Blog
We’re moving into a bigger office
Vector Fabrics is expanding, and as a result we’ll be moving into a bigger and nicer office on January 13th. We’ll still be in Eindhoven, the smartest region in the world, in the city center. Restaurants, hotels and the famous PSV soccer stadium are within a few walking minutes. The railway station is roughly a 10-minute walk. Perhaps it goes without saying, but there will be no disruption in our service.
Happy new year!
We’re excited about 2012. More and more multicore products are being introduced, alleviating the need to optimize software applications for these high-performance processors. We’re moving into a new and larger office, as we’re expanding our team. We have many more good things to come in 2012. You can try our latest release, with many new features, and sign up for a free trial here. Click on the image below to view our happy new year wish. Multicore is our bread and butter, and thus our message can be viewed in single-core or multi-core fashion. Enjoy!
Multithreading examples for C programs, Part 1
Have you ever parallelized a C or C++ program? Then you got a slowdown first, didn’t you? Parallelization is a tough call because of data dependencies hiding behind pointers, unclear multithreading overheads, unexpected processor and OS behaviour, let alone potential starvation and deadlocks. Luckily, in the past 40 years computer engineers have been constantly accumulating best practices of industry-proven solutions to common concurrency challenges. They call them parallelization patterns.
Measuring Power Consumption of the OMAP4430 using the PandaBoard
Recently, I performed a series of power measurements on Texas Instrument’s OMAP4430 mobile SoC. The aim of this experiment was to confirm that we can perform such energy consumption measurements with sufficient accuracy for software source-code optimization. The software parallelization tools of Vector Fabrics can be used to lower power consumption, and measurements like these are needed to verify our results. Furthermore, it was my goal to evaluate speed/power trade offs in re-mapping a compute kernel from the multi-core ARM Cortex-A9 host processor onto the PowerVR SGX graphics processing unit.
Parallelizing a sequential C program using vfEmbedded
In this blog post I will explain how to use vfEmbedded to parallelize a small sequential C program. The program in question calculates the electric potential in a two dimensional plane created by a discrete set of point charges. The analysis and implementation of the recipes provided by vfEmbedded were done in three hours, speeding up the program with a factor 3.4.