Performance Benchmarking

suraj · March 10, 2023, 10:16pm

Hi All,

Starting with Topaz Video AI v3.1.9 you will be able to run the models against standard inputs to able to compare results across different machines. This option doesn’t transmit any data and is completely up to the user to share. Once benchmarking is finished you can copy and paste results here.

Example results, may not be accurate:

Topaz Video AI ALPHA v3.2.0.0.a
System Information
OS: Mac v13.0201
CPU: Apple M1 Max 64 GB
GPU: Apple M1 Max 48 GB
Settings: device: 0 vram: 1 instances: 1
Benchmark Results
Artemis 1X: 7.52 fps 2X: 5.41 fps 4X: 2.01 fps
Proteus 1X: 8.49 fps 2X: 5.58 fps 4X: 1.92 fps
Gaia 1X: 2.63 fps 2X: 1.82 fps 4X: 1.52 fps
4X Slowmo Apollo: 8.01 fps Chronos: 2.72 fps Chronos Fast: 4.26 fps

Running the benchmarks will download the corresponding models for 1080p input and store them in the models directory. If you do not use the above models and want to change other models. Feel free to modify the benchmarks.json file in the models directory.

Currently we use very little frames to get the fps numbers to keep the benchmark quick. If you see discrepancies between the benchmark numbers and your experience running the models let us know.
The benchmark doesn’t decode/encode or do any disk access so numbers would vary.

If you have any questions or issues let us know.

JakSpoon · March 10, 2023, 10:28pm

This is great! Thanks a lot!

avid-1676 · March 10, 2023, 11:58pm

Topaz Video AI Alpha v3.2.0.1.a
System Information
OS: Windows v11.2009
CPU: Intel(R) Core™ i7-8700 CPU @ 3.20GHz 63.826 GB
GPU: NVIDIA GeForce RTX 2070 7.8311 GB
GPU: Intel(R) UHD Graphics 630 0.125 GB
Settings: device: 0 vram: 1 instances: 0
Benchmark Results
Artemis 1X: 8.52 fps 2X: 5.42 fps 4X: 1.62 fps
Proteus 1X: 7.57 fps 2X: 4.92 fps 4X: 1.58 fps
Gaia 1X: 2.99 fps 2X: 1.99 fps 4X: 1.33 fps
4X Slowmo Apollo: 7.4 fps Chronos: 2.63 fps Chronos Fast: 11.35 fps

bnumerick · March 11, 2023, 12:46am

Topaz Video AI Alpha v3.2.0.1.a
System Information
OS: Mac v13.0201
CPU: Apple M1 Ultra 64 GB
GPU: Apple M1 Ultra 48 GB
Settings: device: 0 vram: 1 instances: 1
Benchmark Results
Artemis 1X: 13.13 fps 2X: 6.95 fps 4X: 2.53 fps
Proteus 1X: 12.7 fps 2X: 6.09 fps 4X: 2.44 fps
Gaia 1X: 4.7 fps 2X: 3.4 fps 4X: 2.45 fps
4X Slowmo Apollo: 8.34 fps Chronos: 3.62 fps Chronos Fast: 5.55 fps

Topaz Video AI Alpha v3.2.0.1.a
System Information
OS: Mac v13.0201
CPU: Apple M1 Max 32 GB
GPU: Apple M1 Max 21.333 GB
Settings: device: 0 vram: 1 instances: 1
Benchmark Results
Artemis 1X: 8.35 fps 2X: 5.41 fps 4X: 1.93 fps
Proteus 1X: 7.84 fps 2X: 5.52 fps 4X: 1.76 fps
Gaia 1X: 2.72 fps 2X: 1.84 fps 4X: 1.44 fps
4X Slowmo Apollo: 8.6 fps Chronos: 2.7 fps Chronos Fast: 4.2 fps

Topaz Video AI Alpha v3.2.0.1.a
System Information
OS: Mac v13.0201
CPU: Apple M1 Pro 16 GB
GPU: Apple M1 Pro 10.667 GB
Settings: device: 0 vram: 1 instances: 1
Benchmark Results
Artemis 1X: 5.72 fps 2X: 3.96 fps 4X: 0.83 fps
Proteus 1X: 5.67 fps 2X: 4 fps 4X: 1.4 fps
Gaia 1X: 1.98 fps 2X: 1.41 fps 4X: 1.04 fps
4X Slowmo Apollo: 4.47 fps Chronos: 1.73 fps Chronos Fast: 2.84 fps

Topaz Video AI Alpha v3.2.0.1.a

System Information
OS: Mac v13.02
CPU: Apple M1 8 GB
GPU: Apple M1 5.3333 GB
Settings: device: 0 vram: 1 instances: 1
Benchmark Results
Artemis 1X: 3.68 fps 2X: 2.46 fps 4X: 0.35 fps
Proteus 1X: 3.75 fps 2X: 2.5 fps 4X: 0.27 fps
Gaia 1X: 1.21 fps 2X: 0.89 fps 4X: 0.65 fps
4X Slowmo Apollo: 4.28 fps Chronos: 1.19 fps Chronos Fast: 1.91 fps

