slower than expected benchmark for RTX4090

[staticmotion]

Hi. I've just upgraded from a RTX2080 Ti to a Gigabyte RTX4090 Gaming OC. The Benchmark5 speed result is lower than the c.120fps I was expecting and I don't know why. I'm getting c.76fps, and the rest of my system is pretty decent:

Ryzen Threadripper 3970X
64GB DDR4 RAM
SABRENT 1TB Rocket Nvme PCIe 4.0 M.2

I've uninstalled and re-installed the GPU drivers, and re-installed NeatVideo 5 (I've got the Virtualdub version) so that's the latest version. Is there something else I can do to get the full use out of my 4090?

Log:

Neat Bench (Neat Image 9.1.0, Neat Video 5.5.11) Windows x64
Copyright (c) 1999-2023 Neat Image team, Neat Video team, ABSoft.
All Rights Reserved.

GPU detection log:

CUDA driver version: 12020
NVIDIA CUDA initialized successfully.
Checking CUDA GPU 1:
GPU device name is: NVIDIA GeForce RTX 4090
24563 MB total (23008 MB available during initialization)
Check passed - will attempt to use the device

Checking OpenCL platform 1 (NVIDIA Corporation):
The platform is not supported.
OpenCL initialized successfully.


Neat Video benchmark:

Frame Size: 1920x1080 progressive
Bitdepth: 32 bits per channel
Mix with Original: Disabled
Temporal Filter: Enabled
Quality Mode: Normal
Radius: 2 frames
Dust and Scratches: Disabled
Repeat Rate: 0% of repeated frames
Jitter Filtration: Normal
Spatial Filter: Enabled
Quality Mode: Normal
Frequencies: High, Mid, Low, Very Low
Artifact Removal: Enabled
Edge Smoothing: Disabled
Sharpening: Disabled


Detecting the best combination of performance settings:
running the test data set on up to 64 CPU cores and on up to 1 GPU

CPU Model: AMD Ryzen Threadripper 3970X 32-Core Processor
GPU 1: NVIDIA GeForce RTX 4090 (CUDA): 24563 MB total (23008 MB currently available), using up to 100%

