CUDA ISP for NVIDIA Jetson/Performance: Difference between revisions

From RidgeRun Developer Wiki
(Blanked the page)
Tag: Blanking
 
(73 intermediate revisions by 3 users not shown)
Line 1: Line 1:
<noinclude>
{{CUDA ISP for NVIDIA Jetson/Head|previous=|next=|metakeywords=|metadescription=}}
</noinclude>


{{DISPLAYTITLE:CUDA ISP for NVIDIA Jetson: Performance|noerror}}
= GStreamer elements performance =
To measure the performance, we have used two of our GStreamer tools: [[GstShark]] and [https://github.com/RidgeRun/gst-perf GstPerf].
For testing purposes, take into account the following points:
* Maximun performance mode enabled: all cores and Jetson clocks enabled.
* Jetpack 4.6
* FPS is equal to 1/processing time
== '''Jetson Xavier NX''' ==
In the following sections you will see the performance for each of the elements.
=== '''cudashift element''' ===
The following pipeline was used to measure the processing time and FPS for the cudashift element with an input image with 4K resolution coming from a camera sensor. 
<source lang=bash>
GST_DEBUG="GST_TRACER:7" GST_TRACERS="proctime" gst-launch-1.0 -ve v4l2src io-mode=userptr ! 'video/x-bayer, bpp=10, format=rggb' ! cudashift shift=5 ! fakesink
</source>
The results obtained:
<center>
{| class="wikitable"
|-
! Measurement (Average) !! Jetson Xavier NX
|-
| FPS || 396
|-
| Processing time (seconds) || 0.002522
|}
</center>
<br>
<br><br>
[[File:Shift4kproctimenx.png|1000px|frameless|center|CUDA ISP library ]]
<br>
<br><br>
=== '''cudadebayer element''' ===
'''RGB Output'''
The following pipeline was used to measure the processing time and FPS for the cudadebayer element with an input image with 4K resolution coming from a camera sensor to an RGB output image.
<source lang=bash>
GST_DEBUG="GST_TRACER:7" GST_TRACERS="proctime" gst-launch-1.0 -ve v4l2src io-mode=userptr ! 'video/x-bayer, bpp=10, width=3840, height=2160' ! cudadebayer ! fakesink
</source>
The results obtained:
<center>
{| class="wikitable"
|-
! Measurement (Average) !! Jetson Xavier NX
|-
| FPS || 228
|-
| Processing time (seconds) || 0.004389
|}
</center>
<br>
<br><br>
[[File:Debayer4krgbproctimenx.png|1000px|frameless|center|CUDA ISP library ]]
<br>
<br><br>
'''I420 Output'''
The following pipeline was used to measure the processing time and FPS for the cudadebayer element with an input image with 4K resolution coming from a camera sensor to an I420 output image.
<source lang=bash>
GST_DEBUG="GST_TRACER:7" GST_TRACERS="proctime" gst-launch-1.0 -ve v4l2src io-mode=userptr ! 'video/x-bayer, bpp=10, width=3840, height=2160' ! cudadebayer ! fakesink
</source>
The results obtained:
<center>
{| class="wikitable"
|-
! Measurement (Average) !! Jetson Xavier NX
|-
| FPS || 187
|-
| Processing time (seconds) || 0.005353
|}
</center>
<br>
<br><br>
[[File:Debayer4ki420proctimenx.png|1000px|frameless|center|CUDA ISP library ]]
<br>
<br><br>
=== '''cudaawb element''' ===
'''RGB Output'''
The following pipeline was used to measure the processing time and FPS for the cudaawb element with an input image with 4K resolution coming from a camera sensor to an RGB output image.
<source lang=bash>
GST_DEBUG="GST_TRACER:7" GST_TRACERS="proctime" gst-launch-1.0 -ve v4l2src io-mode=userptr ! 'video/x-bayer, bpp=10, width=3840, height=2160' ! cudadebayer ! cudaawb ! fakesink
