Get a quick overview of the latest and greatest technologies for application acceleration!
Ready to reach an order of magnitude or more in performance increase? Tired of software optimizations, hardware updates and algorithm tweaking that only give you incremental improvements in performance? Think out of the box and put some FPGAs or GPUs at work and you will get:
- Increased computational speed with 10x-100x or
- Reduced system size with 10x or
- Reduced power consumption with 10x-100x
Synective offers solutions that give you tomorrow’s performance already today!
Reducing system sizes, power consumption and overall cost has become increasingly important to stay competitive in today’s global economy. The focus on “green computing” also asks for new solutions. Traditional software optimization, hardware updates and system tuning are not enough. By using hybrid computing, where the CPU cycle hungry parts of your software are off-loaded to “accelerators”, your application will in many cases get a dramatic speed-up. This technique works in both embedded systems as well as for desktop applications and for cluster solutions in HPC.
Reduce power consumption
When computations are moved from a CPU to an accelerator, with a power consumption being less or equal to the CPU, but performing 10x to 100x, the total power consumption for a specific computation is reduced with the same factor. And the power required for cooling is reduced at the same time!
Do more in less time
Introducing accelerators in a system of a specific size will bring down the computation time from maybe days to hours – which will open up completely new possibilities! Accelerated research, higher accuracy in calculations or higher throughput!
Reduce system footprint – desktop supercomputing
With a magnitude or two in increased system performance, all of a sudden something that previously required a compute cluster can be replaced with an accelerated desktop computer. That can give your PC-based product compute cluster capabilities or each researcher a personal supercomputer!
The two most common types of accelerator are FPGAs and GPUs. FPGAs are a type of devices that can be seen as reprogrammable hardware – by programming them with “a dedicated hardware solution, they become “co-processors” tailored for your specific algorithm. Read more about FPGA based solutions here.
GPUs, Graphics Processing Units, i.e. the ordinary devices used to render the graphics on your computer’s screen – are simple miniaturized, massively parallel processing systems that can also make generic calculations and therefore be used as accelerators. Read more about GPU based solutions here