CPU only (1 core): 3.43 frames/sec
CPU only (2 cores): 6.83 frames/sec
CPU only (3 cores): 9.05 frames/sec
CPU only (4 cores): 11.1 frames/sec
CPU only (5 cores): 12.2 frames/sec
CPU only (6 cores): 13.3 frames/sec
CPU only (7 cores): 14 frames/sec
CPU only (8 cores): 14.7 frames/sec
CPU only (9 cores): 15.5 frames/sec
CPU only (10 cores): 16 frames/sec
CPU only (11 cores): 16.5 frames/sec
CPU only (12 cores): 16.7 frames/sec
CPU only (13 cores): 17 frames/sec
CPU only (14 cores): 17 frames/sec
CPU only (15 cores): 17.1 frames/sec
CPU only (16 cores): 17.2 frames/sec
CPU only (17 cores): 17 frames/sec
CPU only (18 cores): 17.1 frames/sec
CPU only (19 cores): 17.3 frames/sec
CPU only (20 cores): 17.5 frames/sec
CPU only (21 cores): 16.9 frames/sec
CPU only (22 cores): 17.3 frames/sec
CPU only (23 cores): 17.2 frames/sec
CPU only (24 cores): 17.2 frames/sec
CPU only (25 cores): 16.9 frames/sec
CPU only (26 cores): 17.1 frames/sec
CPU only (27 cores): 17 frames/sec
CPU only (28 cores): 16.7 frames/sec
CPU only (29 cores): 16.3 frames/sec
CPU only (30 cores): 16.5 frames/sec
CPU only (31 cores): 15.6 frames/sec
CPU only (32 cores): 15.4 frames/sec
CPU only (33 cores): 15.3 frames/sec
CPU only (34 cores): 15.2 frames/sec
CPU only (35 cores): 13.9 frames/sec
CPU only (36 cores): 13.8 frames/sec
CPU only (37 cores): 13.1 frames/sec
CPU only (38 cores): 13.7 frames/sec
CPU only (39 cores): 12.3 frames/sec
CPU only (40 cores): 12.8 frames/sec
CPU only (41 cores): 12.7 frames/sec
CPU only (42 cores): 12.6 frames/sec
CPU only (43 cores): 12.2 frames/sec
CPU only (44 cores): 12.1 frames/sec
CPU only (45 cores): 12.2 frames/sec
CPU only (46 cores): 12.1 frames/sec
CPU only (47 cores): 11.8 frames/sec
CPU only (48 cores): 12.9 frames/sec
CPU only (49 cores): 12.8 frames/sec
CPU only (50 cores): 12.4 frames/sec
CPU only (51 cores): 12.1 frames/sec
CPU only (52 cores): 11.6 frames/sec
CPU only (53 cores): 11 frames/sec
CPU only (54 cores): 10.8 frames/sec
CPU only (55 cores): 10.3 frames/sec
CPU only (56 cores): 10.2 frames/sec
CPU only (57 cores): 10.1 frames/sec
CPU only (58 cores): 9.39 frames/sec
CPU only (59 cores): 9.12 frames/sec
CPU only (60 cores): 8.97 frames/sec
CPU only (61 cores): 8.36 frames/sec
CPU only (62 cores): 8.11 frames/sec
CPU only (63 cores): 8.04 frames/sec
CPU only (64 cores): 7.89 frames/sec
GPU only (NVIDIA GeForce RTX 4090): 76.2 frames/sec
CPU (2 cores) and GPU (NVIDIA GeForce RTX 4090): 16.8 frames/sec
CPU (3 cores) and GPU (NVIDIA GeForce RTX 4090): 20.5 frames/sec
CPU (4 cores) and GPU (NVIDIA GeForce RTX 4090): 20.4 frames/sec
CPU (5 cores) and GPU (NVIDIA GeForce RTX 4090): 22.7 frames/sec
CPU (6 cores) and GPU (NVIDIA GeForce RTX 4090): 23.8 frames/sec
CPU (7 cores) and GPU (NVIDIA GeForce RTX 4090): 22.9 frames/sec
CPU (8 cores) and GPU (NVIDIA GeForce RTX 4090): 23.3 frames/sec
CPU (9 cores) and GPU (NVIDIA GeForce RTX 4090): 22.8 frames/sec
CPU (10 cores) and GPU (NVIDIA GeForce RTX 4090): 22.9 frames/sec
CPU (11 cores) and GPU (NVIDIA GeForce RTX 4090): 23.7 frames/sec
CPU (12 cores) and GPU (NVIDIA GeForce RTX 4090): 22.8 frames/sec
CPU (13 cores) and GPU (NVIDIA GeForce RTX 4090): 22 frames/sec
CPU (14 cores) and GPU (NVIDIA GeForce RTX 4090): 25 frames/sec
