BI 6 not using Nvidia Card

I just realized that if you highlight the first or latest clip of any cam, then hold control key down and dbl click the alert, it will bring up the .dat analysis window. This is EVEN when I do not have the .dat option checked for that cam. It appears to only be the latest clip for every cam. Not sure when this change happened. Running 6.1.2.0

Screenshot 2026-09-14 112436.jpgScreenshot 2026-09-14 112522.jpg
 
  • Like
Reactions: Flintstone61
Tinman said:
I just realized that if you highlight the first or latest clip of any cam, then hold control key down and dbl click the alert, it will bring up the .dat analysis window. This is EVEN when I do not have the .dat option checked for that cam. It appears to only be the latest clip for every cam. Not sure when this change happened. Running 6.1.2.0

View attachment 249062View attachment 249063
Click to expand...
Version 5.9.9.28 (Feb 2025)
See this post
I requested this feature.
It took awhile longer (Aug 2025) before the Ctrl-double-click worked. Link
 
Last edited:
  • Like
Reactions: Flintstone61, observant1, looney2ns and 2 others
cool.
I just tried it.
I'm getting some decent millisecond times out of the 6.1.2.0 update with the use of nvidia Tensor cores. I am pleased to report that sometimes I see where the GPU is busier than the CPU. like about CPU 8% and GPU 14% ratio
1789540329589.png
 
  • Like
Reactions: observant1
ELA said:
I am experiencing something similar with my GT1030 vs Intel UHD 630 GPUs. It appears to work better when the 630 is selected for AI use than the GT1030.
I know the GT1030 is a weak competitor but should do slightly better than the 630?. I upgraded to the very latest Ver6 release but did not help.
I wonder if there is a bug in which GPU is actually being used for AI? When I use task manager I can see that the GT1030 is being more heavily utilized than the 630, even though the main setup AI has the 630 selected for AI use.
I am leaving the 630 selected as I get the best roundtrip times that way whatever is happening under the hood.
Click to expand...
I have exactly the same problem, persisting also with the latest version and I don't see a CUDA option for my 1050 ti. Performance is far better with 630 selected than 1050 (like 10x faster) which seems unlikely. Tried to use Driver Uninstaller and disable the 630 from windows devices, but after a reboot everything comes back to this situation.

I've tried to reach out to the support team , to no avail. Anyone else please?

Screenshot 2026-09-16 alle 10.24.13.png

EDIT: I've done some testing with a CPU Load Chart and I confirm that selecting DirectML: Inter UHD 630 from the list the Nvidia card is actually getting CPU load and no load to the intel internal GPU. Selecting GTX 1050 option is putting load on the Intel GPU. Something is wrong in Blue Iris?
 
Last edited:
its fixed in 6.1.2.0 I believe. maybe earlier. But thats the one I tried and it straightened out. i also had to get the latest nvidia driver and do a clean install, and reboot, and the therewas NO nvidia in the list! so did an "end task" in task mgr, and blue iris found it then. then i went to bed. after i got up i still had lousy millisecond times, so I shit canned all the excess models, and am running it leaner. now its good.
1789553044014.png1789553074373.png
 
Last edited:
Flintstone61 said:
its fixed in 6.1.2.0 I believe. maybe earlier. But thats the one I tried and it straightened out. i also had to get the latest nvidia driver and do a clean install, and reboot, and the therewas NO nvidia in the list! so did an "end task" in task mgr, and blue iris found it then. then i went to bed. after i got up i still had lousy millisecond times, so I shit canned all the excess models, and am running it leaner. now its good.
View attachment 249170View attachment 249171
Click to expand...
Thanks for your reply. Unfortunately I'm on v 6.1.3 and it's not fixed. I don't know if there's something to do to reset that combo box.... I think it's dynamically loaded at startup.
 
Il_Pres said:
Thanks for your reply. Unfortunately I'm on v 6.1.3 and it's not fixed. I don't know if there's something to do to reset that combo box.... I think it's dynamically loaded at startup.
Click to expand...
Hello II_Pres,
I have not installed the latest with MicrosoftML & Tensor core support so I cannot say for sure in that version. However when using pre-6.1.2.0 versions and selecting DirectML the NVidia and Intel GPUs enumeration order is reversed. As you found you need to select the Intel GPU in order to take advantage of the NVidia GPU. The newer versions with MicrosoftML support uses a different driver that appears to get the enumeration order correct for those selections. Most likely the older directML enumeration order remains reversed. BI Support has said it is out of their control as the issue is at the DirectML driver level.
 
  • Like
Reactions: Il_Pres and Flintstone61
maybe when blue iris was installed it detected the GPU's in a certain order, and enumerated them 0 and 1. perhaps if you uninstall the Nvidia drivers and select " Clean Install" on re-install, it may ( as in my case) straighten itself out.
 
For Me GPU 0 is intel and GPU 1 is Nvidia. I can't see any Windows.ML option in the combobox like shown in the screenshot above. Is this normal? For now, I'll simply stay with the inverted Direct.ML option that seems to load the Nvidia GPU. Thanks all for contributing !
 
  • Like
Reactions: Flintstone61
I’m keeping some processing power available in hopes they come out with the “ipcam-Alien” and “ipcam-UFOabuction” models
. IMG_5012.jpegIMG_5017.png
 
  • Haha
Reactions: samplenhold and Il_Pres
You must log in or register to reply here.
Top Bottom