• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1Platform: Apple
2  Device: AMD Radeon HD - FirePro D700 Compute Engine
3    Driver version  : 1.2 (Mar 16 2017 18:19:56) (Macintosh)
4    Compute units   : 32
5    Clock frequency : 850 MHz
6
7    Global memory bandwidth (GBPS)
8      float   : 182.39
9      float2  : 189.53
10      float4  : 195.73
11      float8  : 101.98
12      float16 : 53.33
13
14    Single-precision compute (GFLOPS)
15      float   : 2593.97
16      float2  : 2591.22
17      float4  : 2584.96
18      float8  : 2572.53
19      float16 : 2543.34
20
21    No half precision support! Skipped
22
23    Double-precision compute (GFLOPS)
24      double   : 654.65
25      double2  : 654.62
26      double4  : 653.86
27      double8  : 652.82
28      double16 : 650.63
29
30    Transfer bandwidth (GBPS)
31      enqueueWriteBuffer         : 11.12
32      enqueueReadBuffer          : 11.92
33      enqueueMapBuffer(for read) : 94.12
34        memcpy from mapped ptr   : 6.79
35      enqueueUnmap(after write)  : 7550.93
36        memcpy to mapped ptr     : 7.64
37
38    Kernel launch latency : 9.63 us
39
40  Device: AMD Radeon HD - FirePro D700 Compute Engine
41    Driver version  : 1.2 (Mar 16 2017 18:19:56) (Macintosh)
42    Compute units   : 32
43    Clock frequency : 850 MHz
44
45    Global memory bandwidth (GBPS)
46      float   : 184.77
47      float2  : 191.74
48      float4  : 197.64
49      float8  : 102.48
50      float16 : 53.56
51
52    Single-precision compute (GFLOPS)
53      float   : 2599.49
54      float2  : 2594.67
55      float4  : 2590.64
56      float8  : 2576.32
57      float16 : 2547.82
58
59    No half precision support! Skipped
60
61    Double-precision compute (GFLOPS)
62      double   : 654.91
63      double2  : 654.89
64      double4  : 654.75
65      double8  : 654.34
66      double16 : 651.34
67
68    Transfer bandwidth (GBPS)
69      enqueueWriteBuffer         : 10.05
70      enqueueReadBuffer          : 9.31
71      enqueueMapBuffer(for read) : 85.90
72        memcpy from mapped ptr   : 6.73
73      enqueueUnmap(after write)  : 7389.83
74        memcpy to mapped ptr     : 7.70
75
76    Kernel launch latency : 10.21 us
77