</source>
The results obtained:
<center>
{| class="wikitable"
|-
! Measurement (Average) !! Jetson Xavier NX
|-
| FPS || 370
|-
| Processing time (seconds) || 0.002698
|}
</center>
<br>
<br><br>
[[File:Awb4krgbproctimenx.png|1000px|frameless|center|CUDA ISP library ]]
<br>
<br><br>
'''I420 Output'''
The following pipeline was used to measure the processing time and FPS for the cudaawb element with an input image with 4K resolution coming from a camera sensor to an I420 output image.
<source lang=bash>
GST_DEBUG="GST_TRACER:7" GST_TRACERS="proctime" gst-launch-1.0 -ve v4l2src io-mode=userptr ! 'video/x-bayer, bpp=10, width=3840, height=2160' ! cudadebayer ! cudaawb ! 'video/x-raw, format=I420' ! fakesink
</source>
The results obtained:
<center>
{| class="wikitable"
|-
! Measurement (Average) !! Jetson Xavier NX
|-
| FPS || 202
|-
| Processing time (seconds) || 0.004952
|}
</center>
<br>
<br><br>
[[File:Awb4ki420proctimenx.png|1000px|frameless|center|CUDA ISP library ]]
<br>
<br><br>
== '''Jetson Xavier AGX''' ==
In the following sections you will see the performance the elements.
=== cudadebayer ===
'''RGB Output'''
The following pipeline was used to measure the processing time and FPS for the cudadebayer element with an input image with 1920x1200 resolution coming from a camera sensor to an RGB output image.
<source lang=bash>
GST_DEBUG="GST_TRACER:7" GST_TRACERS="proctime" gst-launch-1.0 -ve rrv4l2src io-mode=userptr ! 'video/x-bayer, bpp=10, width=1920, height=1200, format=grbg' ! cudadebayer ! 'video/x-raw, format=RGB' ! fakesink
</source>
The results obtained:
<center>
{| class="wikitable"
|-
! Measurement (Average) !! Jetson Xavier AGX
|-
| FPS || 539
|-
| Processing time (seconds) || 0.001854
|}
</center>
<br>
<br><br>
[[File:Debayer1920x1200rgbproctimev1.png|1000px|frameless|center|CUDA ISP library ]]
<br>
<br><br>
'''I420 Output'''
The following pipeline was used to measure the processing time and FPS for the cudadebayer element with an input image with 1920x1200 resolution coming from a camera sensor to an I420 output image.
<source lang=bash>
GST_DEBUG="GST_TRACER:7" GST_TRACERS="proctime" gst-launch-1.0 -ve rrv4l2src io-mode=userptr ! 'video/x-bayer, bpp=10, width=1920, height=1200, format=grbg' ! cudadebayer ! 'video/x-raw, format=I420' ! fakesink
</source>
The results obtained:
<center>
{| class="wikitable"
|-
! Measurement (Average) !! Jetson Xavier AGX
|-
| FPS || 458
|-
| Processing time (seconds) || 0.002183
|}
</center>
<br>
<br><br>
[[File:Debayer1920x1200i420proctime.png|1000px|frameless|center|CUDA ISP library ]]
<br>
<br><br>
=== cudaawb ===
'''RGB Output'''
The following pipeline was used to measure the processing time and FPS for the cudaawb element with an input image with 1920x1200 resolution coming from a camera sensor to an RGB output image.
<source lang=bash>
GST_DEBUG="GST_TRACER:7" GST_TRACERS="proctime" gst-launch-1.0 -ve rrv4l2src io-mode=userptr ! 'video/x-bayer, bpp=10, width=1920, height=1200, format=grbg' ! cudadebayer ! cudaawb ! 'video/x-raw, format=RGB' ! fakesink
</source>
The results obtained:
<center>
{| class="wikitable"
|-
! Measurement (Average) !! Jetson Xavier NX
