intoPIX lance les SDK FastTICO-XS et FastTICO-RAW pour le GPU Nvidia Jetson

09.03.21 12:33 PM By Justine

Nvidia Jetson is a series of embedded computing boards from Nvidia. Jetson is a low-power system and is designed for accelerating machine learning applications at a lower price and with a smaller shape and a very fast GPU.

The applications that need that kind of processing power are often related to image/video capture and processing. Our Nvidia Jetson GPU SDK brings the compression to this platform.

The benefits of the intoPIX Jetson SDKs are multiple 

The intoPIX Jetson SDKs use Nvidia GPUs to encode and decode 4:2:2, 4:4:4 and RAW Bayer images at a much greater speed. Depending on the GPU, the decoder/encoder may be faster. Our FastTICO-XS and FastTICO-RAW SDKs for Nvidia GPU have a really higher performance in terms of frame rate that can be reached. Offering extremely low latency and high-quality processing, the intoPIX SDKs support resolutions up to 8192 x 4096 pixels and real-time operations. 


With this solution, the entire encoding/decoding is offloaded to the GPU, so the CPU remains free to be used for other tasks at the same time. If the user has a processing pipeline where other stages are also performed on the GPU, it is beneficial to use the GPU library. For example, the user could have a setup where a JPEG XS image is decoded using this decoder, then some filter is applied to it on the GPU, and finally it is displayed on the screen. With the CPU library, the system would need to copy each decoded image from the GPU to the CPU, which would take time, and make it slower.


« Having the GPU library on a Jetson is useful because that way the user has a separate piece of hardware dedicated to decoding or encoding JPEG XS. It's not necessary anymore to install it on its own system, worry about compatibility, or have it interfere with other tasks running on its computer » explains Justine Hecq, Product Owner of the intoPIX TICO-XS solutions. It is much cheaper to buy a GPU than to buy a CPU of equivalent speed and it also consumes less energy. So this is really a cost-effective way to develop an application ».

The Nvidia Jetson TX2 and AGX Xavier performance benchmarks for HD & 4K 
(v1.2.5 - Q3.2020)

 PerformancesNvidia Jetson TX2  Nvidia AGX Xavier
 FastTICO-XS - HD
1 stream (1080p50/60)
Up to 2 streams (720p50/60 or 1080i50/60)
Up to 6 streams (1080p50/60) 
Up to 12 streams (720p50/60 or 1080i50/60)
FastTICO-XS - 4K
1 stream (2160p24/25/30)
Up to  4 streams (2160p24/25/30)
Up to 2 streams (2160p50/60)
FastTICO-RAW - HDUp to 110 fpsUp to 400 fps
FastTICO-RAW - 4K Up to 40 fpsUp to 130 fps

intoPIX FastTICO-XS and FastTICO-RAW SDKs are gaining in performance continuously. 

Please contact us to get access the latest updates performances. us to get access the latest updates performances.

CONTACTEZ-NOUS

 Contenus associés

Product
Learn more about the intoPIX FastTICO-XS SDKs.
En savoir plus
Product
Learn more about the intoPIX FastTICO-RAW SDKs.
En savoir plus
Article
​​High quality live production in the LAN, over the WAN and into the CLOUD using JPEG XS...
En savoir plus