby Bindings for C/C++ libraries. It covers a practical example rather than a Hello world! or chairs and tables examples.
I have been working on “Port NMatrix to JRuby” as my GSoC project. I am pleased to announce that JRuby is ready for Nmatrix users.
For my GSoC 2016 project of JRuby port of NMatrix, we worked on testing NMatrix-JRuby with real-life data. We started with mixed-models gem.
NMatrix for MRI has become a fairly well-established project. NMatrix is SciRuby’s numerical matrix core, implementing dense matrices as well as two types of sparse (linked-list-based and Yale/CSR). The backend has been written in C/C++ and relies on ATLAS/ CBLAS / CLAPACK and standard LAPACK for...
I built a ruby-gem “mitab” that helps in parsing a PSI-Mitab file.