MCV/DEVELOP News, events, research and jobs from the games industry
Montreal tech outfit Fabric Engine has made public the beta version of its eponymous web application development platform.
The Fabric Engine itself has been designed to offer web developers and those crafting games for online to bring multi-threaded performance to their projects without resorting to compiled languages. The tech’s creators hope that it will change the fact that despite the acceleration of computer hardware, web applications remain single-threaded, and are as such unable to tap into the computing muscle of multi-core hardware.
“If you want performance in your applications, you have to write native, compiled code. Native apps have constraints around platform, deployment and cost of development that force developers to make tough choices in their development process,” confirmed Paul Doyle, CEO of Fabric Engine. “But the web offers an open platform for application development.
"Ideally, all of our applications would be web-based; they’re quick and easy to build, simple to deploy and they work everywhere. Unfortunately, apps that use dynamic languages like JavaScript are painfully slow when compared to native, multi-threaded applications. Fabric Engine alleviates that problem by allowing developers to create computationally powerful web applications using languages and tools they are familiar with. Whether on the web server or in the browser, Fabric Engine brings native performance where it’s needed.”
Fabric Engine enables the development of web applications and games in two parts; firstly the technologies that define the interface, interaction and layout of the application, such as HTML5, JavaScript or CSS, and secondly the core of the application that is written using the company’s proprietary performance-specific JavaScript variant, known as Fabric KL.
"In order for Fabric-enabled web apps to tap into multi-core processing power, Fabric apps are described as a dependency graph, which defines where tasks can be performed in parallel, and where the same operations can be applied to lots of data at the same time (SIMD)," reads a statement to the press. "Fabric Engine then handles CPU multi-threading automatically and compiles the application locally.
Fabric Engine also includes an embedded compiler that lets developer dynamically compile their applications on target.
