what is the total no. of threads that can be launched by GTX 295 and how do i find it for any card? the matrix multiplication code given in the programming guide stops working for 8192*8192 matrix. what could be the possible reason??

what is the total no. of threads that can be launched by GTX 295 and how do i find it for any card?
maximum number of blocks in a 2D grid: 65535 x 65535
maximum threads per block: 512
maximum number of threads = 65535 x 65535 x 512 
the matrix multiplication code given in the programming guide stops working for 81928192 matrix. what could be the possible reason?
81928192 matrix requires 5.6 s to compute C=A*B, it near threashold of watchdog problem.
You can use cudaGetDeviceProperties to query maximum threads No. for your specific system.
See http://developer.download.nvidia.com/compu…08074d09b7d520c