Skip to content

Instantly share code, notes, and snippets.

@jramapuram

jramapuram/nvidiap2p

Created May 1, 2020 00:28
Show Gist options
  • Save jramapuram/e470421afb3fa6fd6b060057a3b6652e to your computer and use it in GitHub Desktop.
Save jramapuram/e470421afb3fa6fd6b060057a3b6652e to your computer and use it in GitHub Desktop.
Download ZIP
Raw
nvidiap2p
ramapur0@gpu010:~$ ~/samples/bin/x86_64/linux/release/p2pBandwidthLatencyTest
[P2P (Peer-to-Peer) GPU Bandwidth Latency Test]
Device: 0, TITAN Xp, pciBusID: 4, pciDeviceID: 0, pciDomainID:0
Device: 1, TITAN Xp, pciBusID: 5, pciDeviceID: 0, pciDomainID:0
Device: 2, TITAN Xp, pciBusID: 8, pciDeviceID: 0, pciDomainID:0
Device: 3, TITAN Xp, pciBusID: 9, pciDeviceID: 0, pciDomainID:0
Device: 4, TITAN Xp, pciBusID: 83, pciDeviceID: 0, pciDomainID:0
Device: 5, TITAN Xp, pciBusID: 84, pciDeviceID: 0, pciDomainID:0
Device: 6, TITAN Xp, pciBusID: 87, pciDeviceID: 0, pciDomainID:0
Device: 7, TITAN Xp, pciBusID: 88, pciDeviceID: 0, pciDomainID:0
Device=0 CAN Access Peer Device=1
Device=0 CAN Access Peer Device=2
Device=0 CAN Access Peer Device=3
Device=0 CANNOT Access Peer Device=4
Device=0 CANNOT Access Peer Device=5
Device=0 CANNOT Access Peer Device=6
Device=0 CANNOT Access Peer Device=7
Device=1 CAN Access Peer Device=0
Device=1 CAN Access Peer Device=2
Device=1 CAN Access Peer Device=3
Device=1 CANNOT Access Peer Device=4
Device=1 CANNOT Access Peer Device=5
Device=1 CANNOT Access Peer Device=6
Device=1 CANNOT Access Peer Device=7
Device=2 CAN Access Peer Device=0
Device=2 CAN Access Peer Device=1
Device=2 CAN Access Peer Device=3
Device=2 CANNOT Access Peer Device=4
Device=2 CANNOT Access Peer Device=5
Device=2 CANNOT Access Peer Device=6
Device=2 CANNOT Access Peer Device=7
Device=3 CAN Access Peer Device=0
Device=3 CAN Access Peer Device=1
Device=3 CAN Access Peer Device=2
Device=3 CANNOT Access Peer Device=4
Device=3 CANNOT Access Peer Device=5
Device=3 CANNOT Access Peer Device=6
Device=3 CANNOT Access Peer Device=7
Device=4 CANNOT Access Peer Device=0
Device=4 CANNOT Access Peer Device=1
Device=4 CANNOT Access Peer Device=2
Device=4 CANNOT Access Peer Device=3
Device=4 CAN Access Peer Device=5
Device=4 CAN Access Peer Device=6
Device=4 CAN Access Peer Device=7
Device=5 CANNOT Access Peer Device=0
Device=5 CANNOT Access Peer Device=1
Device=5 CANNOT Access Peer Device=2
Device=5 CANNOT Access Peer Device=3
Device=5 CAN Access Peer Device=4
Device=5 CAN Access Peer Device=6
Device=5 CAN Access Peer Device=7
Device=6 CANNOT Access Peer Device=0
Device=6 CANNOT Access Peer Device=1
Device=6 CANNOT Access Peer Device=2
Device=6 CANNOT Access Peer Device=3
Device=6 CAN Access Peer Device=4
Device=6 CAN Access Peer Device=5
Device=6 CAN Access Peer Device=7
Device=7 CANNOT Access Peer Device=0
Device=7 CANNOT Access Peer Device=1
Device=7 CANNOT Access Peer Device=2
Device=7 CANNOT Access Peer Device=3
Device=7 CAN Access Peer Device=4
Device=7 CAN Access Peer Device=5
Device=7 CAN Access Peer Device=6
***NOTE: In case a device doesn't have P2P access to other one, it falls back to normal memcopy procedure.
So you can see lesser Bandwidth (GB/s) and unstable Latency (us) in those cases.
P2P Connectivity Matrix
D\D 0 1 2 3 4 5 6 7
0 1 1 1 1 0 0 0 0
1 1 1 1 1 0 0 0 0
2 1 1 1 1 0 0 0 0
3 1 1 1 1 0 0 0 0
4 0 0 0 0 1 1 1 1
5 0 0 0 0 1 1 1 1
6 0 0 0 0 1 1 1 1
7 0 0 0 0 1 1 1 1
Unidirectional P2P=Disabled Bandwidth Matrix (GB/s)
D\D 0 1 2 3 4 5 6 7
0 455.81 9.88 11.38 11.16 10.99 10.88 11.14 10.83
1 9.91 456.34 11.03 10.05 9.69 8.81 10.01 9.90
