Video AI 5.3.X - User Benchmarking Results

Topaz Video AI  v5.3.6
System Information
OS: Windows v11.23
CPU: 12th Gen Intel(R) Core(TM) i5-12500  63.745 GB
GPU: NVIDIA GeForce RTX 4060  7.7705 GB
Processing Settings
device: 0 vram: 0.9 instances: 1
Input Resolution: 1920x1080
Benchmark Results
Artemis		1X: 	10.63 fps 	2X: 	07.15 fps 	4X: 	02.30 fps 	
Iris		1X: 	10.78 fps 	2X: 	06.31 fps 	4X: 	01.76 fps 	
Proteus		1X: 	10.16 fps 	2X: 	07.40 fps 	4X: 	02.50 fps 	
Gaia		1X: 	03.40 fps 	2X: 	02.27 fps 	4X: 	01.55 fps 	
Nyx		1X: 	04.10 fps 	2X: 	03.54 fps 	
Nyx Fast		1X: 	07.51 fps 	
Rhea		4X: 	01.35 fps 	
4X Slowmo		Apollo: 	15.23 fps 	APFast: 	40.59 fps 	Chronos: 	08.08 fps 	CHFast: 	12.51 fps 	
16X Slowmo		Aion: 	ERR fps 	

Topaz Video AI  v5.3.6
System Information
OS: Windows v11.23
CPU: Intel(R) Xeon(R) w5-2465X  127.25 GB
GPU: NVIDIA GeForce RTX 4070  11.73 GB
GPU: NVIDIA RTX A4000  15.79 GB
Processing Settings
device: 0 vram: 0.9 instances: 1
Input Resolution: 1920x1080
Benchmark Results
Artemis		1X: 	19.81 fps 	2X: 	12.68 fps 	4X: 	03.85 fps 	
Iris		1X: 	18.66 fps 	2X: 	11.37 fps 	4X: 	03.25 fps 	
Proteus		1X: 	17.89 fps 	2X: 	12.28 fps 	4X: 	04.29 fps 	
Gaia		1X: 	06.16 fps 	2X: 	04.24 fps 	4X: 	02.89 fps 	
Nyx		1X: 	07.24 fps 	2X: 	06.13 fps 	
Nyx Fast		1X: 	13.49 fps 	
Rhea		4X: 	02.35 fps 	
4X Slowmo		Apollo: 	24.50 fps 	APFast: 	54.48 fps 	Chronos: 	13.12 fps 	CHFast: 	20.34 fps 	
16X Slowmo		Aion: 	ERR fps 	

Thank you

Topaz Video AI  v5.3.6
System Information
OS: Mac v15.01
CPU: Apple M2 Max  96 GB
GPU: Apple M2 Max  72 GB
Processing Settings
device: -2 vram: 1 instances: 1
Input Resolution: 1920x1080
Benchmark Results
Artemis		1X: 	10.10 fps 	2X: 	06.41 fps 	4X: 	02.13 fps 	
Iris		1X: 	14.05 fps 	2X: 	04.06 fps 	4X: 	01.36 fps 	
Proteus		1X: 	12.34 fps 	2X: 	08.22 fps 	4X: 	01.80 fps 	
Gaia		1X: 	04.10 fps 	2X: 	02.79 fps 	4X: 	01.84 fps 	
Nyx		1X: 	04.72 fps 	2X: 	04.45 fps 	
Nyx Fast		1X: 	11.20 fps 	
Rhea		4X: 	00.85 fps 	
4X Slowmo		Apollo: 	12.38 fps 	APFast: 	40.36 fps 	Chronos: 	03.89 fps 	CHFast: 	05.79 fps 	
16X Slowmo		Aion: 	13.05 fps 	

Topaz Video AI  v5.3.6
System Information
OS: Windows v11.24
CPU: AMD Ryzen 9 3950X 16-Core Processor              31.928 GB
GPU: NVIDIA GeForce RTX 4090  23.576 GB
Processing Settings
device: 0 vram: 1 instances: 1
Input Resolution: 1920x1080
Benchmark Results
Artemis		1X: 	32.38 fps 	2X: 	13.60 fps 	4X: 	03.13 fps 	
Iris		1X: 	34.00 fps 	2X: 	15.16 fps 	4X: 	03.87 fps 	
Proteus		1X: 	40.56 fps 	2X: 	15.37 fps 	4X: 	04.30 fps 	
Gaia		1X: 	16.24 fps 	2X: 	10.78 fps 	4X: 	03.53 fps 	
Nyx		1X: 	18.20 fps 	2X: 	13.52 fps 	
Nyx Fast		1X: 	32.46 fps 	
Rhea		4X: 	03.52 fps 	
4X Slowmo		Apollo: 	34.53 fps 	APFast: 	52.25 fps 	Chronos: 	32.77 fps 	CHFast: 	29.08 fps 	
16X Slowmo		Aion: 	30.76 fps 	

