TensorFlow Lite

This is a benchmark of the TensorFlow Lite implementation focused on TensorFlow machine learning for mobile, IoT, edge, and other cases. The current Linux support is limited to running on CPUs. This test profile is measuring the average inference time.

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

Project Site

tensorflow.org

Source Repository

github.com

Test Created

23 August 2020

Last Updated

19 May 2022

Test Maintainer

Michael Larabel 

Test Type

System

Average Install Time

8 Seconds

Average Run Time

4 Minutes, 17 Seconds

Accolades

10k+ Downloads

Supported Platforms


Public Result Uploads *Reported Installs **Reported Test Completions **Test Profile Page Views ***OpenBenchmarking.orgEventsTensorFlow Lite Popularity Statisticspts/tensorflow-lite2020.082020.092020.102020.112020.122021.012021.022021.032021.042021.052021.062021.072021.082021.092021.102021.112021.122022.012022.022022.032022.042022.052022.062022.072022.082022.092022.102022.112022.122023.012023.022023.032023.042023.052023.062023.072023.082023.095K10K15K20K25K
* 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 updated weekly as of 27 September 2023.
Mobilenet Float19.3%NASNet Mobile13.8%Mobilenet Quant16.8%SqueezeNet17.5%Inception ResNet V215.2%Inception V417.4%Model Option PopularityOpenBenchmarking.org

Revision History

pts/tensorflow-lite-1.1.0   [View Source]   Thu, 19 May 2022 09:57:39 GMT
Update against latest upstream nightly.

pts/tensorflow-lite-1.0.0   [View Source]   Sun, 23 Aug 2020 14:13:10 GMT
TensorFlow Lite initial commit.

Suites Using This Test

Machine Learning

HPC - High Performance Computing


Performance Metrics

Analyze Test Configuration:

TensorFlow Lite 2022-05-18

Model: Mobilenet Float

OpenBenchmarking.org metrics for this test profile configuration based on 617 public results since 19 May 2022 with the latest data as of 22 September 2023.

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
Microseconds (Average)
98th
10
1017 +/- 8
91st
4
1297 +/- 2
88th
9
1406 +/- 30
87th
7
1434 +/- 13
84th
3
1518 +/- 6
84th
3
1527 +/- 5
82nd
3
1606 +/- 113
79th
10
1706 +/- 70
76th
4
1796 +/- 6
Mid-Tier
75th
> 1805
74th
4
1818 +/- 10
74th
7
1823 +/- 142
71st
3
1901 +/- 39
66th
9
2171 +/- 81
65th
4
2216 +/- 18
63rd
3
2339 +/- 38
62nd
6
2345 +/- 7
61st
12
2396 +/- 38
56th
4
2482 +/- 108
55th
3
2542 +/- 10
52nd
4
2655 +/- 12
52nd
3
2692 +/- 53
51st
6
2759 +/- 23
Median
50th
2760
50th
3
2788 +/- 12
49th
22
2814 +/- 136
45th
5
2972 +/- 278
44th
4
3056 +/- 30
43rd
4
3156 +/- 77
42nd
7
3246 +/- 65
37th
16
3526 +/- 522
35th
9
3925 +/- 23
33rd
4
3953 +/- 15
32nd
6
3955 +/- 272
32nd
10
3958 +/- 66
31st
3
4091 +/- 50
31st
3
4110 +/- 99
29th
4
4245 +/- 23
29th
3
4257 +/- 8
26th
9
4541 +/- 219
Low-Tier
25th
> 4549
25th
7
4598 +/- 59
24th
3
4660 +/- 5
20th
3
4909 +/- 23
20th
17
4963 +/- 450
20th
4
4965 +/- 398
19th
4
5387 +/- 53
18th
3
5606 +/- 173
15th
5
5955 +/- 4
14th
3
6049 +/- 99
14th
6
6133 +/- 293
13th
4
6633 +/- 194
12th
3
6917 +/- 5
12th
6
6974 +/- 145
10th
4
7159 +/- 10
9th
4
8069 +/- 17
7th
4
9975 +/- 7
3rd
3
23400 +/- 1608
2nd
5
68651 +/- 3119
OpenBenchmarking.orgDistribution Of Public Results - Model: Mobilenet Float617 Results Range From 940 To 345091 Microseconds940782414708215922847635360422444912856012628966978076664835489043297316104200111084117968124852131736138620145504152388159272166156173040179924186808193692200576207460214344221228228112234996241880248764255648262532269416276300283184290068296952303836310720317604324488331372338256345140120240360480600

Based on OpenBenchmarking.org data, the selected test / test configuration (TensorFlow Lite 2022-05-18 - Model: Mobilenet Float) has an average run-time of 5 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 BenchmarkModel: Mobilenet FloatRun-Time48121620Min: 3 / Avg: 4.49 / Max: 15

Based on public OpenBenchmarking.org results, the selected test / test configuration has an average standard deviation of 0.5%.