2 11.01 11.07 456.87 9.79 10.71 10.80 10.96 10.70
3 11.11 11.24 10.04 457.41 11.01 11.19 11.27 10.81
4 11.20 11.18 10.77 11.08 454.22 9.92 11.26 11.30
5 11.09 11.09 11.19 11.01 10.11 458.48 11.29 10.72
6 10.37 9.64 8.02 9.40 9.51 10.50 457.41 9.75
7 10.90 10.63 10.83 10.87 10.95 10.97 10.02 457.41
Unidirectional P2P=Enabled Bandwidth (P2P Writes) Matrix (GB/s)
D\D 0 1 2 3 4 5 6 7
0 448.99 13.18 10.27 10.27 10.94 10.94 10.91 11.07
1 13.19 455.81 10.27 10.27 11.11 10.85 11.07 10.39
2 9.82 8.04 456.34 11.70 8.37 10.19 10.80 10.78
3 10.27 10.28 13.18 457.41 10.80 10.83 10.76 11.11
4 11.03 11.14 11.09 10.99 458.48 13.19 10.28 10.28
5 11.09 11.20 10.79 10.54 13.05 457.41 6.22 7.11
6 10.12 10.01 10.74 10.87 10.28 10.28 457.41 13.19
7 10.84 10.55 11.16 11.05 10.27 10.28 13.18 456.87
Bidirectional P2P=Disabled Bandwidth Matrix (GB/s)
D\D 0 1 2 3 4 5 6 7
0 457.14 10.34 19.68 18.81 19.39 19.57 19.25 18.93
1 9.89 456.07 17.73 17.96 19.21 18.90 19.16 19.30
2 19.08 19.19 459.02 10.13 19.47 19.50 19.60 19.49
3 19.20 19.64 10.52 456.83 19.36 19.17 18.29 15.32
4 16.23 17.30 19.02 19.26 459.56 10.13 18.95 19.15
5 19.09 19.74 19.68 19.62 10.43 458.75 19.20 18.96
6 16.27 15.83 18.33 18.25 18.85 18.95 455.29 10.14
7 19.15 19.11 18.97 19.10 19.52 19.75 10.49 458.21
Bidirectional P2P=Enabled Bandwidth Matrix (GB/s)
D\D 0 1 2 3 4 5 6 7
0 455.27 25.29 19.66 19.66 19.53 19.09 19.45 17.58
1 22.75 457.40 17.00 18.16 19.36 19.11 18.56 18.72
2 19.69 19.72 455.27 25.39 19.68 19.52 19.36 19.47
3 19.06 19.00 22.55 456.07 15.95 16.50 18.06 18.67
4 18.93 19.06 19.16 19.17 455.81 25.37 19.79 19.81
5 19.56 19.86 19.80 19.65 24.66 457.94 18.63 15.07
6 17.35 17.73 18.88 19.15 19.80 19.76 457.14 25.39
7 18.82 18.81 18.66 19.24 19.80 19.80 25.37 456.37
P2P=Disabled Latency Matrix (us)
GPU 0 1 2 3 4 5 6 7
0 1.27 17.18 14.94 14.81 11.96 12.49 15.13 14.61
1 15.65 1.24 10.43 10.30 10.62 12.47 15.23 10.63
2 12.36 18.42 1.23 18.46 10.49 11.87 11.99 10.68
3 15.44 10.36 13.41 1.23 13.43 12.44 12.17 13.74
4 12.32 14.39 12.17 13.05 1.20 17.34 13.15 11.59
5 10.55 10.44 14.58 13.10 11.02 1.22 18.15 15.22
6 12.49 12.15 10.56 13.12 12.85 12.07 1.27 15.28
7 11.53 12.00 10.63 11.20 12.31 14.17 13.11 1.29
CPU 0 1 2 3 4 5 6 7
0 4.02 9.86 9.66 9.49 10.64 13.26 9.54 9.91
1 10.08 3.44 8.77 8.80 9.93 9.81 8.78 8.92
2 9.81 8.64 3.40 8.56 10.48 9.89 8.82 8.82
3 9.76 8.81 8.54 3.51 11.69 9.70 8.71 8.79
4 10.52 11.54 9.66 9.63 4.00 10.49 12.94 9.82
5 10.64 9.63 11.54 9.59 10.87 3.99 9.84 13.41
6 9.71 8.95 8.59 8.56 12.40 9.90 3.48 8.81
7 10.49 8.79 8.69 8.66 13.63 9.95 9.19 3.46
P2P=Enabled Latency (P2P Writes) Matrix (us)
GPU 0 1 2 3 4 5 6 7
0 1.24 1.09 1.47 1.50 16.67 18.95 11.78 13.32
1 1.03 1.68 1.48 1.47 10.37 10.52 10.99 10.58
2 1.50 1.47 1.24 1.01 10.69 11.37 12.17 11.15
3 1.44 1.50 1.06 1.25 13.67 11.05 11.50 12.66
4 11.53 13.96 13.16 13.26 1.20 1.17 1.51 1.53
5 11.92 14.19 13.69 11.74 1.15 1.25 1.57 1.54
6 10.54 11.31 10.64 13.92 1.51 1.50 1.30 1.12
7 12.73 15.66 10.47 10.67 1.56 1.56 1.08 1.29
CPU 0 1 2 3 4 5 6 7
0 4.03 2.78 2.81 2.75 10.44 13.10 10.33 9.53
1 3.25 5.11 2.67 2.66 10.16 9.90 9.02 9.37
2 2.66 2.67 3.60 2.66 9.55 14.02 8.62 8.90
3 2.59 2.72 2.67 3.52 10.47 9.61 8.65 8.80
4 10.49 9.88 9.66 9.75 4.00 2.85 2.87 2.79
5 10.44 9.62 9.82 9.69 2.86 4.00 2.84 3.52
6 9.83 8.86 8.65 8.74 2.66 2.66 3.59 2.61
7 10.11 9.09 8.68 8.68 2.66 2.76 2.67 3.62
NOTE: The CUDA Samples are not meant for performance measurements. Results may vary when GPU Boost is enabled.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment