Debian Codegraph @

results/NVIDIA_CUDA/GeForce_RTX_2080.log

raw
1
2 Platform: NVIDIA CUDA
3 Device: GeForce RTX 2080
4 Driver version : 410.73 (Linux x64)
5 Compute units : 46
6 Clock frequency : 1815 MHz
7
8 Global memory bandwidth (GBPS)
9 float : 362.93
10 float2 : 382.42
11 float4 : 391.26
12 float8 : 400.79
13 float16 : 364.98
14
15 Single-precision compute (GFLOPS)
16 float : 11258.41
17 float2 : 11248.28
18 float4 : 11228.37
19 float8 : 11166.76
20 float16 : 11064.75
21
22 No half precision support! Skipped
23
24 Double-precision compute (GFLOPS)
25 double : 354.32
26 double2 : 353.24
27 double4 : 351.23
28 double8 : 349.27
29 double16 : 346.67
30
31 Integer compute (GIOPS)
32 int : 11085.63
33 int2 : 11005.45
34 int4 : 11002.92
35 int8 : 10991.37
36 int16 : 10955.21
37
38 Transfer bandwidth (GBPS)
39 enqueueWriteBuffer : 5.88
40 enqueueReadBuffer : 6.49
41 enqueueMapBuffer(for read) : 5.96
42 memcpy from mapped ptr : 14.68
43 enqueueUnmap(after write) : 6.18
44 memcpy to mapped ptr : 14.82
45
46 Kernel launch latency : 3.85 us
47