High CPU and memory Usage

[bphillips]

I built what I consider a beast of a computer to run Blue Iris in my new home

CPU - Intel i9-10850K 3.60ghz
Memory - 32GB of DDR4 2933
512GB SSD to run the OS and handle the Blue Iris DB
Two 8TB Western Digital Purple hard drives in spanned configuration, giving me 14.5TB of usable storage

• I optimized CPU usage for BI using the CPU optimization guide
• I'm using sub-streams set-up exactly as instructed on the sub-stream guide
• I have 9 Amcrest 8K cameras, 3 Amcrest 4MP cameras, and an Amcrest AD110 doorbell cam (2MP, not using a sub-stream on this cam)
• I'm also running deepstack on this computer (set to only analyze on motion) and using ALPR (plate recognizer)
• Blue Iris version 5.5.1.18

My CPU consistently runs at about 35% and routinely spikes to above 80%, even into the high 90%. My memory usage is generally around 10GB but spikes higher at times. Both of these seem high to me. Is this normal or can this be improved somehow? In the Windows 10 task manager BI is usually 90% of my CPU usage with Python (deepstack) sometimes accounting for 25% of my CPU usage at most.

 

[looney2ns]

Did you exclude BI and it's folders in Windows Defender or any other Antivirus as per the BI help file?

 

[bphillips]

Did you exclude BI and it's folders in Windows Defender or any other Antivirus as per the BI help file?

Yep, they have been excluded

 

[bphillips]

I found the culprit. I have Blue Iris integrated with my home automaton system, Home Assistant. Disabling the Home Assistant user in BI dropped my CPU usage to 2%. I'll need to work on limiting the resources that Home Assistant is using.

 

[Flintstone61]

Ok you got it, i was going to say, what are your frame rates.....

 

[bp2008]

I don't know how Home Assistant camera integrations work, but if they are always pulling video from each individual camera that would explain it for sure. Blue Iris can't serve video without transcoding it, although that is a planned feature.

If home assistant can get video from the cameras directly, that would be much more efficient. Alternatively if you can get home assistant to request video from the sub stream, that would keep CPU much lower. Again I don't know how the integration works so I can't be more specific.

 

[bphillips]

I don't know how Home Assistant camera integrations work, but if they are always pulling video from each individual camera that would explain it for sure. Blue Iris can't serve video without transcoding it, although that is a planned feature.

If home assistant can get video from the cameras directly, that would be much more efficient. Alternatively if you can get home assistant to request video from the sub stream, that would keep CPU much lower. Again I don't know how the integration works so I can't be more specific.

Yeah, I need to look into it. I kind of hastily added the integration in order to test using camera motion to trigger motion in Home Assistant. I don't really need camera streams in Home Assistant. I can probably find a better way to send motion alerts, probably through MQTT.

Oddly, I deleted the Blue Iris integration in Home Assistant, restarted Home Assistant and Blue Iris, then re-added the integration in Home Assistant and my CPU is staying below ~15%. Which I'd say is acceptable.