Always good to look at the options and possibilties. I have done some tests, and I don't see that the GPU is being used when directed, and there is no difference in performance.

Thanks for testing. Sorry to hear that it made no difference.

I don't have access to a laptop with a dedicated GPU card so I can't test this out, but I agree that it does sound promising and it might make a huge difference for some users with the right hardware.

It's certainly easy to try, and I would encourage anyone who's been frustrated by PD not being able to use their laptop's nVidia GPU to give it a go and see if it works. Thanks for posting this!

YouTube/optodata

DS365 | Win11 Pro | Ryzen 9 3950X | RTX 4070 Ti | 32GB RAM | 10TB SSDs | 5K+4K HDR monitors

Canon Vixia GX10 (4K 60p) | HF G30 (HD 60p) | Yi Action+ 4K | 360Fly 4K 360°

I am aware of three different methods to get the laptop to use the Nvidia graphics card instead of the Intel integrated graphics with PowerDirector. Each user that claim this show the screen recordings with GPU-Z/sensor tab open or with the task manager open. I am convinced that video editing does use the Nvidia graphics, shortening the battery life and giving little advantage with the change in settings.

Not one of the users above have shown the Nvidia GPU being used while producing to my knowledge like in Nvenc encoding.

I did post one of the methods earlier to a user who asked and have not heard back from this individual : https://forum.cyberlink.com/forum/posts/list/80933.page . Maybe it works maybe it don’t but we will never know if users don’t bother to respond. Using the Windows 10 task manager is a good start.

About using Device Manager to kill (disable) the Intel GPU - in the competitor's program, both GPU's work together (or side by side). Apparently there is some kind of "scheduler" in the program that knows which GPU works best for certain parts of an encode. I don't think PD works that way - it seems all or nothing as far as which GPU it is using. But thought I would add this for completeness.

Pmikep – You have posted a great service to laptop users. Try this: Either open GPU-z and go to the sensor tab or open the task manager to show GPU usage. Open PD18, do some editing and the Nvidia GPU should not be active. Close PD18. Make the changes described in your first post, reopen PD18, do some video editing, and I believe that you will find that the Nvidia card becomes active when you look at the graph in either GPU-z or the task manager. That is what I saw in the author’s posted video using PowerDirector on YouTube.

Please post your findings here.

"That is what I saw in the author’s posted video using PowerDirector on YouTube."

??? Do you mean you? Tried to find you on youtube. No joy. Got a link?

I would just add an 'if it ain't broke don't fix it' observation here. If, like me, you have a good Laptop with a good Nvidia GPU that PD doesn't access, ask yourself if that matters. In my case, PD 18 uses my i7 processor and it renders quickly and accurately. I see no need to try and get it to use the Nvidia 1060. So, my advice would be to only try and force the program to access the GPU if you are unhappy with how it currently renders projects.

Quote I would just add an 'if it ain't broke don't fix it' observation here. If, like me, you have a good Laptop with a good Nvidia GPU that PD doesn't access, ask yourself if that matters. In my case, PD 18 uses my i7 processor and it renders quickly and accurately. I see no need to try and get it to use the Nvidia 1060. So, my advice would be to only try and force the program to access the GPU if you are unhappy with how it currently renders projects.

I fully and wholeheartedly agree, TDK1044!
However, still a kind nudge to Cyberlink to look at the utilization of GPU's on laptops, as most of the other video software manufacturers seem to be able to do the trick.

Quote I would just add an 'if it ain't broke don't fix it' observation here. If, like me, you have a good Laptop with a good Nvidia GPU that PD doesn't access, ask yourself if that matters. In my case, PD 18 uses my i7 processor and it renders quickly and accurately. I see no need to try and get it to use the Nvidia 1060. So, my advice would be to only try and force the program to access the GPU if you are unhappy with how it currently renders projects.

i don't understand? what do you mean?

Quote


I fully and wholeheartedly agree, TDK1044!
However, still a kind nudge to Cyberlink to look at the utilization of GPU's on laptops, as most of the other video software manufacturers seem to be able to do the trick.

Yep. Right there with you.

Quote

i don't understand? what do you mean?

There are two ways for your project to be rendered. PD either uses your onboard CPU, or it uses a dedicated video card (GPU). A lot of laptops are configured to use the CPU for everything except gaming, and when you load a game, the GPU then gets activated. That's how my laptop works. Nvidia has a small icon that changes from black and white to color when the GPU is active. Some people are trying to force PD to use their GPU instead of the default CPU. My post was simply pointing out that many times you don't need to do that if your project is being rendered accurately and quickly. That's the case with my i7 CPU and I have no need to force PD to use the GPU.

If I force PD to use the GPU on my Dell XPS 9560 (i7, NVIDIA 1050), the performance in "timeline editing" is worse than without usage
So I still use my Intel 630 GPU

I just bought the new GTX-1650 Super. (Keyword: Super.) It has the Turing NVENC hardware encoder on it, and gives about the same performance as the more expensive 1660.

I don't have PD 18 at this time (waiting on the first patch before buying the permant version). So I can't test NVENC vs. QS myself. But I expect this is one of those things where YMMV, depeding in GPU.

(In fact, I'm still running PD 15 with the old Nvidia 411.x drivers. So I can't even test PD 15 using my new GPU.)

P.S. The guys in a competitor's forum say to use the Stuido version of Nvidia's drivers.

I haven't seen anyone here say that. And I don't know what the difference is between Studio and Game Ready.