Topaz Video AI Alpha v3.2.0.1.a
System Information
OS: Mac v13.01
CPU: Apple M1 16 GB
GPU: Apple M1 10.667 GB
Settings: device: 0 vram: 1 instances: 1
Benchmark Results
Artemis 1X: 3.3 fps 2X: 2.15 fps 4X: 0.92 fps
Proteus 1X: 3.34 fps 2X: 2.23 fps 4X: 0.93 fps
Gaia 1X: 1.07 fps 2X: 0.81 fps 4X: 0.6 fps
4X Slowmo Apollo: 3.73 fps Chronos: 1.02 fps Chronos Fast: 1.75 fps

Topaz Video AI Alpha v3.2.0.1.a
System Information
OS: Mac v13.0201
CPU: Apple M2 Max 32 GB
GPU: Apple M2 Max 21.333 GB
Settings: device: 0 vram: 1 instances: 1
Benchmark Results
Artemis 1X: 11.1 fps 2X: 6.9 fps 4X: 2.35 fps
Proteus 1X: 10.94 fps 2X: 7 fps 4X: 2.25 fps
Gaia 1X: 3.38 fps 2X: 2.46 fps 4X: 1.81 fps
4X Slowmo Apollo: 10.44 fps Chronos: 3.24 fps Chronos Fast: 5.24 fps

ibobbyts · March 11, 2023, 12:56am

Here’s a function I think would be helpful when using multiple models
Input a video and set the output config, the program will detect which model is the bottleneck and we can (or the program automatically decide) which model to add extra instance.

michael.carter-7687 · March 11, 2023, 12:58am

Topaz Video AI Alpha v3.2.0.1.a
System Information
OS: Windows v11.2009
CPU: 13th Gen Intel(R) Core™ i9-13900K 31.747 GB
GPU: NVIDIA GeForce RTX 4090 23.59 GB
Settings: device: 0 vram: 1 instances: 0
Benchmark Results
Artemis 1X: 32.48 fps 2X: 15.92 fps 4X: 4.74 fps
Proteus 1X: 26.99 fps 2X: 14.46 fps 4X: 4.52 fps
Gaia 1X: 13.86 fps 2X: 9.11 fps 4X: 4.82 fps
4X Slowmo Apollo: 26.5 fps Chronos: 28.23 fps Chronos Fast: 36.35 fps

twpotato606 · March 11, 2023, 2:02am

Topaz Video AI Beta v3.1.9.0.b
System Information
OS: Windows v10.2009
CPU: AMD Ryzen 5 5600X 6-Core Processor 31.914 GB
GPU: NVIDIA GeForce RTX 3090 23.77 GB
Settings: device: 0 vram: 1 instances: 1
Benchmark Results
Artemis 1X: 14.29 fps 2X: 6.04 fps 4X: 1.47 fps
Proteus 1X: 10.94 fps 2X: 4.83 fps 4X: 1.44 fps
Gaia 1X: 7.43 fps 2X: 4.81 fps 4X: 1.97 fps
4X Slowmo Apollo: 14.23 fps Chronos: 14.79 fps Chronos Fast: 19.65 fps

Topaz Video AI Alpha v3.2.0.1.a
System Information
OS: Windows v10.2009
CPU: AMD Ryzen 5 5600X 6-Core Processor 31.914 GB
GPU: NVIDIA GeForce RTX 3090 23.77 GB
Settings: device: 0 vram: 1 instances: 0
Benchmark Results
Artemis 1X: 17 fps 2X: 7.44 fps 4X: 2.04 fps
Proteus 1X: 14.29 fps 2X: 6.83 fps 4X: 1.95 fps
Gaia 1X: 7.17 fps 2X: 4.74 fps 4X: 2.29 fps
4X Slowmo Apollo: 14.25 fps Chronos: 14.73 fps Chronos Fast: 20.33 fps

m75 · March 11, 2023, 2:03am

Topaz Video AI Alpha v3.2.0.1.a
System Information
OS: Windows v11.2009
CPU: Intel(R) Core™ i7-10870H CPU @ 2.20GHz 31.757 GB
GPU: NVIDIA GeForce RTX 3070 Laptop GPU 7.8447 GB
GPU: Intel(R) UHD Graphics 0.125 GB
Settings: device: 0 vram: 1 instances: 0
Benchmark Results
Artemis 1X: 7.73 fps 2X: 4.88 fps 4X: 1.37 fps
Proteus 1X: 7.09 fps 2X: 4.47 fps 4X: ERR fps
Gaia 1X: 2.73 fps 2X: 1.88 fps 4X: 1.2 fps
4X Slowmo Apollo: 7.67 fps Chronos: 6.23 fps Chronos Fast: 10.3 fps