I noticed your Mac version has been affected by a previous bug which sets the processing device to ā€œautoā€ (shown as -2 in the benchmark) without you knowing. If you are only working with HD sources, then you can leave it as is. However, if you are upscaling SD sources then processing could take much longer than necessary.

Although Windows and Intel Mac users can change the processor in the TVAI UI, Apple Silicon users canā€™t. So currently the only solutions are to either use a binary Plist editor to edit your preferences file or otherwise delete the file if you are willing to set your preferences again.

EDIT: There is another option - use the terminal to run plutil to convert the binary plist to xml:

cd ~/library/preferences
plutil -convert xml1 ā€˜com.topazlabs.topaz video ai.plistā€™

You can then edit the plist file in a text editor. Search for ā€œGPUDeviceIDā€ and youā€™ll see its value set to -2. Change to 0 and save. Use plutil to convert back to binary:

plutil -convert binary1 ā€˜com.topazlabs.topaz video ai.plistā€™

Regards

Andy

1 Like

Windows has defaulted to -2 as wellā€”but not being able to change it easily, seems mean.
I only noticed because Proteus 4X randomly took forever and got a score of 0.21 fps.

Thanks much @Chamberhouse
Fixed it now:

Topaz Video AI  v5.3.6
System Information
OS: Mac v15.01
CPU: Apple M2 Max  96 GB
GPU: Apple M2 Max  72 GB
Processing Settings
device: 0 vram: 1 instances: 1
Input Resolution: 1920x1080
Benchmark Results
Artemis		1X: 	12.00 fps 	2X: 	07.05 fps 	4X: 	02.27 fps 	
Iris		1X: 	07.11 fps 	2X: 	04.01 fps 	4X: 	01.58 fps 	
Proteus		1X: 	11.10 fps 	2X: 	07.92 fps 	4X: 	02.49 fps 	
Gaia		1X: 	03.32 fps 	2X: 	02.31 fps 	4X: 	01.68 fps 	
Nyx		1X: 	04.54 fps 	2X: 	04.08 fps 	
Nyx Fast		1X: 	10.78 fps 	
Rhea		4X: 	00.90 fps 	
4X Slowmo		Apollo: 	12.89 fps 	APFast: 	44.60 fps 	Chronos: 	03.75 fps 	CHFast: 	06.05 fps 	
16X Slowmo		Aion: 	12.93 fps 	

Youā€™re most welcome.

Regards

Andy

Topaz Video AI  v5.3.6
System Information
OS: Windows v11.23
CPU: 13th Gen Intel(R) Core(TM) i7-13700K  31.747 GB
GPU: NVIDIA GeForce RTX 4070 Ti  11.715 GB
GPU: Intel(R) UHD Graphics 770  0.125 GB
Processing Settings
device: 0 vram: 1 instances: 1
Input Resolution: 1920x1080
Benchmark Results
Artemis		1X: 	26.34 fps 	2X: 	17.74 fps 	4X: 	05.03 fps 	
Iris		1X: 	26.89 fps 	2X: 	16.19 fps 	4X: 	04.07 fps 	
Proteus		1X: 	24.66 fps 	2X: 	15.74 fps 	4X: 	05.32 fps 	
Gaia		1X: 	08.78 fps 	2X: 	06.05 fps 	4X: 	03.97 fps 	
Nyx		1X: 	10.32 fps 	2X: 	08.65 fps 	
Nyx Fast		1X: 	18.27 fps 	
Rhea		4X: 	03.39 fps 	
4X Slowmo		Apollo: 	34.92 fps 	APFast: 	71.89 fps 	Chronos: 	19.34 fps 	CHFast: 	29.22 fps 	
16X Slowmo		Aion: 	32.40 fps 	

