Vector Fabrics has added vfLabs to its cloud-based tool suite for parallelizing sequential C code that gives software engineers an early opportunity to try out functionality as it is developed.
It allows users to selectively augment the capabilities of the vfAnalyst tool, launched in May 2010, which enables software engineers to create multicore implementations of sequential C programs.
vfAnalyst is undergoing continuous improvement based on user feedback, with features being added much more regularly than can be done with standard tool distributions due to its cloud implementation. vfLabs provides access to pre-release versions of these features, allowing users to experiment with them and provide ongoing feedback.
At its launch, vfLabs includes access to a new vfStream library of streaming data communications routines. When parallelizing a program, provisions must be made for threads to communicate with each other if streaming dependencies are cut. vfAnalyst shows where those dependences are and where the communications are needed.
The vfStream library provides a means of implementing the communications, saving design and debugging time as compared to writing such a library from scratch.
Also included in this release of vfLabs are scheduling and resource utilization views that make it easier to see where performance blockages are and what resources are being used as the program executes.
Over time, features initially provided in vfLabs will migrate to full production status, being replaced by new features for customer evaluation. vfLabs will also provide pre-release access to features of future Vector Fabrics tools as they become available.