Sure, I ran some tests using the benchmarking project. Here are the results:
 

User Name: JaredF
Machine: Dell XPS
VP Version: 18
VP Build Number: 482
CPU: Intel i7 11700K
Cores: 8
GPU: Nvidia RTX 3060 TI
iGPU: UHD 750
OC: No
Render Frame Size: FHD
HW Acc: Nvidia
Render as: Magix AVC
Envode Mode: NVENC
Render time: 0:48

User Name: JaredF
Machine: Dell XPS
VP Version: 18
VP Build Number: 482
CPU: Intel i7 11700K
Cores: 8
GPU: Nvidia RTX 3060 TI
iGPU: UHD 750
OC: No
Render Frame Size: FHD
HW Acc: Nvidia
Render as: Magix AVC
Endode Move: Intel QSV
Render time 0:46

User Name: JaredF
Machine: Dell XPS
VP Version: 18
VP Build Number: 482
CPU: Intel i7 11700K
Cores: 8
GPU: Nvidia RTX 3060 TI
iGPU: UHD 750
OC: No
Render Frame Size: UHD
HW Acc: Nvidia
Render as: Magix AVC
Envode Mode: NVENC
Render time: 1:20

User Name: JaredF
Machine: Dell XPS
VP Version: 18
VP Build Number: 482
CPU: Intel i7 11700K
Cores: 8
GPU: Nvidia RTX 3060 TI
iGPU: UHD 750
OC: No
Render Frame Size: UHD
HW Acc: Nvidia
Render as: Magix AVC
Endode Move: Intel QSV
Render time 1:19

So at this point it looks like render speed of NVENC vs. QSV is basically a wash. Is there any consensus about whether one results in a higher quality render than the other?

0:48 for FHD & 1:20 for UHD match the results I got testing a Nvidia 3060 Ti in my 9900K workstation:
https://www.vegascreativesoftware.info/us/forum/nvidia-3060-ti-vegas-18-benchmarks--126075/#ca786277

It looks like the 8-core 11700K & 3060 Ti are a good modern & affordable combo for Vegas 18 since Vegas has always responded better to core speed vs. core count...

However, the scarcity of the 3xxx GPUs is forcing people to choose ready-built systems vs. DIY.. For others considering the Dell XPS SE desktop, note that there is (1) M.2 for the OS and (2) 2.5" and (1) 3.5" available bays/SATA ports for internal storage. The (1) x16 PCIe is for the GPU, leaving you a PCIe x1 and PCIe x4 for internal expansion and (1) USB-C and (7) USB 3.1 ports for external expansion.

Explanation of the XPS' storage configuration: https://www.dell.com/support/manuals/en-us/xps-8940-desktop/xps-8940-setup-and-specifications/storage?guid=guid-a00459e0-2a64-449c-9f02-581bf2e3c2cf&lang=en-us

Forum discussing upgrading the XPS with better cooling so it does not throttle-down during long renders:
https://www.dell.com/community/XPS-Desktops/XPS-8940-Better-cooler/td-p/7690322

@JaredF If you would like your results added to the Benchmarking project, just do a copy and paste to that thread. If you do that, consider also getting your Average FPS results for the "region 1" section in the project, it will be the same result for all 4 entries, so you only need to do it once, and so enter the same value in all of your 4 entries.

Study carefully the steps needed to do so. Note also to set Dynamic Ram Preview to zero, from it's normal default of 200 for the duration of the test. Attempt to get the result with as few passes as possible, as even with DR set to 0, it will still "learn" from previous round trips.

The link is in my signature.

There's also a link in my signature to render quality metrics tables that shows Nvenc is better than Qsv.

However, this mainly applies to UHD testing. In the FHD testing there's not that much in it, Nvenc hevc still has the edge though, see below.

UHD ...

FHD ...

@JaredF OTOH All of my render quality comparisons were at best done with Intels UHD 630 iGpu. It’s possible that the UHD 750 has improved. If you are interested in finding this out, I can send you via dropbox the FHD source file that I used for previous tests.

It would require you to render the 27 second source file out using QSV to a fixed data rate and file size, as close as is possible, to figures I will give you, I understand that it’s never possible to get it to perfectly match data rates and file sizes.

You would then make available the 27s QSV rendered out file for download and I would then do the Render Quality testing, and update my tables accordingly.

There is a quality setting, 1 to 7, for QSV in the most recent iGpu's. I always made sure to set this to highest quality, the scale may have changed?

It would be very interesting to know if the quality has changed.

A little busy the rest of the week, but I'll look into the FPS test and QSV render test when I have time later this month. Thanks for all the info.

A question for JN, Rhino and others: At what Dynamic Ram Preview do you generally edit with for best timeline playback and overall performance?

