Difference between revisions of "CUDA ISP for NVIDIA Jetson/Performance/Library"

From RidgeRun Developer Connection
Jump to: navigation, search
Line 104: Line 104:
 
|- style="text-align:right;"
 
|- style="text-align:right;"
 
| style="text-align:left; font-weight:bold;" | GPU usage (%)
 
| style="text-align:left; font-weight:bold;" | GPU usage (%)
| style="background-color:#ffd6a5;" |  
+
| style="background-color:#ffd6a5;" | 20.68
| style="background-color:#ffd6a5;" |  
+
| style="background-color:#ffd6a5;" | 27.06
 
| style="background-color:#ffadad;" |  
 
| style="background-color:#ffadad;" |  
 
| style="background-color:#ffadad;" |  
 
| style="background-color:#ffadad;" |  

Revision as of 10:37, 29 March 2023


  Index  






Library API performance

To measure the CUDA ISP API performance, we built a simple example (not included in the production code) that iterates over the apply methods and records performance metrics for each iteration. We recorded the duration of each apply method separately. We then recorded the CPU and GPU usage, as well as the CPU RAM and GPU RAM usage for the complete processing pipeline running at 30 fps. We recorded the performance statistics for 1080p and 4K buffers. We recorded the performance on a Jetson Nano, Jetson Xavier NX, Jetson Xavier AGX, and Jetson Orin.

The following table summarizes CUDA ISP's performance results:

Platform Jetson Orin Jetson Xavier AGX Jetson Xavier NX Jetson Nano
Buffer size 1080p 4K 1080p 4K 1080p 4K 1080p 4K
Processing time by algorithm (microseconds)
CudaShift 60 51 135 131 93 93 135 147
CudaDebayer 22 20 48 39 39 31 53 55
CudaWhiteBalancer 4056 5966 4844 8091 1360 4249 5071 18903
CudaColorSpaceConverter 20 17 45 52 35 34 55 57
Resource consumption profile
CPU usage (%) 0.211111 0.129419 0.491435 0.458062 0.523657 0.477216 0.836478 0.819940
CPU RAM (kB) 160295 157636 173613 173477 173539 171987 146295 147580
GPU usage (%) 20.68 27.06 5.48 17.91 25.12 94.6
GPU RAM (kB) 86700 135899 105247 107641 100387 106288 91733 116833



  Index