CPU (15 cores) and GPU (NVIDIA GeForce RTX 4090): 25.6 frames/sec
CPU (16 cores) and GPU (NVIDIA GeForce RTX 4090): 25.5 frames/sec
CPU (17 cores) and GPU (NVIDIA GeForce RTX 4090): 25.9 frames/sec
CPU (18 cores) and GPU (NVIDIA GeForce RTX 4090): 25 frames/sec
CPU (19 cores) and GPU (NVIDIA GeForce RTX 4090): 24.3 frames/sec
CPU (20 cores) and GPU (NVIDIA GeForce RTX 4090): 23.6 frames/sec
CPU (21 cores) and GPU (NVIDIA GeForce RTX 4090): 23.7 frames/sec
CPU (22 cores) and GPU (NVIDIA GeForce RTX 4090): 23.4 frames/sec
CPU (23 cores) and GPU (NVIDIA GeForce RTX 4090): 23 frames/sec
CPU (24 cores) and GPU (NVIDIA GeForce RTX 4090): 22.5 frames/sec
CPU (25 cores) and GPU (NVIDIA GeForce RTX 4090): 23 frames/sec
CPU (26 cores) and GPU (NVIDIA GeForce RTX 4090): 22.8 frames/sec
CPU (27 cores) and GPU (NVIDIA GeForce RTX 4090): 22 frames/sec
CPU (28 cores) and GPU (NVIDIA GeForce RTX 4090): 22.3 frames/sec
CPU (29 cores) and GPU (NVIDIA GeForce RTX 4090): 21.5 frames/sec
CPU (30 cores) and GPU (NVIDIA GeForce RTX 4090): 19.9 frames/sec
CPU (31 cores) and GPU (NVIDIA GeForce RTX 4090): 21 frames/sec
CPU (32 cores) and GPU (NVIDIA GeForce RTX 4090): 19.9 frames/sec
CPU (33 cores) and GPU (NVIDIA GeForce RTX 4090): 20.9 frames/sec
CPU (34 cores) and GPU (NVIDIA GeForce RTX 4090): 20.3 frames/sec
CPU (35 cores) and GPU (NVIDIA GeForce RTX 4090): 18.5 frames/sec
CPU (36 cores) and GPU (NVIDIA GeForce RTX 4090): 20.1 frames/sec
CPU (37 cores) and GPU (NVIDIA GeForce RTX 4090): 19.5 frames/sec
CPU (38 cores) and GPU (NVIDIA GeForce RTX 4090): 19.7 frames/sec
CPU (39 cores) and GPU (NVIDIA GeForce RTX 4090): 19.5 frames/sec
CPU (40 cores) and GPU (NVIDIA GeForce RTX 4090): 19.7 frames/sec
CPU (41 cores) and GPU (NVIDIA GeForce RTX 4090): 19.7 frames/sec
CPU (42 cores) and GPU (NVIDIA GeForce RTX 4090): 19.3 frames/sec
CPU (43 cores) and GPU (NVIDIA GeForce RTX 4090): 18.7 frames/sec
CPU (44 cores) and GPU (NVIDIA GeForce RTX 4090): 19 frames/sec
CPU (45 cores) and GPU (NVIDIA GeForce RTX 4090): 19.2 frames/sec
CPU (46 cores) and GPU (NVIDIA GeForce RTX 4090): 19.6 frames/sec
CPU (47 cores) and GPU (NVIDIA GeForce RTX 4090): 19 frames/sec
CPU (48 cores) and GPU (NVIDIA GeForce RTX 4090): 19 frames/sec
CPU (49 cores) and GPU (NVIDIA GeForce RTX 4090): 19 frames/sec
CPU (50 cores) and GPU (NVIDIA GeForce RTX 4090): 18.2 frames/sec
CPU (51 cores) and GPU (NVIDIA GeForce RTX 4090): 18.7 frames/sec
CPU (52 cores) and GPU (NVIDIA GeForce RTX 4090): 17.8 frames/sec
CPU (53 cores) and GPU (NVIDIA GeForce RTX 4090): 17.8 frames/sec
CPU (54 cores) and GPU (NVIDIA GeForce RTX 4090): 17.7 frames/sec
CPU (55 cores) and GPU (NVIDIA GeForce RTX 4090): 17.2 frames/sec
CPU (56 cores) and GPU (NVIDIA GeForce RTX 4090): 17 frames/sec
CPU (57 cores) and GPU (NVIDIA GeForce RTX 4090): 16 frames/sec
CPU (58 cores) and GPU (NVIDIA GeForce RTX 4090): 16.1 frames/sec
CPU (59 cores) and GPU (NVIDIA GeForce RTX 4090): 15.2 frames/sec
CPU (60 cores) and GPU (NVIDIA GeForce RTX 4090): 15 frames/sec
CPU (61 cores) and GPU (NVIDIA GeForce RTX 4090): 14.8 frames/sec
CPU (62 cores) and GPU (NVIDIA GeForce RTX 4090): 14 frames/sec
CPU (63 cores) and GPU (NVIDIA GeForce RTX 4090): 14 frames/sec
CPU (64 cores) and GPU (NVIDIA GeForce RTX 4090): 13.9 frames/sec

