1Platform: NVIDIA CUDA 2 Device: GeForce GTX TITAN 3 Driver version : 352.21 (Linux x64) 4 Compute units : 14 5 Clock frequency : 928 MHz 6 7 Global memory bandwidth (GBPS) 8 float : 227.55 9 float2 : 235.48 10 float4 : 244.46 11 float8 : 203.42 12 float16 : 150.25 13 14 Single-precision compute (GFLOPS) 15 float : 3205.36 16 float2 : 4043.81 17 float4 : 4002.08 18 float8 : 3948.40 19 float16 : 3705.41 20 21 Double-precision compute (GFLOPS) 22 double : 1629.84 23 double2 : 1628.86 24 double4 : 1625.99 25 double8 : 1619.69 26 double16 : 1606.94 27 28 Integer compute (GIOPS) 29 int : 815.49 30 int2 : 814.82 31 int4 : 814.34 32 int8 : 812.88 33 int16 : 814.81 34 35 Transfer bandwidth (GBPS) 36 enqueueWriteBuffer : 4.08 37 enqueueReadBuffer : 3.71 38 enqueueMapBuffer(for read) : 6.04 39 memcpy from mapped ptr : 6.47 40 enqueueUnmap(after write) : 6.35 41 memcpy to mapped ptr : 6.45 42 43 Kernel launch latency : 6.58 us 44