@JaredF I always leave it at 200, unless on say my 4 core laptop for example, which has to be set to zero to complete the benchmarking uhd part, or similar media.

@JaredF I see you have posted your results in the benchmarking thread. Ok, I’ll hold of posting them until you update your benchmarking FPS results, thanks.

In the meantime I’ll upload the spreadsheets to the download zip location with your results posted today.

I will post a message in the benchmarking thread later when done.

Sure, no problem, I ran the FPS test with dynamic ram preview at 0 and the best result I got was an AVG of 12.5. But honestly, the results seemed all over the place each time I tried it. Often I got AVG of 9.5. Not sure why the difference.

I suppose the number that really matters to me is real-time playback of highest quality Canon C70 footage, which is XF-AVC, All-I, 10-bit,4096x2160 at 23.97 frame rate, with a basic Camera LUT applied. With this I'm getting no worse than 11 to 12fps playback on the timeline all the time in Best Full (same for Best Auto interestingly) and sometimes it goes all the way up to 23.97.

I don't think there is any setting in Vegas that will further improve performance. Perhaps future updates or versions of Vegas will change that.

Otherwise, bypass the LUT, switch to preview/full for fluid playback, create a proxy file, etc.

@JaredF “Often I got AVG of 9.5. Not sure why the difference.” Anyway, pick a number, it’s difficult to pin down, we have no more scientific, programming way of detecting the avg. fps values than you eyeballing it. Give it your best shot and post, update in the benchmarking thread when u have the time.

It appears likely that the next update will bring improvement to XAVC playback.

A big update on my Vegas Pro 18 tests using the new Intel 11th Gen processor:

I was impressed with the rendering test results I was getting on the new computer, but surprised at the mediocre real-time playback I was getting with 4K footage on my timeline. So I started to think through bottlenecks and whether this could be a case of user error.

I’m happy to report that it was a case of user error.

What I finally realized is that while I’ve been playing back 4K footage off a very capable external SSD, it’s connected to the computer using a USB-C to USB-A adapter cord, as my old system didn’t have a USB-C connection. This adapter cord turned out to be the bottleneck. 

I transferred the same 4K All-I 23.98 footage to the computer’s internal SSD drive, and I now get perfect 4K playback at 23.98 frames per second. I also get perfect or very close to perfect playback off the internal HDD.

This leaves me much more satisfied with both the new computer and Vegas Pro 18.

Out of curiosity, I reran all the rendering tests using the benchmark file. The results didn’t change using the internal SSD except for one second faster using QSV encoder. So apparently there wasn’t much of a bottleneck there. (Results updated above.)

One final note to this complicated saga is that a USB-C to USB-A adapter cord isn't necessarily unusable. In my case, the cord was USB-C to the slower USB-A Gen2, which I never even noticed until it created this problem. I just tested it and an adapter cord with USB3 seems to work fine. Moral is, mind your cords.

@JaredF curious if you tried selecting different "hardware decoder to use" choices in the File I/O tab (Nvidia NVDEC, Intel QSV, or None) and whether any cause decoder utilization to show up in the Task Manager performance charts for Intel or Nvidia gpus. I don't think Vegas can use decoder hardware for XF-AVC footage but if it can, that would be great. And might give you better performance if the decoding and encoding were not done at the same time on the same gpu.

Interesting question, and not something I would have known to check.

Under "Hardware decoder to use" the auto selection defaults to "Intel QSV." When I check the task manager during timeline playback, GPU shows up as 5% or 6%, fluctuating a bit. During rendering this goes up to 20% to 30%. There's also a note under GPU engine that says "GPU 1- copy" or "GPU 1-3D" or "GPU 1-Video encode" the last only while rendering.

When trying the test above with "Nvidia" or "off" selected in Hardware Decoder to Use," including after restarting the program as directed, the percentage readings appear to be exactly the same in task manager.

Can't say I understand most of this...

Any chance you can upload one of your sample c70 clips so I can try it on my Dell xps15 laptop? I found some sample c70 clips here I'm not sure if they're similar to yours.

@JaredF @RogerS Ok, Jared has kindly rendered out a FHD clip using Intel QSV UHD 750, Xe.

This was with a view to see if there was any quality change one way or the other to the previous UHD 630 iGPU's etc.

Although it weighs in at the lower end of the "Size" and "Data range", it is within the range, as I requested, thanks Jared.

To me, quality wise, it appears about the same as previous Intel iGPU's. See item 25 in the table below.

Thanks for the quality comparison. This is great info that really isn't readily available anywhere!

@RogerS Jared, in the interests of science, has re-done it, so it's now very much in the middle of the pack with respect to file size and data rate, will update later. Not expecting much change, but thank you Jared.

@JaredF @RogerS Is updated now, item no. 25 in the table posted earlier.