Best combination: GPU only (NVIDIA GeForce RTX 4090): 76.2 frames/sec

[NVTeam]

What is strange in these results is that the CPU shows less than half of its normal speed (normally it should give about 40 frames/sec from the CPU alone). It is possible that the overall speed is lower because of that as well.

Please run the bandwidthTest utility and post its results.

Please also check the settings of the motherboard and Windows to make sure it all works at full speed, not in some economy/green mode.

Thank you,
Vlad

[staticmotion]

Thanks Vlad. Here are the results from the bandwidth test:

[CUDA Bandwidth Test] - Starting...
Running on...

Device 0: NVIDIA GeForce RTX 4090
Quick Mode

Host to Device Bandwidth, 1 Device(s)
PINNED Memory Transfers
Transfer Size (Bytes) Bandwidth(MB/s)
33554432 12531.9

Device to Host Bandwidth, 1 Device(s)
PINNED Memory Transfers
Transfer Size (Bytes) Bandwidth(MB/s)
33554432 10829.6

Device to Device Bandwidth, 1 Device(s)
PINNED Memory Transfers
Transfer Size (Bytes) Bandwidth(MB/s)
33554432 2049434.6

Result = PASS


How do these results look? The Windows power setting is on high performance, and I used optimised defaults in the BIOS.

[NVTeam]

The first two figures are important:

Host to Device Bandwidth, 1 Device(s)
PINNED Memory Transfers
Transfer Size (Bytes) Bandwidth(MB/s)
33554432 12531.9

Device to Host Bandwidth, 1 Device(s)
PINNED Memory Transfers
Transfer Size (Bytes) Bandwidth(MB/s)
33554432 10829.6

and for some reason they are significantly lower than they should be.

Here is a result from another system:

Host to Device Bandwidth, 1 Device(s)
PINNED Memory Transfers
Transfer Size (Bytes) Bandwidth(MB/s)
33554432 24953.2

Device to Host Bandwidth, 1 Device(s)
PINNED Memory Transfers
Transfer Size (Bytes) Bandwidth(MB/s)
33554432 25065.2

Perhaps there is some issue with the RAM speed of your main system, which then slows down both CPU itself and exchange of data with the GPU.
You may want to double-check the RAM setup (the hardware part) and compare it with the recommendations for that motherboard (which slots should better be populated to achieve the full speed).

Hope this helps,
Vlad

[staticmotion]

Wow, I'm only getting half...

Thanks Vlad, I will investigate!

[staticmotion]

So... Somehow in the BIOS it can gone back to non-XMP RAM timings so I've set that correctly to 3600. Bandwith test definitely an improvement!

Device 0: NVIDIA GeForce RTX 4090
Quick Mode

Host to Device Bandwidth, 1 Device(s)
PINNED Memory Transfers
Transfer Size (Bytes) Bandwidth(MB/s)
33554432 24257.2

Device to Host Bandwidth, 1 Device(s)
PINNED Memory Transfers
Transfer Size (Bytes) Bandwidth(MB/s)
33554432 24764.6

Device to Device Bandwidth, 1 Device(s)
PINNED Memory Transfers
Transfer Size (Bytes) Bandwidth(MB/s)
33554432 2128676.8

Result = PASS



NeatBench test shows a slight improvement, but still not getting 120fps on the GPU, and nowhere near 40fps from the CPU:

Neat Bench (Neat Image 9.1.0, Neat Video 5.5.11) Windows x64
Copyright (c) 1999-2023 Neat Image team, Neat Video team, ABSoft.
All Rights Reserved.

