caffeinade on 15 March 2017
http://www.nvidia.com/object/tegra-x1-processor.html
You can calculate the FLOPs by multiplying clockrate then further multiplying by two for the Maxwell architecture.
the stock x1 is ~ 1GHz.
So 1000 * 256 (cores) * 2 = 0.512 TFLOPs
This is of couse for single precision.