Quote There are two ways for your project to be rendered. PD either uses your onboard CPU, or it uses a dedicated video card (GPU). A lot of laptops are configured to use the CPU for everything except gaming, and when you load a game, the GPU then gets activated. That's how my laptop works. Nvidia has a small icon that changes from black and white to color when the GPU is active. Some people are trying to force PD to use their GPU instead of the default CPU. My post was simply pointing out that many times you don't need to do that if your project is being rendered accurately and quickly. That's the case with my i7 CPU and I have no need to force PD to use the GPU.

I'd just like to point out that while the gist of what you're saying is true, it's not quite an accurate description of the choices on most systems.* For a computer with an Intel CPU and a dedicated AMD or nVidia GPU, there are actually three producing options:

1. Produce with hardware acceleration (aka "Fast video rendering technology") using the dedicated GPU (with nVidia cards, the silicon used is called NVENC and it's completely separate from the gaming section of the chip)


2. Produce with hardware acceleration using the Intel CPU's built in video encoding silicon, called QuickSync.


3. Produce without any hardware acceleration using only software instructions by the CPU. This is used when the Fast video rendering box is disabled and also when SVRT is chosen.

The encoding/producing performance is listed here from most powerful/fastest to least, and I believe your description was referring to the first two items while referring to QuickSync hardware producing as "CPU."

*AMD CPUs do not have dedicated video encoding silicon on board, so only options 1 and 3 are available, and that does match your description.

Quote I just bought the new GTX-1650 Super. (Keyword: Super.) It has the Turing NVENC hardware encoder on it, and gives about the same performance as the more expensive 1660.

I don't have PD 18 at this time (waiting on the first patch before buying the permant version). So I can't test NVENC vs. QS myself. But I expect this is one of those things where YMMV, depeding in GPU.

(In fact, I'm still running PD 15 with the old Nvidia 411.x drivers. So I can't even test PD 15 using my new GPU.)

P.S. The guys in a competitor's forum say to use the Stuido version of Nvidia's drivers.

I haven't seen anyone here say that. And I don't know what the difference is between Studio and Game Ready.

The Studio version was referenced here https://forum.cyberlink.com/forum/posts/list/80290.page#post_box_330215 by PDM with presumably info from CLD for suggestion but a few users experience was negative. Would have been nice if PDM stated CLD view for suggestion, but more than likely just fishing for feedback.

Value of any of the Turing GPU's is currently somewhat mute with PD up to version PD18 v2028 as CL never incorporated any of the new features or the proper SDK. So one does not get the speed/quality for Turing GPUs, HEVC B-frame support (1650 does not support anyhow) ,...so on. Additionally, one loses a few things vs prior Nvidia GPU architecture, hardware encode H.264 60i for instance if important to some users based on current camera suite perhaps.

Jeff

Regarding HEVC B-frame support: One of the cool things about the GTX-1650 Super is that it now has the Turing NEVC hardware. (In contrast with the plain 1650, which did not.)

Some interesting observations on my trick - albeit reverse from what the laptop users want. Still, it might be insightful.

I installed PD15 on my Win10 box last night. (I know it's only PD15, and this is the PD18 forum. But on the assumption that the "wiring" for GPU's is the same in both ...)

My Win10 box has an Intel CPU with a UHD 630 internal GPU and an Nvidia GTX-1650 Super for an external GPU. But the driver for the 1650 Super one of the new ones, with motion estimation NVENC, which is not supported by PD15.

So at first, PD did not offer me Hardware Acceleration on the Produce page. (That is, on install and sniffing my graphics cards, PD opted for the Nvidia external GPU. But it couldn't use it because of the too new driver.) But after playing my trick, I now have the Intel QuickSync for HA.

Now, before going further, I should tell you that I apparently have my monitors hooked up different than the norm. I have my Primary monitor hooked up the external GPU. And I have my Secondary monitor hooked up the Intel GPU.

I thought this was the smart way to go because neither GPU would be overly "burdened" by driving two monitors at the same time. (Burdened in quotes because I'm not running any high Res (4K) displays. And because my old GTX-960 on my Win8.1 box does just fine driving two monitors simultaneously.) But I was surprised when I tried using OBS to capture my Second Display. OBS couldn't see it, apparently because OBS was running on the monitor driven by the external GPU, but my other monitor by the internal GPU.

Still, this might offer a trick for the laptop bound crowd trying to get PD to see their external GPU.

So after my first run with PD15, I went to the Graphic Settings panel in Windows, added PD as a Classic App, and set it for the Low Power GPU. Then I restarted PD.

As expected, QuickSync was now offered for Hardware Acceleration.

And also as I've learned to expect, when I went back a third time and set Graphics Setting to the external GPU, PD still offered QuickSync.

(I have seen this behavior before in PD. It seems to get "stuck" on GPU's. It takes some kind of forced reassessment if it's going to change its mind.)

So I am thinking that perhaps the laptop guys could temporairly plug an external monitor into their laptop, driven by their external GPU. (I presume laptops still have a port for an exteral display?) Then change Graphics Setting to High Power.

Unless someone here knows what file to delete to force PD to do a re-sniff of GPU's, users might have to hook up an external monitor before their first run of PD, so that PD sees the external GPU and locks on to it.

But once done, the setting should persist, even with the external monitor disconnected. (Perhaps.)

Worth a try?

Thanks for the detailed update! I hope it helps people out.

With regard to multiple montors and video cards, I'm not sure there is a "normal" way to connect them, and on my system I have two HDMI switch boxes so I can independently switch monitors/video cards on the fly. The only thing I need to do is close PD before switching, and I don't have to reboot for PD to "see" the new monitor when I relaunch it.

I've also found that it's usually better to run both monitors off of the same card when editing. It's actually possible for PD's About screen to show the nVidia icon but then also show the QuickSync option on the Produce page. Running both monitors from the same card eliminates that disphoric behavior.