1 2Platform: AMD Accelerated Parallel Processing 3 Device: gfx1012:xnack- (RX 5500XT) 4 Driver version : 3361.0 (HSA1.1,LC) (Linux x64) 5 Compute units : 11 6 Clock frequency : 1900 MHz 7 8 Global memory bandwidth (GBPS) 9 float : 190.36 10 float2 : 182.03 11 float4 : 171.64 12 float8 : 157.88 13 float16 : 154.64 14 15 Single-precision compute (GFLOPS) 16 float : 5046.67 17 float2 : 4936.51 18 float4 : 4887.78 19 float8 : 4871.37 20 float16 : 4796.57 21 22 Half-precision compute (GFLOPS) 23 half : 2544.83 24 half2 : 9875.69 25 half4 : 9771.45 26 half8 : 9731.20 27 half16 : 9533.27 28 29 Double-precision compute (GFLOPS) 30 double : 323.84 31 double2 : 323.33 32 double4 : 322.61 33 double8 : 321.09 34 double16 : 318.06 35 36 Integer compute (GIOPS) 37 int : 1025.25 38 int2 : 1025.22 39 int4 : 1021.86 40 int8 : 1018.56 41 int16 : 1012.24 42 43 Integer compute Fast 24bit (GIOPS) 44 int : 4738.49 45 int2 : 4805.52 46 int4 : 4799.40 47 int8 : 4682.88 48 int16 : 4766.09 49 50 Transfer bandwidth (GBPS) 51 enqueueWriteBuffer : 15.64 52 enqueueReadBuffer : 15.34 53 enqueueWriteBuffer non-blocking : 15.70 54 enqueueReadBuffer non-blocking : 15.44 55 enqueueMapBuffer(for read) : 613566.81 56 memcpy from mapped ptr : 15.37 57 enqueueUnmap(after write) : 1227133.62 58 memcpy to mapped ptr : 15.70 59 60 Kernel launch latency : 12.77 us 61 62