|-
| FPS || 752
|-
| Processing time (seconds) || 0.001329
|}
</center>
<br>
<br><br>
[[File:Debayer1920x1200rgbproctime.png|1000px|frameless|center|CUDA ISP library ]]
<br>
<br><br>
'''I420 Output'''
The following pipeline was used to measure the processing time and FPS for the cudaawb element with an input image with 1920x1200 resolution coming from a camera sensor to an I420 output image.
<source lang=bash>
GST_DEBUG="GST_TRACER:7" GST_TRACERS="proctime" gst-launch-1.0 -ve rrv4l2src io-mode=userptr ! 'video/x-bayer, bpp=10, width=1920, height=1200, format=grbg' ! cudadebayer ! cudaawb ! 'video/x-raw, format=I420' ! fakesink
</source>
The results obtained:
<center>
{| class="wikitable"
|-
! Measurement (Average) !! Jetson Xavier NX
|-
| FPS || 473
|-
| Processing time (seconds) || 0.002111
|}
</center>
<br>
<br><br>
[[File:Awb1920x1200i420proctimeagx.png|1000px|frameless|center|CUDA ISP library ]]
<br>
<br><br>
== More cameras ==
This section shows the performance results for the elements running at the same time on more than one camera on a Jetson XavierAGX. For all the tests done with an RGB output image, the following pipeline was used to measure the processing time and FPS for the cudaawb and the cudadebayer element with an input image with 1920x1200 resolution coming from multiple camera sensor.
<source lang=bash>
GST_DEBUG="GST_TRACER:7" GST_TRACERS="proctime" gst-launch-1.0 -ve rrv4l2src device=/dev/video0 io-mode=userptr ! 'video/x-bayer, bpp=10, width=1920, height=1200, format=grbg' ! cudadebayer ! cudaawb ! 'video/x-raw, format=RGB' ! fakesink
</source>
The same way, for all the test done with an I420 output image. the following pipeline was used to measure the processing time and FPS for the cudaawb and the cudadebayer element with an input image with 1920x1200 resolution coming from multiple camera sensor
<source lang=bash>
GST_DEBUG="GST_TRACER:7" GST_TRACERS="proctime" gst-launch-1.0 -ve rrv4l2src device=/dev/video1 io-mode=userptr ! 'video/x-bayer, bpp=10, width=1920, height=1200, format=grbg' ! cudadebayer ! cudaawb ! 'video/x-raw, format=I420' ! fakesink
</source>
=== Two cameras ===
'''RGB Output'''
The results obtained:
<center>
{| class="wikitable"
|-
! Measurement (Average) !! cudadebayer !! cudaawb
|-
| FPS || 412 || 397
|-
| Processing time (seconds) || 0.002426 || 0.002521
|}
</center>
'''I420'''
The results obtained:
<center>
{| class="wikitable"
|-
! Measurement (Average) !! cudadebayer !! cudaawb
|-
| FPS || 464 || 429
|-
| Processing time (seconds) || 0.002154 || 0.002330
|}
</center>
=== Three cameras ===
'''RGB Output'''
The results obtained:
<center>
{| class="wikitable"
|-
! Measurement (Average) !! cudadebayer !! cudaawb
|-
| FPS || 429 || 374
|-
| Processing time (seconds) || 0.002354 || 0.002672
|}
</center>
'''I420'''
The results obtained:
<center>
{| class="wikitable"
|-
! Measurement (Average) !! cudadebayer !! cudaawb
|-
| FPS || 402 || 450
|-
| Processing time (seconds) || 0.002486 || 0.002220
|}
</center>
<noinclude>
{{CUDA ISP for NVIDIA Jetson/Foot||}}
</noinclude>

Latest revision as of 13:17, 28 March 2023

Retrieved from "https://developer.ridgerun.com/wiki/index.php?title=CUDA_ISP_for_NVIDIA_Jetson/Performance&oldid=48440"