OpenBenchmarking.orgPercent, Fewer Is BetterAverage Deviation Between RunsModel: Mobilenet FloatDeviation48121620Min: 0 / Avg: 0.47 / Max: 13

Does It Scale Well With Increasing Cores?

No, based on the automated analysis of the collected public benchmark data, this test / test settings does not generally scale well with increasing CPU core counts. Data based on publicly available results for this test / test settings, separated by vendor, result divided by the reference CPU clock speed, grouped by matching physical CPU core count, and normalized against the smallest core count tested from each vendor for each CPU having a sufficient number of test samples and statistically significant data.

IntelAMDOpenBenchmarking.orgRelative Core Scaling To BaseTensorFlow Lite CPU Core ScalingModel: Mobilenet Float2468101216641.352.74.055.46.75

Notable Instruction Set Usage

Notable instruction set extensions supported by this test, based on an automatic analysis by the Phoronix Test Suite / OpenBenchmarking.org analytics engine.

Instruction Set
Support
Instructions Detected
SSE2 (SSE2)
Used by default on supported hardware.
 
MOVDQA PMULUDQ PSHUFD PSRLDQ MOVD CVTSI2SD ADDSD MULSD SUBSD DIVSD MOVAPD CVTSS2SD CVTTSD2SI SQRTSD UCOMISD XORPD CVTSD2SS CVTTPS2DQ CVTDQ2PS PADDQ MOVDQU PUNPCKLQDQ UNPCKLPD CVTDQ2PD MULPD CVTPD2PS ANDPD MAXSD PSUBQ CVTPS2PD MINSD MOVUPD UNPCKHPD ADDPD PUNPCKHQDQ CVTPS2DQ
Used by default on supported hardware.
Found on Intel processors since Sandy Bridge (2011).
Found on AMD processors since Bulldozer (2011).

 
VBROADCASTF128 VZEROUPPER VMASKMOVPS VEXTRACTF128 VBROADCASTSS VINSERTF128 VPERMILPS
Advanced Vector Extensions 512 (AVX512)
Used by default on supported hardware.
 
(ZMM REGISTER USE)
FMA (FMA)
Used by default on supported hardware.
Found on Intel processors since Haswell (2013).
Found on AMD processors since Bulldozer (2011).

 
VFMADD132PS VFMADD231PS VFMADD213PS
Last automated analysis: 18 January 2022

This test profile binary relies on the shared libraries libm.so.6, libpthread.so.0, libdl.so.2, librt.so.1, libc.so.6.

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)
ARMv8 64-bit
aarch64
ARMv8 9-Core, ARMv8 Cortex-A53 2-Core, ARMv8 Cortex-A53 4-Core, ARMv8 Cortex-A72, ARMv8 Cortex-A72 4-Core, ARMv8 Cortex-A78E 6-Core, ARMv8 Neoverse-N1, ARMv8 Neoverse-V1, Ampere ARMv8 Neoverse-N1 160-Core, Apple, Apple M1, Apple M2

