The specifics of IT projects related to the software development, porting and optimization for graphics accelerators and systems with heterogeneous architecture hampers a prior prediction of their particular parameters. However, to make a well-informed decision on initiating such a project, the customer needs reliable preliminary estimates of the potential increase in application performance when compared to the CPU-only platform as well as of resources that should be allocated for the project.
ttgLabs company has unique techniques that enable (i) to analyze the potential of effective parallelization of computational core and (ii) to estimate with high reliability the range of application performance gain after its porting to graphics accelerators and its further optimization. Our experts will fulfill such estimation practically without coding thus allowing our customers to save a lot of money on preliminary studies before giving ‘green light’ to GPGPU software projects. The estimates will account for the specifics of computational task and the configuration of a chosen hardware platform.
