Editor’s note: DevOps Dossier is a new feature that highlights the tools, people and companies that make-up the emerging DevOps culture. Stories cross agile development practices with enterprise class, mission critical operations standards. Fabric Engine is a multithreading engine that integrates with dynamic scripting languages. It can enable applications written in JavaScript (and, by extension, Node.js), PHP, Python and Ruby will soon be able to take advantage of the same sort of multithreaded performance that C++ does by calling Fabric Engine from within the application. That means these sort of languages could soon be used for the sort of computation intensive applications that have been typically been reserved for C++ and Java.

“Everyone knows you don’t use Node.js for compute run problems,” Fabric Engine CEO Paul Doyle told me. “When you’re doing heavy computation, compiled code is 3-5 times faster.” According to benchmarking data released by the company, JavaScript application using Fabric Engine can run nearly as fast as a C++ application. “We had a lot of people looking at the code thinking we’d just written bad C++ code or were misrepresenting the results,” Doyle says. “We were able to make some improvements to the C++ code but we were still able to get JavaScript to run almost as well.”

For example, LinkedIn is using Node.js as the main backend for its mobile applications, but it is still not using Node.js for its recommendation engine. Fabric Engine could make that possible one day.

The only catch is that you need to use KL, Fabric Engine’s JavaScript-like kernel language, for the parts of the application that would otherwise be performance constrained. Fabric Engine then takes those parts and uses LLVM to compile them to bitcode. The KL elements can be embedded inside the application. Fabric Engine doesn’t really accelerate the dynamic scripting languages, it just handles some of the heavy lifting, and does so in way that should be much easier than writing C++ by hand.

The company is planning a relaunch at the end of the month, moving from a model that was focused on helping JavaScript developers get more performance within the browser to a model more focused on helping server side dynamic language developers get more performance.

Doyle said that although the company hasn’t settled on its exact business model or pricing, it will likely release at least part of it as open source and sell a separate commercial version.

Fabric Engine was founded in 2010 by Doyle, Peter Zion and Phil Taylor and is funded by Real Ventures. The three worked together at a 3D animation software company called Softimage, which was acquired by Autodesk in 2008. Softimage abstracted much of the low level programming elements such as thread management away from developers, allowing them to focus on solving other problems. That showed Doyle that there was a lot of value in giving developers more performance by abstracting away difficult problems.

The project is already gaining a lot of buzz. The company presented at NodeSummit and won the Judge’s Choice award in the NodeJam startup competition. So far it works with V8 (the JavaScript engine used by Node.js) and Python. Ruby and PHP will be added soon.