Regarding NvBuffer

Autonomous Machines Jetson & Embedded Systems Jetson Nano
1

Hello,

The memcpy speed couldn’t keep up with the dqBuffer speed
Is this because NvBuffer is dma_buf?

The memcpy speed couldn’t keep up with the dqBuffer speed
Is this because NvBuffer is dma_buf?
Is it the impact of multithreading?
Is it the effect of dma_buf?

Thank you.

3

Hi,
If you call memcpy(), it consumes CPU loading. A better way is to leverage hardware converter. You can create NVvBuffer and call NvBufferTransform() for copying data from one NvBuffer to the other.

1 Like
4

Hello,

You tell me the usage of using NvBufferTransform() functions is in 12_camera_v4l2_cuda sample app.

Related to Using the function similar to what you were trying to use memcpy,
Can I see this in different example?

of you can give me some code sniffet?

Thank you.

5

Hi,
You may also refer to 00_video_decode or 07_video_convert.