• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1Platform: Portable Computing Language
2  Device: TITAN V
3    Driver version  : 3.0-rc2 (Linux x64)
4    Compute units   : 80
5    Clock frequency : 1455 MHz
6
7    Global memory bandwidth (GBPS)
8      float   : 559.25
9      float2  : 582.61
10      float4  : 605.03
11      float8  : 563.36
12      float16 : 466.89
13
14    Single-precision compute (GFLOPS)
15      float   : 13650.31
16      float2  : 13676.21
17      float4  : 13654.21
18      float8  : 13605.12
19      float16 : 13497.45
20
21    No half precision support! Skipped
22
23    Double-precision compute (GFLOPS)
24      double   : 6851.16
25      double2  : 6839.41
26      double4  : 6821.27
27      double8  : 6802.04
28      double16 : 6750.90
29
30    Integer compute (GIOPS)
31      int   : 13622.78
32      int2  : 13650.92
33      int4  : 13649.84
34      int8  : 13652.66
35      int16 : 13646.51
36
37    Integer compute Fast 24bit (GIOPS)
38      int   : 13622.20
39      int2  : 13650.49
40      int4  : 13648.90
41      int8  : 13653.82
42      int16 : 13646.95
43
44    Transfer bandwidth (GBPS)
45      enqueueWriteBuffer              : 9.75
46      enqueueReadBuffer               : 9.60
47      enqueueWriteBuffer non-blocking : 9.76
48      enqueueReadBuffer non-blocking  : 9.61
49      enqueueMapBuffer(for read)      : 121195.77
50        memcpy from mapped ptr        : 9.11
51      enqueueUnmap(after write)       : 8.42
52        memcpy to mapped ptr          : 9.45
53
54    Kernel launch latency : 27.61 us
55
56  Device: TITAN V
57    Driver version  : 3.0-rc2 (Linux x64)
58    Compute units   : 80
59    Clock frequency : 1455 MHz
60
61    Global memory bandwidth (GBPS)
62      float   : 559.64
63      float2  : 582.45
64      float4  : 604.99
65      float8  : 562.63
66      float16 : 469.28
67
68    Single-precision compute (GFLOPS)
69      float   : 13646.88
70      float2  : 13671.93
71      float4  : 13649.91
72      float8  : 13599.42
73      float16 : 13493.17
74
75    No half precision support! Skipped
76
77    Double-precision compute (GFLOPS)
78      double   : 6843.55
79      double2  : 6834.11
80      double4  : 6820.15
81      double8  : 6797.91
82      double16 : 6743.83
83
84    Integer compute (GIOPS)
85      int   : 13618.75
86      int2  : 13618.60
87      int4  : 13639.15
88      int8  : 13639.72
89      int16 : 13633.81
90
91    Integer compute Fast 24bit (GIOPS)
92      int   : 13612.85
93      int2  : 13642.47
94      int4  : 13639.65
95      int8  : 13641.46
96      int16 : 13634.24
97
98    Transfer bandwidth (GBPS)
99      enqueueWriteBuffer              : 8.94
100      enqueueReadBuffer               : 10.66
101      enqueueWriteBuffer non-blocking : 8.94
102      enqueueReadBuffer non-blocking  : 10.69
103      enqueueMapBuffer(for read)      : 13.16
104        memcpy from mapped ptr        : 8.71
105      enqueueUnmap(after write)       : 12.39
106        memcpy to mapped ptr          : 8.97
107
108    Kernel launch latency : 193.03 us
109