ttgLabs, a Russian software startup with a focus on development and optimization of HPC software, is announcing the pre-release version of the TTG Apptimizer toolkit. This software dramatically facilitates the optimization of computationally-demanding applications on graphics accelerators and heterogeneous HPC systems enabling the developers to easily work around the most tedious tasks of programming for graphics accelerators.
As opposed to traditional optimization techniques, TTG Apptimizer actively uses the concept of software autotuning. After a short self-learning cycle, an application developed with TTG Apptimizer will tailor itself to the hardware platform it is running on and to the data it has to process thus approaching the highest possible performance for a particular computational environment.
Such an autotuning is dynamic by its nature, i.e. the application permanently monitors the hardware configuration and data parameters and readapts itself to any changes. Currently, TTG Apptimizer is available in CUDA and OpenCL editions for Microsoft Windows and Linux and can be used with graphics processors from Intel, NVIDIA, and AMD.
The pre-release version, available at no charge upon request, supports only one processing node and has a limited functionality. A commercial version of TTG Apptimizer will be available by the year end. The company also plans to release a mini-cluster and supercomputer versions of the toolkit supporting up to 20 and up to 1,000 nodes, respectively.