Topaz Video AI  v5.3.6
System Information
OS: Windows v11.23
CPU: 13th Gen Intel(R) Core(TM) i7-13700K  31.747 GB
GPU: NVIDIA GeForce RTX 4070 Ti  11.715 GB
GPU: Intel(R) UHD Graphics 770  0.125 GB
Processing Settings
device: 0 vram: 1 instances: 1
Input Resolution: 3840x2160
Benchmark Results
Artemis		1X: 	05.61 fps 	2X: 	03.63 fps 	4X: 	01.00 fps 	
Iris		1X: 	05.71 fps 	2X: 	03.39 fps 	4X: 	00.87 fps 	
Proteus		1X: 	05.25 fps 	2X: 	03.40 fps 	4X: 	01.14 fps 	
Gaia		1X: 	01.88 fps 	2X: 	01.29 fps 	4X: 	00.84 fps 	
Nyx		1X: 	01.71 fps 	2X: 	02.14 fps 	
Nyx Fast		1X: 	03.02 fps 	
Rhea		4X: 	00.73 fps 	
4X Slowmo		Apollo: 	09.83 fps 	APFast: 	23.97 fps 	Chronos: 	04.28 fps 	CHFast: 	07.93 fps 	
16X Slowmo		Aion: 	ERR fps 	

In practical terms, it took 7 hours 5 minutes to take a 2 hour 4k resolution video at 23.97fps encoded with H265 Main 10 at 16Mb/sec bit rate, and enhance it with Proteus and encoding it to AV1 10bit at 12Mb/sec bitrate.

CPU usage ranges around 83-89% and obviously GPU is at 90-100%.

Hello.

Which driver are you guys utilizing?

It seems to me that many of you with an RTX3080/RTX RTX4070;RTX3090/4070Ti (SUPER) up to a RTX4090 should be getting at least an extra 2 up to 6fps more that whatā€™s in here.

I have 566.03 nvidia game drivers. I usually use NVCleaninstall to install the bare minimum components necessary. Is there a difference in using the Nvidia Studio drivers?

Oh, alright.

No. GRD & Studio drivers are the exact same today.

Always the latest Studio drivers for me. The result can vary from run to run a bit and can also depend on memory tunings/speed and some slight CPU overclocking.

Itā€™s never idea to OC inference workloads. Sure. Things can give the ILLUSION a bump in performance but often the negatives always outweighs the positives.

Always keep an eye on those metricsā€¦ ALWAYS!

Has anyone here got the new Intel 24 Core Ultra 9 285K?

I would like to see some benchmarks with it using a 4080 or 4090 to determine whether getting it is worth it over the 14th gen.

I heard itā€™s bad for gaming but good for creative workflows, canā€™t seem to any benchmarks when it comes to Topaz.

Thanks.

Should be interesting when paired with tuned CUDIMM 8000 RAM for scaling tasks.

RAM latency is worse than on previous Intel generations due to the chiplet design (which is the primary reason gaming performance suffers) but support for faster RAM combined with generational IPC improvements should bring it ahead in productivity.

Canā€™t wait to see some benchmarks from machines with M4 Pro and M4 MAX CPUs

Im curious, how would those insanely expense AMD Threadripper Pro with 96 cores or Intel Xeon processors handle encoding, upscaling, and etc?

It will be one of the first things Iā€™ll be doing when I get my M4 Pro Mini tomorrow. On the one hand, Iā€™m not getting my hopes too high - Iā€™m afraid Topaz donā€™t seem to be able to get the most out of Apple Silicon, especially when it comes to upscaling standard definition. However, on the other hand, Iā€™m really hoping that Iā€™ll be able to at last upscale SD in real-time using Artemis 2x - but probably not Iris 2x. Perhaps one day Topaz will finally sort out Iris on Apple Silicon and get it up to around Artemis and Proteus speeds.

Andy

Looking forward to, thanks, here is my chart pulled from Apple website, only for 14 inches model, the test Apple provided was slow motion x2 on 1920x800 not saying which model, so itā€™s quite unclear if that can be equivalent to upscale model, very rough comparisons.

I doubt I will eventually get any mac with M4 chip cuz I have a 14700k + 4090FE + 16GBX2 DDR5-6400Mhz PC, the potential performance might not catch up that build but, it might be better from my M1 Pro MBP 14 inch base model.