GPU detection log:

CUDA driver version: 12020
NVIDIA CUDA initialized successfully.
Checking CUDA GPU 1:
GPU device name is: NVIDIA GeForce RTX 4090
24563 MB total (23008 MB available during initialization)
Check passed - will attempt to use the device

Checking OpenCL platform 1 (NVIDIA Corporation):
The platform is not supported.
OpenCL initialized successfully.


Neat Video benchmark:

Frame Size: 1920x1080 progressive
Bitdepth: 32 bits per channel
Mix with Original: Disabled
Temporal Filter: Enabled
Quality Mode: Normal
Radius: 2 frames
Dust and Scratches: Disabled
Repeat Rate: 0% of repeated frames
Jitter Filtration: Normal
Spatial Filter: Enabled
Quality Mode: Normal
Frequencies: High, Mid, Low, Very Low
Artifact Removal: Enabled
Edge Smoothing: Disabled
Sharpening: Disabled


Detecting the best combination of performance settings:
running the test data set on up to 64 CPU cores and on up to 1 GPU

CPU Model: AMD Ryzen Threadripper 3970X 32-Core Processor
GPU 1: NVIDIA GeForce RTX 4090 (CUDA): 24563 MB total (23008 MB currently available), using up to 100%

CPU only (1 core): 3.76 frames/sec
CPU only (2 cores): 7.64 frames/sec
CPU only (3 cores): 10.1 frames/sec
CPU only (4 cores): 12.3 frames/sec
CPU only (5 cores): 13.7 frames/sec
CPU only (6 cores): 15 frames/sec
CPU only (7 cores): 16.2 frames/sec
CPU only (8 cores): 17 frames/sec
CPU only (9 cores): 17.5 frames/sec
CPU only (10 cores): 18.2 frames/sec
CPU only (11 cores): 18.6 frames/sec
CPU only (12 cores): 19.1 frames/sec
CPU only (13 cores): 19.3 frames/sec
CPU only (14 cores): 19.3 frames/sec
CPU only (15 cores): 19.6 frames/sec
CPU only (16 cores): 19.4 frames/sec
CPU only (17 cores): 19.7 frames/sec
CPU only (18 cores): 20.2 frames/sec
CPU only (19 cores): 20.1 frames/sec
CPU only (20 cores): 19.7 frames/sec
CPU only (21 cores): 19.8 frames/sec
CPU only (22 cores): 19.6 frames/sec
CPU only (23 cores): 19.6 frames/sec
CPU only (24 cores): 19.5 frames/sec
CPU only (25 cores): 19.5 frames/sec
CPU only (26 cores): 18.8 frames/sec
CPU only (27 cores): 18.9 frames/sec
CPU only (28 cores): 19.2 frames/sec
CPU only (29 cores): 19.2 frames/sec
CPU only (30 cores): 18.8 frames/sec
CPU only (31 cores): 17.8 frames/sec
CPU only (32 cores): 17.1 frames/sec
CPU only (33 cores): 16.7 frames/sec
CPU only (34 cores): 16.4 frames/sec
CPU only (35 cores): 15 frames/sec
CPU only (36 cores): 14.8 frames/sec
CPU only (37 cores): 14.7 frames/sec
CPU only (38 cores): 14.8 frames/sec
CPU only (39 cores): 13.8 frames/sec
CPU only (40 cores): 13.2 frames/sec
CPU only (41 cores): 12.4 frames/sec
CPU only (42 cores): 12.8 frames/sec
CPU only (43 cores): 11.5 frames/sec
CPU only (44 cores): 11.8 frames/sec
