# Streaming test (RTSP, MJPG)
### Test streaming server on PC
* server and environment

0. hardware specs:
    cpu:
        Architecture:        x86_64
        CPU op-mode(s):      32-bit, 64-bit
        Byte Order:          Little Endian
        CPU(s):              12
        On-line CPU(s) list: 0-11
        Thread(s) per core:  2
        Core(s) per socket:  6
        Socket(s):           1
        NUMA node(s):        1
        Vendor ID:           GenuineIntel
        CPU family:          6
        Model:               158
        Model name:          Intel(R) Core(TM) i7-8700 CPU @ 3.20GHz
        Stepping:            10
        CPU MHz:             3212.542
        CPU max MHz:         4300.0000
        CPU min MHz:         800.0000
        BogoMIPS:            6384.00
        Virtualization:      VT-x
        L1d cache:           32K
        L1i cache:           32K
        L2 cache:            256K
        L3 cache:            12288K
        NUMA node0 CPU(s):   0-11
    mem:
        Memory block size:       128M
        Total online memory:      32G
        Total offline memory:      0B
    gpu:
        Nvidia 1080Ti

1. use jagereye_ng.io.streaming StreamReaderThread 30fps, 6 frames buffer,
    which keep fetch newest frame from video source. StreamReader gstreamger is default pipeline setting.
2. streaming server with 1 client connection  
ps:rtsp server cpu usage grows when multiple client connection establish.
3. server test on 10.10.0.39
4. ipcam 10.10.0.89
5. rtsp server 4 threads

* decode video streaming + encode streaming performance
| |mjpeg |rtsp|
|-|-|-|
|cpu |55*-80% |75*-115%|
|memory |1% |1%|
*cpu lower bound is both mjpeg rtsp streamReader connect to same source at the same time

* encode simple image streaming performance
| |mjpeg |rtsp|
|-|-|-|
|cpu |55-70% |60-80%|
|memory |0.3% |0.4%|

#### client media player (on pc)
* connect to rtsp server (30 fps 1080p h.264) performance
| |nx | vlc|
|-|-|-|
|cpu |15-25% | 3-5%|
|gpu mem | 89mb | 98mb |
|memory |1.5% | 0.6%|
|latency |0.5-1s | 1-1.5s|
|fps |28-29 | 30|
|frame drop |sometimes | often|
|bandwidth |1.5-2.5Mbps|NA|

*fps: real fps is lower than the number
* connect to mjpeg server (30 fps 1080p) performance
| |nx | vlc|
|-|-|-|
|cpu |45-60% | 7-12%|
|gpu mem | 74mb | 12m |
|memory |1.8% | 0.7%|
|latency |0.5-1s | 0.5-1s|
|fps |*30 | NA|
|frame drop |sometimes | often|
|bandwidth |~75Mbps|NA|

*fps: real fps is lower than the number

--------------------------------------------
## Test streaming server on tx2
* server and environment

    0. hardware specs:
        cpu:
            Architecture:         aarch64
            Byte Order:           Little Endian
            CPU(s):               6
            On-line CPU(s) list:  0,3-5
            Off-line CPU(s) list: 1,2
            Thread(s) per core:   1
            Core(s) per socket:   4
            Socket(s):            1
            Vendor ID:            ARM
            Model:                3
            Model name:           Cortex-A57
            Stepping:             r1p3
            CPU max MHz:          2035.2000
            CPU min MHz:          345.6000
            BogoMIPS:             62.50
            L1d cache:            32K
            L1i cache:            48K
            L2 cache:             2048K
            Flags:                fp asimd evtstrm aes pmull sha1 sha2 crc32
        mem:
            Memory block size:         1G
            Total online memory:       8G
            Total offline memory:      0B
        gpu:
            Nvidia 256 cuda core
    1. use jagereye_ng.io.streaming StreamReaderThread 30fps, 6 frames buffer,
        which keep fetch newest frame from video source. StreamReader gstreamger is use gpu decode.
    2. streaming server with 1 client connection  
    ps:rtsp server cpu usage grows when multiple client connection establish.
    3. server test on 10.10.0.69
    4. ipcam 10.10.0.188
    5. rtsp server 4 threads

* decode video streaming + encode streaming performance
| |mjpeg | rtsp|
|-|-|-|
|cpu |80-90% | 38-180%|
|memory |2.3% | 1.9-3.9%|
|gpu |*0% | *0%|

*gpu 0% might not be sampled while decode 1080p video

* encode simple image streaming performance
| |mjpeg | rtsp|
|-|-|-|
|cpu |50-55% | ~130%|
|memory |0.9% | 1.6%|
|gpu |0% | 0%|

### client media player (on pc)

* connect to rtsp server (30 fps 1080p h.264) performance
| |nx | vlc|
|-|-|-|
|cpu |37% | 6.9%|
|gpu mem |78mb | 98mb |
|memory |1.4% | 0.6%|
|latency |0.5-1s | NA (too lag to measure, maybe 0-1s)|
|fps |20 | *30|
|frame drop |sometimes | frequently|
|bandwidth |1.5Mbps|NA|

*fps: real fps is lower than the number

* connect to mjpeg server (30 fps 1080p) performance
| |nx | vlc|
|-|-|-|
|cpu |65-72% | 41-59%|
|gpu mem | 43mb | 13m |
|memory |1.4% | 0.8%|
|latency |<0.5s | <0.5s|
|fps |29.29 | NA|
|frame drop |seldom | seldom|
|bandwidth |~75Mbps|NA|

*fps: real fps is lower than the number

---------------------------------------------
### client media player (on PC)

* doing nothing resource usage
| |nx | vlc|
| - | - | -|
|cpu |5-9% | 0%|
|gpu | 40m | 0m |
|memory |1.4% | 0.5%|

* connect to ipcam 10.10.0.89 (30 fps 1080p h.264) performance
| |nx | vlc|
| - | - | -|
|cpu |40-50% | 5-10%|
|gpu | 47m | 98m |
|memory |1.7% | 0.4%|
|latency |0.1-0.5s | 0.1-0.5s|
|frame drop |sometimes | seldom|
|bandwidth |5 Mbps|NA|

---------------------------------------------
## NX recording
file ext: mkv
codec: same as streaming input format (H.264 MPEG-4 AVC, MJPG)