Hi Marcin. Unfortunately, there is no workaround other than to use a GPU with compute capability 1.3 or higher, as the error message states.
The compute capability restriction is in place for a number of reasons. First, compute capability 1.3 cards were the first to offer double precision support, which is the default precision in MATLAB. Furthermore, starting with compute capability 1.3, NVIDIA GPUs offer greater single precision arithmetic accuracy and IEEE compliance for all computations.
Regards,
Jill
Best Answer