CPU only (45 cores): 11.1 frames/sec
CPU only (46 cores): 10.6 frames/sec
CPU only (47 cores): 10.5 frames/sec
CPU only (48 cores): 12.5 frames/sec
CPU only (49 cores): 13.1 frames/sec
CPU only (50 cores): 12.4 frames/sec
CPU only (51 cores): 11.5 frames/sec
CPU only (52 cores): 12.2 frames/sec
CPU only (53 cores): 12.3 frames/sec
CPU only (54 cores): 11 frames/sec
CPU only (55 cores): 7.85 frames/sec
CPU only (56 cores): 7.94 frames/sec
CPU only (57 cores): 7.68 frames/sec
CPU only (58 cores): 7.9 frames/sec
CPU only (59 cores): 7.82 frames/sec
CPU only (60 cores): 7.89 frames/sec
CPU only (61 cores): 7.79 frames/sec
CPU only (62 cores): 7.93 frames/sec
CPU only (63 cores): 7.85 frames/sec
CPU only (64 cores): 7.73 frames/sec
GPU only (NVIDIA GeForce RTX 4090): 81.3 frames/sec
CPU (2 cores) and GPU (NVIDIA GeForce RTX 4090): 19.3 frames/sec
CPU (3 cores) and GPU (NVIDIA GeForce RTX 4090): 25.1 frames/sec
CPU (4 cores) and GPU (NVIDIA GeForce RTX 4090): 24.5 frames/sec
CPU (5 cores) and GPU (NVIDIA GeForce RTX 4090): 29.6 frames/sec
CPU (6 cores) and GPU (NVIDIA GeForce RTX 4090): 25.5 frames/sec
CPU (7 cores) and GPU (NVIDIA GeForce RTX 4090): 27.4 frames/sec
CPU (8 cores) and GPU (NVIDIA GeForce RTX 4090): 28.2 frames/sec
CPU (9 cores) and GPU (NVIDIA GeForce RTX 4090): 28 frames/sec
CPU (10 cores) and GPU (NVIDIA GeForce RTX 4090): 27.9 frames/sec
CPU (11 cores) and GPU (NVIDIA GeForce RTX 4090): 26.7 frames/sec
CPU (12 cores) and GPU (NVIDIA GeForce RTX 4090): 27.7 frames/sec
CPU (13 cores) and GPU (NVIDIA GeForce RTX 4090): 29.1 frames/sec
CPU (14 cores) and GPU (NVIDIA GeForce RTX 4090): 28.6 frames/sec
CPU (15 cores) and GPU (NVIDIA GeForce RTX 4090): 28 frames/sec
CPU (16 cores) and GPU (NVIDIA GeForce RTX 4090): 27.5 frames/sec
CPU (17 cores) and GPU (NVIDIA GeForce RTX 4090): 27.3 frames/sec
CPU (18 cores) and GPU (NVIDIA GeForce RTX 4090): 26.4 frames/sec
CPU (19 cores) and GPU (NVIDIA GeForce RTX 4090): 26.9 frames/sec
CPU (20 cores) and GPU (NVIDIA GeForce RTX 4090): 26.2 frames/sec
CPU (21 cores) and GPU (NVIDIA GeForce RTX 4090): 26.1 frames/sec
CPU (22 cores) and GPU (NVIDIA GeForce RTX 4090): 26.1 frames/sec
CPU (23 cores) and GPU (NVIDIA GeForce RTX 4090): 25.3 frames/sec
CPU (24 cores) and GPU (NVIDIA GeForce RTX 4090): 25.4 frames/sec
CPU (25 cores) and GPU (NVIDIA GeForce RTX 4090): 25.3 frames/sec
CPU (26 cores) and GPU (NVIDIA GeForce RTX 4090): 25.3 frames/sec
CPU (27 cores) and GPU (NVIDIA GeForce RTX 4090): 24.3 frames/sec
CPU (28 cores) and GPU (NVIDIA GeForce RTX 4090): 23.6 frames/sec
CPU (29 cores) and GPU (NVIDIA GeForce RTX 4090): 24 frames/sec
CPU (30 cores) and GPU (NVIDIA GeForce RTX 4090): 23.8 frames/sec
