alastairstevenson
Staff member
Hello,
Both NVRs show similar utilisation - and both remain very responsive at both the GUI and command line and VGA/HDMI interfaces.
One has 6 IP cameras recording 1080P continuous and motion/intrusion etc detection, and another 2 720P cameras continuous recording, and is also feeding from PoE ports 5 1080P streams to another NVR and 3 1080P streams to a QNAP NAS with Surveillance Station. And it runs absolutely fine, no recording glitches, responsive UI, responsive at the command line, no errors in the logs, zero crashes.
As I said - I'm impressed with how well it performs.
I understand where you are coming from in your comments about the load averages.
But I believe you are comparing the figures with what you'd see on a Linux box sitting waiting for your next keystroke, or maybe browsing the net or downloading an update.
Compared with that, the NVR is a very busy box indeed, and it's normal to have a large number of items queued up ready for processing. Remember it's a real-time processing environment.
That does not necessarily mean it's overloaded - though it could do in other circumstances.
Remember also that the SoC has a fast and busy DSP on board that is used to do the hardware encoding and decoding of video, and is under control of the main CPUs which queue up its tasks.
To reinforce the statement that what's being seen is a consequence of the type of job the CPU is performing - take a look at the figures for an IP camera.
You will see a very similar picture, showing what you might consider some scarily high load averages. Though on the Hik cameras, the CPU is much more utilised, depending somewhat on what analytics have been configured and enabled.
Here is an example:
Both NVRs show similar utilisation - and both remain very responsive at both the GUI and command line and VGA/HDMI interfaces.
One has 6 IP cameras recording 1080P continuous and motion/intrusion etc detection, and another 2 720P cameras continuous recording, and is also feeding from PoE ports 5 1080P streams to another NVR and 3 1080P streams to a QNAP NAS with Surveillance Station. And it runs absolutely fine, no recording glitches, responsive UI, responsive at the command line, no errors in the logs, zero crashes.
As I said - I'm impressed with how well it performs.
I understand where you are coming from in your comments about the load averages.
But I believe you are comparing the figures with what you'd see on a Linux box sitting waiting for your next keystroke, or maybe browsing the net or downloading an update.
Compared with that, the NVR is a very busy box indeed, and it's normal to have a large number of items queued up ready for processing. Remember it's a real-time processing environment.
That does not necessarily mean it's overloaded - though it could do in other circumstances.
Remember also that the SoC has a fast and busy DSP on board that is used to do the hardware encoding and decoding of video, and is under control of the main CPUs which queue up its tasks.
To reinforce the statement that what's being seen is a consequence of the type of job the CPU is performing - take a look at the figures for an IP camera.
You will see a very similar picture, showing what you might consider some scarily high load averages. Though on the Hik cameras, the CPU is much more utilised, depending somewhat on what analytics have been configured and enabled.
Here is an example:
Mem: 87608K used, 7828K free, 0K shrd, 500K buff, 46728K cached
CPU: 30.2% usr 35.9% sys 0.0% nic 12.0% idle 0.0% io 0.0% irq 21.7% sirq
Load average: 6.04 5.85 6.20 1/152 30642
PID PPID USER STAT VSZ %VSZ CPU %CPU COMMAND
842 1 root S 406m435.2 0 85.6 {main} /home/davinci
30641 30639 root R 3072 3.2 0 0.9 top
4 2 root SW 0 0.0 0 0.5 [kworker/0:0]
826 1 root S 14972 15.6 0 0.1 /home/process/net_process
821 1 root S 6404 6.7 0 0.1 /home/process/daemon_fsp_app
5 2 root SW 0 0.0 0 0.1 [kworker/u:0]
3 2 root SW 0 0.0 0 0.1 [ksoftirqd/0]
30639 30638 root S 3072 3.2 0 0.0 -sh
610 1 root S 3072 3.2 0 0.0 -/bin/sh
1 0 root S 3068 3.2 0 0.0 init
609 1 root S 3068 3.2 0 0.0 /sbin/inetd -f -e /etc/inetd.conf
611 1 root S 3068 3.2 0 0.0 init
30638 613 root S 2856 2.9 0 0.0 /sbin/dropbear
613 1 root S 2500 2.6 0 0.0 /sbin/dropbear
825 1 root S 2072 2.1 0 0.0 /bin/execSystemCmd
384 1 root S < 1996 2.0 0 0.0 /sbin/udevd -d
712 2 root DW 0 0.0 0 0.0 [dsplogd]
60 2 root SW 0 0.0 0 0.0 [sync_supers]
343 2 root DW 0 0.0 0 0.0 [rxdma_check]
62 2 root SW 0 0.0 0 0.0 [bdi-default]
#