Performance¶

DeepStream application is benchmarked across various NVIDIA TAO Toolkit and open source models. The measured performance represents end-to-end performance of the entire video analytic application considering video capture and decode, pre-processing, batching, inference, and post-processing to generate metadata. The output rendering is turned off to achieve peak inference performance. For information on disabling the output rendering, see DeepStream Reference Application - deepstream-app chapter.

TAO Pre-trained models¶

TAO toolkit has a set of pretrained models listed in the table below. If the models below satisfy your requirement, you should start with one of them. These could be used for various applications in smart city or smart places. If your application is beyond the scope of these models, you may re-train one of the popular model architecture using TAO toolkit. The table below shows the end-to-end performance on highly accurate pre-trained models from TAO toolkit. All models are available on NGC. These models are natively integrated with DeepStream and the instructions to run these models are in /opt/nvidia/deepstream/deepstream-6.2/samples/configs/tao_pretrained_models/. The following numbers are obtained with sample_1080p_h265.mp4.

Performance jetson- pretrained models¶

Jetson

Orin

Jetson

Xavier

NX

Jetson

AGX

Xavier

Model Arch

Inference resolution

Precision

GPU

(FPS)

DLA1 (FPS)

DLA2 (FPS)

GPU

(FPS)

DLA1 (FPS)

DLA2 (FPS)

GPU (FPS)

DLA1 (FPS)

DLA2 (FPS)

PeopleNet- ResNet34

960x544

INT8

455

NA

NA

79

22

22

137

28

28

TrafficCamNet – ResNet18 License Plate Detection License Plate Recognition

960x544 640x480 96x48

INT8

379

NA

NA

79

NA

NA

139

NA

NA

TrafficCamNet – ResNet18

960x544

INT8

1112

NA

NA

287

84

84

488

155

155

DashCamNet – ResNet18

960x544

INT8

1120

NA

NA

275

75

75

466

119

119

FaceDetectIR- ResNet18

384x240

INT8

1142

NA

NA

1178

357

357

2018

613

613

All the models in the table above can run solely on DLA. This saves valuable GPU resources to run more complex models.

Note

Running inference simultaneously on multiple models is not supported on the DLA. You can only run one model at a time on the DLA.
NA : Not available for Jetson

Performance dgpu- pretrained models¶
			T4	A100 PCIe	A30	A2	A10
Model Arch	Inference resolution	Precision	GPU (FPS)	GPU (FPS)	GPU (FPS)	GPU (FPS)	GPU (FPS)
PeopleNet- ResNet34	960x544	INT8	431	2380	1466	237	993
TrafficCamNet – ResNet18 License Plate Detection License Plate recognition	960x544 640x480 96x48	INT8	481	2196	1356	310	1169
TrafficCamNet – ResNet18	960x544	INT8	1362	5202	3657	888	2481
DashCamNet – ResNet18	960x544	INT8	1325	5281	3630	852	2382
FaceDetectIR- ResNet18	384x240	INT8	2521	5655	5621	3061	3122

Note

NA : Not available

DeepStream reference model and tracker¶

DeepStream SDK ships with a reference DetectNet_v2-ResNet10 model and three ResNet18 classifier models. The detailed instructions to run these models with DeepStream are provided in the next section. DeepStream provides three reference trackers: IOU, NvDCF and DeepSORT. For more information about trackers, See the Gst-nvtracker section.

Configuration File Settings for Performance Measurement¶

To achieve peak performance, make sure the devices are properly cooled. For Turing and Ampere GPUs, make sure you use a server that meets the thermal and airflow requirements. Along with the hardware setup, a few other options in the config file need to be set to achieve the published performance. Make the required changes to one of the config files from DeepStream SDK to replicate the peak performance.

Turn off output rendering, OSD, and tiler

OSD (on-screen display) is used to display bounding box, masks, and labels on the screen. If output rendering is disabled, creating bounding boxes is not required unless the output needs to be streamed over RTSP or saved to disk. Tiler is used to display the output in NxM tiled grid. It is not needed if rendering is disabled. Output rendering, OSD and tiler use some percentage of compute resources, so it can reduce the inference performance.

To disable OSD, tiled display and output sink, make the following changes in the DeepStream config file.
To disable OSD, change enable to 0
[osd]
enable=0
To disable tiling, change enable to 0
[tiled-display]
enable=0
To turn-off output rendering, change the sink to fakesink.
[sink0]
enable=1
#Type - 1=FakeSink 2=EglSink 3=File
type=1
sync=0

Use the max_perf setting for tracker

DeepStream SDK 6.2 introduces a new reference low-level tracker library, NvMultiObjectTracker, along with a set of configuration files:

config_tracker_IOU.yml

config_tracker_NvDCF_max_perf.yml

config_tracker_NvDCF_perf.yml

config_tracker_NvDCF_accuracy.yml

To achieve the peak performance shown in the table above when using the NvDCF tracker, make sure the max_perf configuration is used with video frame resolution matched to that of the inference module. If the inference module uses 480x272 resolution, for example, it would be recommended to use a reduced resolution (e.g., 480x288) for the tracker module like the following:

[tracker]
enable=1
tracker-width=480
tracker-height=288
ll-lib-file=/opt/nvidia/deepstream/deepstream/lib/libnvds_nvmultiobjecttracker.so
#ll-config-file=/opt/nvidia/deepstream/deepstream/samples/configs/deepstream-app/config_tracker_IOU.yml
ll-config-file=/opt/nvidia/deepstream/deepstream/samples/configs/deepstream-app/config_tracker_NvDCF_max_perf.yml
gpu-id=0
enable-batch-process=1
display-tracking-id=1

When the IOU tracker is used, the video frame resolution doesn’t matter, and the default config_tracker_IOU.yml can be used.

