I Requested eight camera v4l2 buffer in the function of request_camera_buff,then these eight v4l2 buffer was be used for two global g_buff （the g_buff0 and g_buff1）of two converter,the g_buff0 and g_buff1 mapped separately,but every converter upto 15FPS about,while the camera single converter can upto 30FPS.How can I realised two converter both upto 30FPS at the same time when I Requested four camera v4l2 buffer?
camera_v4l2_cuda.cpp (27.3 KB)
Are you on r28.1?
Do you obseeve the issue by running default camera_v4l2_cuda:
./camera_v4l2_cuda -d /dev/video0 -s 640x480 -f YUYV & ./camera_v4l2_cuda -d /dev/video1 -s 640x480 -f YUYV &
Thank you very much,I had dealed the problem by used as flowlling code:
ctx->conv0_output_plane_buf_queue = new queue < nv_buffer * >;
ctx->conv1_output_plane_buf_queue = new queue < nv_buffer * >;
memcpy(ctx->g_buff1[v4l2_buf.index].start, ctx->g_buff0[v4l2_buf.index].start, ctx->g_buff0[v4l2_buf.index].size);
camera_v4l2_cuda.cpp (31.4 KB)