CPU (31 cores) and GPU (NVIDIA GeForce RTX 4090): 23.1 frames/sec
CPU (32 cores) and GPU (NVIDIA GeForce RTX 4090): 21.2 frames/sec
CPU (33 cores) and GPU (NVIDIA GeForce RTX 4090): 21.4 frames/sec
CPU (34 cores) and GPU (NVIDIA GeForce RTX 4090): 21.5 frames/sec
CPU (35 cores) and GPU (NVIDIA GeForce RTX 4090): 20.3 frames/sec
CPU (36 cores) and GPU (NVIDIA GeForce RTX 4090): 20.6 frames/sec
CPU (37 cores) and GPU (NVIDIA GeForce RTX 4090): 20.4 frames/sec
CPU (38 cores) and GPU (NVIDIA GeForce RTX 4090): 19.9 frames/sec
CPU (39 cores) and GPU (NVIDIA GeForce RTX 4090): 19.5 frames/sec
CPU (40 cores) and GPU (NVIDIA GeForce RTX 4090): 17.2 frames/sec
CPU (41 cores) and GPU (NVIDIA GeForce RTX 4090): 18.4 frames/sec
CPU (42 cores) and GPU (NVIDIA GeForce RTX 4090): 17.6 frames/sec
CPU (43 cores) and GPU (NVIDIA GeForce RTX 4090): 17.5 frames/sec
CPU (44 cores) and GPU (NVIDIA GeForce RTX 4090): 18.4 frames/sec
CPU (45 cores) and GPU (NVIDIA GeForce RTX 4090): 16.6 frames/sec
CPU (46 cores) and GPU (NVIDIA GeForce RTX 4090): 16.4 frames/sec
CPU (47 cores) and GPU (NVIDIA GeForce RTX 4090): 17 frames/sec
CPU (48 cores) and GPU (NVIDIA GeForce RTX 4090): 16.9 frames/sec
CPU (49 cores) and GPU (NVIDIA GeForce RTX 4090): 17.1 frames/sec
CPU (50 cores) and GPU (NVIDIA GeForce RTX 4090): 18.2 frames/sec
CPU (51 cores) and GPU (NVIDIA GeForce RTX 4090): 17.4 frames/sec
CPU (52 cores) and GPU (NVIDIA GeForce RTX 4090): 18.2 frames/sec
CPU (53 cores) and GPU (NVIDIA GeForce RTX 4090): 16.9 frames/sec
CPU (54 cores) and GPU (NVIDIA GeForce RTX 4090): 16.5 frames/sec
CPU (55 cores) and GPU (NVIDIA GeForce RTX 4090): 15.8 frames/sec
CPU (56 cores) and GPU (NVIDIA GeForce RTX 4090): 15.2 frames/sec
CPU (57 cores) and GPU (NVIDIA GeForce RTX 4090): 15.2 frames/sec
CPU (58 cores) and GPU (NVIDIA GeForce RTX 4090): 15.1 frames/sec
CPU (59 cores) and GPU (NVIDIA GeForce RTX 4090): 14.9 frames/sec
CPU (60 cores) and GPU (NVIDIA GeForce RTX 4090): 14.3 frames/sec
CPU (61 cores) and GPU (NVIDIA GeForce RTX 4090): 14.1 frames/sec
CPU (62 cores) and GPU (NVIDIA GeForce RTX 4090): 13.6 frames/sec
CPU (63 cores) and GPU (NVIDIA GeForce RTX 4090): 13.5 frames/sec
CPU (64 cores) and GPU (NVIDIA GeForce RTX 4090): 13.1 frames/sec

Best combination: GPU only (NVIDIA GeForce RTX 4090): 81.3 frames/sec

[NVTeam]

What motherboard model do you use? Does it have any NUMA-related settings in BIOS/EEFI? It is possible that something is not set right there.

Please run the Geekbench 6 test and compare your results with those listed in their page.

Please check whether any background processes use the CPU when running the Optimize Settings test in Neat Video.

Thank you,
Vlad