To use DLA on Jetson AGX Xavier and Xavier NX for performance measurement, please refer to “Using DLA for inference” section in the Quickstart Guide.

DeepStream reference model¶

Data center GPU - GA100¶

This section describes configuration and settings for the DeepStream SDK on NVIDIA Data center GPU - GA100.

System Configuration¶

The system configuration for the DeepStream SDK is listed below:

GA100 System configuration¶

System Configuration

Specification

CPU

AMD EPYC 7742 @ 2.25GHz 3.4GHz Turbo (Rome) HT Off

GPU

A100-PCIE-40GB(GA100) 1*40537 MiB 1*108 SM

Ubuntu

Ubuntu 20.04

GPU Driver

525.60.13

CUDA

11.8

TensorRT

8.5.1.7

GPU clock frequency

1410 MHz

Application Configuration¶

Config file: source4_1080p_dec_infer-resnet_tracker_sgie_tiled_display_int8.txt

Change the following items in the config file:

The inference resolution of Primary GIE is specified in the samples/models/Primary_detector/resnet10.prototxt. In this file, change the dim (i.e. height and width of input tensor) from 368x640 to 272x480.
Change batch size under streammux and primary-gie to match the number of streams.
Disable tiled display and rendering using instructions above.
Enable IoU tracker.

The application configuration for the DeepStream SDK is listed below:

GA100 application configuration¶

Application Configuration

Specification

N×1080p 30 fps stream

sample_1080p_h265.mp4 (provided with the SDK) N=187

sample_1080p_h264.mp4 (provided with the SDK) N=98

Primary GIE

Resnet10 (480×272)

Batch Size = N

Interval=0

Tracker

Enabled. Processing at 480×272 resolution, IOU tracker enabled.

3 × Secondary GIEs

All batches size 32. Asynchronous mode enabled.

Secondary_VehicleTypes (224×224—Resnet18)

Secondary_CarColor (224×224—Resnet18)

Secondary_CarMake (224×224—Resnet18)

Tiled Display

Disabled

Rendering

Disabled

Achieved Performance The table below shows the achieved performance of the DeepStream SDK under the specified system and application configuration:

Stream type	No. of Stream @ 30 FPS	CPU Utilization	GPU Utilization
H.265	187	4.16%	49.85%
H.264	98	2.16%	22.1%

Data center GPU - T4¶

This section describes configuration and settings for the DeepStream SDK on NVIDIA Data center GPU - T4.

System Configuration¶

The system configuration for the DeepStream SDK is listed below:

T4 System configuration¶

System Configuration

Specification

CPU

Dual Intel® Xeon® CPU E5-2650 v4 @ 2.20GHz (48 threads total)

GPU

Tesla T4*

System Memory

360448Mb (22x16384) DDR42666, 2400MHz

Ubuntu

Ubuntu 20.04

GPU Driver

525.60.13

CUDA

11.8

TensorRT

8.5.1.7

GPU clock frequency

1513 MHz

Application Configuration¶

Config file: source4_1080p_dec_infer-resnet_tracker_sgie_tiled_display_int8.txt

Change the following items in the config file:

The inference resolution of Primary GIE is specified in the samples/models/Primary_detector/resnet10.prototxt. In this file, change the dim (i.e. height and width of input tensor) from 368x640 to 272x480.
Change batch size under streammux and primary-gie to match the number of streams.
Disable tiled display and rendering using instructions above.
Enable IoU tracker.

The application configuration for the DeepStream SDK is listed below:

T4 application configuration¶

Application Configuration

Specification

N×1080p 30 fps stream

sample_1080p_h265.mp4 (provided with the SDK) N=75

sample_1080p_h264.mp4 (provided with the SDK) N=43

Primary GIE

Resnet10 (480×272)

Batch Size = N

Interval=0

Tracker

Enabled. Processing at 480×272 resolution, IOU tracker enabled.

3 × Secondary GIEs

All batches size 32. Asynchronous mode enabled.

Secondary_VehicleTypes (224×224—Resnet18)

Secondary_CarColor (224×224—Resnet18)

Secondary_CarMake (224×224—Resnet18)

Tiled Display

Disabled

Rendering

Disabled

Achieved Performance The table below shows the achieved performance of the DeepStream SDK under the specified system and application configuration:

Stream type	No. of Stream @ 30 FPS	CPU Utilization	GPU Utilization
H.265	75	6.34%	57.94%
H.264	43	3.99%	28.65%

Jetson¶

This section describes configuration and settings for the DeepStream SDK on NVIDIA Jetson™ platforms. JetPack X [TBD] is used for software installation.

