Debian Codegraph @

results/NVIDIA_CUDA/GeForce_GTX_465.log

raw
1
2 Platform: NVIDIA CUDA
3 Device: GeForce GTX 465
4 Driver version : 325.15 (Linux x86)
5 Compute units : 11
6
7 Global memory bandwidth (GBPS)
8 float : 85.28
9 float2 : 86.04
10 float4 : 87.34
11 float8 : 45.51
12 float16 : 22.70
13
14 Single-precision compute (GFLOPS)
15 float : 843.19
16 float2 : 835.39
17 float4 : 836.46
18 float8 : 831.69
19 float16 : 827.41
20
21 Double-precision compute (GFLOPS)
22 double : 106.75
23 double2 : 106.66
24 double4 : 106.41
25 double8 : 106.04
26 double16 : 105.21
27
28 Integer compute (GIOPS)
29 int : 426.07
30 int2 : 425.49
31 int4 : 426.23
32 int8 : 426.26
33 int16 : 426.24
34
35 Transfer bandwidth (GBPS)
36 enqueueWriteBuffer : 0.69
37 enqueueReadBuffer : 0.47
38 enqueueMapBuffer(for read) : 0.35
39 memcpy from mapped ptr : 0.53
40 enqueueUnmap(after write) : 1.57
41 memcpy to mapped ptr : 0.49
42
43 Kernel launch latency : 15.87 us
44