NVIDIA GPU Cloud TensorFlow

This test profile uses the NVIDIA GPU Cloud (NGC/nvcr.io) for running the TensorFlow image inside Docker for benchmarking. You must have already signed into NGC for this test profile to work.

This test profile requires you have already signed into the NVIDIA GPU Cloud (https://ngc.nvidia.com/) via Docker on this system/user (via docker login nvcr.io) in order for this test profile to work.

To run this test with the Phoronix Test Suite, the basic command is: phoronix-test-suite benchmark ngc-tensorflow.

Use with caution this test profile is currently marked Experimental.

Project Site

ngc.nvidia.com

Test Created

7 October 2018

Test Maintainer

Michael Larabel 

Test Type

Graphics

Average Install Time

2 Seconds

Average Run Time

4 Minutes, 6 Seconds

Accolades

20k+ Downloads

Supported Platforms


Public Result Uploads *Reported Installs **Reported Test Completions **Test Profile Page Views ***OpenBenchmarking.orgEventsNVIDIA GPU Cloud TensorFlow Popularity Statisticspts/ngc-tensorflow2018.102018.122019.022019.042019.062019.082019.102019.122020.022020.042020.062020.082020.102020.122021.022021.042021.062021.082021.102021.122022.022022.042022.062022.082004006008001000
* Uploading of benchmark result data to OpenBenchmarking.org is always optional (opt-in) via the Phoronix Test Suite for users wishing to share their results publicly.
** Data based on those opting to upload their test results to OpenBenchmarking.org and users enabling the opt-in anonymous statistics reporting while running benchmarks from an Internet-connected platform.
*** Test profile page view reporting began March 2021.
Data current as of 21 September 2022.
AlexNet, FP1613.9%VGG-16, FP329.2%ResNet-50, FP329.7%ResNet-50, FP1613.6%Googlenet, FP1612.9%Inception v4, FP1614.2%VGG-16, FP1614.2%AlexNet, FP3212.3%Test Option PopularityOpenBenchmarking.org

Revision History

pts/ngc-tensorflow-1.0.0   [View Source]   Sun, 07 Oct 2018 09:49:47 GMT
Initial commit of NVIDIA GPU Cloud TensorFlow Docker-based compute test.


Performance Metrics

Analyze Test Configuration:

NVIDIA GPU Cloud TensorFlow 18.09

Test: AlexNet, FP16

OpenBenchmarking.org metrics for this test profile configuration based on 55 public results since 7 October 2018 with the latest data as of 2 December 2019.

Below is an overview of the generalized performance for components where there is sufficient statistically significant data based upon user-uploaded results. It is important to keep in mind particularly in the Linux/open-source space there can be vastly different OS configurations, with this overview intended to offer just general guidance as to the performance expectations.

Component
Percentile Rank
# Compatible Public Results
Images Per Second (Average)
Mid-Tier
75th
< 4330
41st
4
1865 +/- 12
Low-Tier
25th
< 1566
18th
4
1556 +/- 10
OpenBenchmarking.orgDistribution Of Public Results - Test: AlexNet, FP1655 Results Range From 1045 To 5476 Images Per Second1045113412231312140114901579166817571846193520242113220222912380246925582647273628252914300330923181327033593448353736263715380438933982407141604249433844274516460546944783487249615050513952285317540654953691215

Based on OpenBenchmarking.org data, the selected test / test configuration (NVIDIA GPU Cloud TensorFlow 18.09 - Test: AlexNet, FP16) has an average run-time of 2 minutes. By default this test profile is set to run at least 3 times but may increase if the standard deviation exceeds pre-defined defaults or other calculations deem additional runs necessary for greater statistical accuracy of the result.

OpenBenchmarking.orgMinutesTime Required To Complete BenchmarkTest: AlexNet, FP16Run-Time246810Min: 1 / Avg: 1.3 / Max: 2

Tested CPU Architectures

This benchmark has been successfully tested on the below mentioned architectures. The CPU architectures listed is where successful OpenBenchmarking.org result uploads occurred, namely for helping to determine if a given test is compatible with various alternative CPU architectures.

CPU Architecture
Kernel Identifier
Verified On
Intel / AMD x86 64-bit
x86_64
(Many Processors)