System Configuration¶

For the performance test:

Max power mode is enabled: $ sudo nvpmodel -m 0. For Jetson NX, use $ sudo nvpmodel -m 8
The GPU clocks are stepped to maximum: $ sudo jetson_clocks

For information about supported power modes, see the “Supported Modes and Power Efficiency” section in the power management topics of NVIDIA Tegra Linux Driver Package Development Guide, e.g., “Power Management for Jetson AGX Xavier Devices.”

Jetson AGX Xavier¶

Config file: source4_1080p_dec_infer-resnet_tracker_sgie_tiled_display_int8.txt

Change the following items in the config file:

The inference resolution of Primary GIE is specified in the samples/models/Primary_detector/resnet10.prototxt. In this file, change the dim (i.e. height and width of input tensor) from 368x640 to 272x480.
Change batch size under streammux and primary-gie to match the number of streams.
Disable tiled display and rendering using instructions above.
Enable IOU tracker.

The following tables describe performance results for the NVIDIA Jetson AGX Xavier™.

Jetson Nano Pipeline Configuration (deepstream-app)¶

Application Configuration

Specification

N×1080p 30 fps streams

sample_1080p_h265.mp4 (provided with the SDK) N=46

sample_1080p_h264.mp4 (provided with the SDK) N=34

Primary GIE

Resnet10 (480×272) Asynchronous mode enabled

Batch Size = N

Interval = 0

Tracker

Enabled; processing at 480×272 resolution, IOU tracker enabled.

3× secondary GIEs

All batches are size 32.

Secondary_VehicleTypes (224×224—Resnet18)

Secondary_CarColor (224×224—Resnet18)

Secondary_CarMake (224×224—Resnet18)

OSD/tiled display

Disabled

Renderer

Disabled

Achieved Performance

Stream type	No. of Stream @ 30 FPS	CPU Utilization	GPU Utilization
H.265	46	53.95%	93.88%
H.264	34	40.70%	72.05%

Jetson NX¶

Config file: source4_1080p_dec_infer-resnet_tracker_sgie_tiled_display_int8.txt

Change the following items in the config file:

The inference resolution of Primary GIE is specified in the samples/models/Primary_detector/resnet10.prototxt. In this file, change the dim (i.e. height and width of input tensor) from 368x640 to 272x480.
Change batch size under streammux and primary-gie to match the number of streams.
Disable tiled display and rendering using instructions above.
Enable IOU tracker.

The following tables describe performance results for the NVIDIA Jetson NX™.

Jetson NX Pipeline Configuration (`deepstream-app`)¶
Application Configuration	Specification
N×1080p 30 fps streams	`sample_1080p_h265.mp4` (provided with the SDK) N=27 `sample_1080p_h264.mp4` (provided with the SDK) N=22
Primary GIE	Resnet10 (480×272) Asynchronous mode enabled Batch Size = N Interval = 0
Tracker	Enabled; processing at 480×272 resolution, IOU tracker enabled.
3× secondary GIEs	All batches are size 32. Secondary_VehicleTypes (224×224—Resnet18) Secondary_CarColor (224×224—Resnet18) Secondary_CarMake (224×224—Resnet18)
OSD/tiled display	Disabled
Renderer	Disabled

Achieved Performance

Stream type	No. of Stream @ 30 FPS	CPU Utilization	GPU Utilization
H.265	27	68.04%	93.85%
H.264	22	59.33%	83.03%

Jetson Orin¶

Config file: source4_1080p_dec_infer-resnet_tracker_sgie_tiled_display_int8.txt

Change the following items in the config file:

The inference resolution of Primary GIE is specified in the samples/models/Primary_detector/resnet10.prototxt. In this file, change the dim (i.e. height and width of input tensor) from 368x640 to 272x480.
Change batch size under streammux and primary-gie to match the number of streams.
Disable tiled display and rendering using instructions above.
Enable IOU tracker.

The following tables describe performance results for the NVIDIA Jetson Orin™.

Jetson Orin Pipeline Configuration (`deepstream-app`)¶
Application Configuration	Specification
N×1080p 30 fps streams	`sample_1080p_h265.mp4` (provided with the SDK) N=37 `sample_1080p_h264.mp4` (provided with the SDK) N=15
Primary GIE	Resnet10 (480×272) Asynchronous mode enabled Batch Size = N Interval = 0
Tracker	Enabled; processing at 480×272 resolution, IOU tracker enabled.
3× secondary GIEs	All batches are size 32. Secondary_VehicleTypes (224×224—Resnet18) Secondary_CarColor (224×224—Resnet18) Secondary_CarMake (224×224—Resnet18)
OSD/tiled display	Disabled
Renderer	Disabled

Achieved Performance

Stream type	No. of Stream @ 30 FPS	CPU Utilization	GPU Utilization
H.265	37	16.14%	31%
H.264	15	10.55%	26.25%