Zlib slow on TX2

Hi, I am running a ROS node on TX2 which uses zlib. I add the following line to ROS package CMakefile to include the zlib:

target_link_libraries(TrackRun yaml-cpp uvc jetson-utils z ${catkin_LIBRARIES))

I found that to compress an 640K data block, it took 140ms on TX2. I need something much faster than that, like 25ms. Any one know a special zlib build for TX2 which would utilize the computation resource on the device, such as GPU?

Hi AutoCar,

Have you found the special zlib build for TX2 to optimize the performance? Any result can be shared? Thanks

No, I did not find anything.