Hi platinunjsi
I'm a novice here, but I've managed to get multiple GPU WUs working on my machine (through the helpful comments of others on this forum).
You need to create a file called app_config.xml (using e.g. Notepad) and put it in the /BOINC/projects/www.worldcommunitygrid.org/ folder.
My file says this:
<app_config>
<app>
<name>opng</name>
<gpu_versions>
<gpu_usage>0.125</gpu_usage>
<cpu_usage>0.5</cpu_usage>
</gpu_versions>
</app>
<report_results_immediately/>
</app_config>

The number 0.125 means I can run up to 8 WUs simultaneously (i.e. 1/0.125). The 0.5 means those WUs can each share a thread (I think!).
Then you need to go into the BOINC Manager, Options, and click Read config file.
I think that should do it.
My GTX1660S will crunch 6 or 8 GPU WUs simultaneously, so your GPU will no doubt do more. Perhaps start with 0.125 and see how it goes. A bit of trial and error looking at GPU-Z got me to the right place.
Hope this helps.
Let me know if you have problems.

Oh dear.
I wonder if that's because you have set your system to run only the GPU WUs.
I'm running GPU WUs and ordinary CPU WUs at the same time. My system is currently running 6 GPU tasks and 9 CPU tasks (with those app_config.xml settings).
Perhaps try e.g. CPU = 0.25 to see if that would run more?

EDIT: If you do have CPU WUs running at the same time, I wonder whether some of them may need to finish before the new GPU tasks start.

And its suddenly just kicked in doing 8 units, which seems to be enough, ill check the times for each unit though and make sure its worth it, a single unit was completed in just over 6 mins.

I didn't see any WUs utilizing Nvidia card so far.
BOINC reports this

06/05/2021 09:42:38 |  | CUDA: NVIDIA GPU 0: GeForce GTX 950 (driver version 461.40, CUDA version 11.2, compute capability 5.2, 2048MB, 1678MB available, 1964 GFLOPS peak)
06/05/2021 09:42:38 |  | OpenCL: NVIDIA GPU 0: GeForce GTX 950 (driver version 461.40, device version OpenCL 1.2 CUDA, 2048MB, 1678MB available, 1964 GFLOPS peak)
06/05/2021 09:42:38 |  | OpenCL: Intel GPU 0: Intel(R) HD Graphics 4600 (driver version 20.19.15.4531, device version OpenCL 1.2, 1630MB, 1630MB available, 192 GFLOPS peak)
06/05/2021 09:42:38 |  | OpenCL CPU: Intel(R) Core(TM) i7-4770 CPU @ 3.40GHz (OpenCL driver vendor: Intel(R) Corporation, driver version 5.2.0.10094, device version OpenCL 1.2 (Build 10094))

What exactly do I put in app_config.xml to use only Nvidia card?

UPDATE:

Option <use_all_gpus>1</use_all_gpus> is invalid

06/05/2021 09:42:38 | World Community Grid | Unknown tag in app_config.xml: use_all_gpus

the use_all_gpus flag goes in the cc_config.xml file.

however, if you want to use ONLY the nvidia GPU for OPNG, you can certainly do that too. just drop the intel GPU into an exclude_gpu block into the cc_config.xml so that WCG doesnt use it at all.

read the documentation from BOINC. it lists all the available options and what they do.

https://boinc.berkeley.edu/wiki/Client_configuration

something like this in the options section of your cc_config.xml would exclude your Intel GPU, leaving only the nvidia GPU to crunch

<exclude_gpu>
   <url>http://www.worldcommunitygrid.org/</url>
   <type>intel_gpu</type>
</exclude_gpu>