Recent Test Results

OpenBenchmarking.org Results Compare

1 System - 33 Benchmark Results

AMD Ryzen 7 3800X 8-Core - ASRock B550 Phantom Gaming 4 - AMD Starship

EndeavourOS rolling - 6.4.12-1-cachyos-tt - Xfce 4.18

1 System - 33 Benchmark Results

AMD Ryzen 7 3800X 8-Core - ASRock B550 Phantom Gaming 4 - AMD Starship

EndeavourOS rolling - 6.4.12-273-tkg-tt - Xfce 4.18

1 System - 33 Benchmark Results

AMD Ryzen 7 3800X 8-Core - ASRock B550 Phantom Gaming 4 - AMD Starship

EndeavourOS rolling - 6.4.12-273-tkg-bore - Xfce 4.18

1 System - 33 Benchmark Results

AMD Ryzen 7 3800X 8-Core - ASRock B550 Phantom Gaming 4 - AMD Starship

EndeavourOS rolling - 6.4.12-1-cachyos-pds - Xfce 4.18

1 System - 33 Benchmark Results

AMD Ryzen 7 3800X 8-Core - ASRock B550 Phantom Gaming 4 - AMD Starship

EndeavourOS rolling - 6.4.12-273-tkg-pds - Xfce 4.18

1 System - 195 Benchmark Results

1 System - 195 Benchmark Results

1 System - 33 Benchmark Results

AMD Ryzen 7 3800X 8-Core - ASRock B550 Phantom Gaming 4 - AMD Starship

EndeavourOS rolling - 6.4.12-1-cachyos-bore - Xfce 4.18

1 System - 33 Benchmark Results

AMD Ryzen 7 3800X 8-Core - ASRock B550 Phantom Gaming 4 - AMD Starship

EndeavourOS rolling - 6.4.12-273-tkg-eevdf - Xfce 4.18

1 System - 33 Benchmark Results

AMD Ryzen 7 3800X 8-Core - ASRock B550 Phantom Gaming 4 - AMD Starship

EndeavourOS rolling - 6.4.12-1-cachyos-eevdf - Xfce 4.18

1 System - 33 Benchmark Results

AMD Ryzen 7 3800X 8-Core - ASRock B550 Phantom Gaming 4 - AMD Starship

EndeavourOS rolling - 6.4.12-273-tkg-bmq - Xfce 4.18

1 System - 33 Benchmark Results

AMD Ryzen 7 3800X 8-Core - ASRock B550 Phantom Gaming 4 - AMD Starship

EndeavourOS rolling - 6.4.12-1-cachyos-bmq - Xfce 4.18

1 System - 33 Benchmark Results

AMD Ryzen 7 3800X 8-Core - ASRock B550 Phantom Gaming 4 - AMD Starship

EndeavourOS rolling - 6.4.12-1-cachyos-cfs - Xfce 4.18

1 System - 33 Benchmark Results

AMD Ryzen 7 3800X 8-Core - ASRock B550 Phantom Gaming 4 - AMD Starship

EndeavourOS rolling - 6.4.12-273-tkg-cfs - Xfce 4.18

1 System - 33 Benchmark Results

AMD Ryzen 7 3800X 8-Core - ASRock B550 Phantom Gaming 4 - AMD Starship

EndeavourOS rolling - 6.4.12-arch1-1 - Xfce 4.18

Most Popular Test Results

OpenBenchmarking.org Results Compare

5 Systems - 94 Benchmark Results

AMD EPYC 7R13 - Amazon EC2 c6a.4xlarge - Intel 440FX 82441FX PMC

Ubuntu 22.04 - 5.15.0-1004-aws - GCC 11.2.0

13 Systems - 333 Benchmark Results

Intel Core i9-12900K - ASUS ROG STRIX Z690-E GAMING WIFI - Intel Device 7aa7

Ubuntu 22.04 - 6.0.0-060000rc1daily20220820-generic - GNOME Shell 42.2

7 Systems - 293 Benchmark Results

AMD Ryzen 7 4700U - LENOVO LNVNB161216 - AMD Renoir

Pop 22.04 - 5.17.5-76051705-generic - GNOME Shell 42.1

4 Systems - 171 Benchmark Results

AMD Ryzen 5 5600G - Gigabyte A520M DS3H - AMD Renoir

Ubuntu 22.04 - 5.15.0-41-generic - GNOME Shell 42.2

2 Systems - 219 Benchmark Results

AMD Ryzen 9 5900HX - ASUS G513QY v1.0 - AMD Renoir

Arch rolling - 5.18.16-arch1-1 - KDE Plasma 5.25.4

2 Systems - 86 Benchmark Results

Intel Xeon Silver 4216 - TYAN S7100AG2NR - Intel Sky Lake-E DMI3 Registers

Debian 11 - 5.10.0-10-amd64 - X Server

16 Systems - 333 Benchmark Results

Intel Core i9-12900K - ASUS ROG STRIX Z690-E GAMING WIFI - Intel Device 7aa7

Ubuntu 22.04 - 6.0.0-060000rc1daily20220820-generic - GNOME Shell 42.2

3 Systems - 190 Benchmark Results

AMD Ryzen 7 PRO 6850U - LENOVO 21CM0001US - AMD Device 14b5

Arch rolling - 5.18.16-arch1-1 - KDE Plasma 5.25.4

3 Systems - 18 Benchmark Results

AMD Ryzen Threadripper 3990X 64-Core - Gigabyte TRX40 AORUS PRO WIFI - AMD Starship

Pop 22.04 - 5.17.5-76051705-generic - GNOME Shell 42.0

2 Systems - 218 Benchmark Results

Intel Core i7-1280P - MSI MS-14C6 - Intel Alder Lake PCH

Arch Linux - 5.18.16-arch1-1 - KDE Plasma 5.25.4

Find More Test Results