bnumerick · March 11, 2023, 2:14am

Topaz Video AI Alpha v3.2.0.1.a
System Information
OS: Windows v10.2009
CPU: Intel(R) Core™ i7-9700F CPU @ 3.00GHz 31.918 GB
GPU: NVIDIA GeForce RTX 3080 Ti 11.816 GB
Settings: device: 0 vram: 1 instances: 0
Benchmark Results
Artemis 1X: 12.89 fps 2X: 6.37 fps 4X: 2.1 fps
Proteus 1X: 12.25 fps 2X: 6.45 fps 4X: 1.97 fps
Gaia 1X: 6.16 fps 2X: 4.13 fps 4X: 2.28 fps
4X Slowmo Apollo: 14.35 fps Chronos: 14.55 fps Chronos Fast: 20.2 fps

20rushtonj · March 11, 2023, 2:47am

RTX 4090

Topaz Video AI Beta v3.1.9.0.b

System Information
OS: Windows v11.2009
CPU: AMD Ryzen 9 5950X 16-Core Processor              63.928 GB
GPU: NVIDIA GeForce RTX 4090  23.59 GB
Settings: device: 0 vram: 1 instances: 1
Benchmark Results
Artemis		1X: 19.75 fps 	2X: 7.76 fps 	4X: 1.93 fps 	
Proteus		1X: 14.37 fps 	2X: 5.77 fps 	4X: 1.87 fps 	
Gaia		1X: 13.4 fps 	2X: 8.21 fps 	4X: 2.53 fps 	
4X Slowmo		Apollo: 16.94 fps 	Chronos: 25.18 fps 	Chronos Fast: 25.09 fps

Topaz Video AI Alpha v3.2.0.1.a

System Information
OS: Windows v11.2009
CPU: AMD Ryzen 9 5950X 16-Core Processor              63.928 GB
GPU: NVIDIA GeForce RTX 4090  23.59 GB
Settings: device: 0 vram: 1 instances: 0
Benchmark Results
Artemis		1X: 25.73 fps 	2X: 11 fps 	4X: 2.74 fps 	
Proteus		1X: 21.37 fps 	2X: 9.23 fps 	4X: 2.6 fps 	
Gaia		1X: 13.34 fps 	2X: 8.49 fps 	4X: 3.36 fps 	
4X Slowmo		Apollo: 17.69 fps 	Chronos: 28.18 fps 	Chronos Fast: 30.6 fps

M1 Pro, 14 core GPU

Notes:

This is in a MacBook Pro Chassis, the MacBook was plugged in, and it was in “normal power” mode.
The MacBook was allowed 5 minutes to cool down between testing each version.

Topaz Video AI Beta v3.1.9.0.b

System Information
OS: Mac v13.0201
CPU: Apple M1 Pro  16 GB
GPU: Apple M1 Pro  10.667 GB
Settings: device: 0 vram: 1 instances: 1
Benchmark Results
Artemis		1X: 5.69 fps 	2X: 4.03 fps 	4X: 1.55 fps 	
Proteus		1X: 5.6 fps 	2X: 3.84 fps 	4X: 1.55 fps 	
Gaia		1X: 1.86 fps 	2X: 1.38 fps 	4X: 1.03 fps 	
4X Slowmo		Apollo: 4.44 fps 	Chronos: 1.84 fps 	Chronos Fast: 2.93 fps

Topaz Video AI Alpha v3.2.0.1.a

System Information
OS: Mac v13.0201
CPU: Apple M1 Pro  16 GB
GPU: Apple M1 Pro  10.667 GB
Settings: device: 0 vram: 1 instances: 1
Benchmark Results
Artemis		1X: 5.71 fps 	2X: 4.02 fps 	4X: 1.58 fps 	
Proteus		1X: 5.64 fps 	2X: 3.92 fps 	4X: 1.6 fps 	
Gaia		1X: 1.86 fps 	2X: 1.35 fps 	4X: 1.02 fps 	
4X Slowmo		Apollo: 4.44 fps 	Chronos: 1.83 fps 	Chronos Fast: 2.92 fps

david.123 · March 11, 2023, 3:24am

Fantastic feature! Finally we don’t have to wonder about the specifics of people’s testing!

I noticed one thing, on Windows 11, the Gaia results don’t line up in neat columns on the GUI like the rest of the results.

Edit: I have noticed that the lining up in columns issue on the GUI is not specific to Gaia results as I see on other screenshots that other rows sometimes are also affected.

david.123 · March 11, 2023, 3:27am

Topaz Video AI Alpha  v3.2.0.1.a
System Information
OS: Windows v11.2009
CPU: Intel(R) Core(TM) i7-6700K CPU @ 4.00GHz  15.959 GB
GPU: NVIDIA GeForce GTX 1070  7.8916 GB
Settings: device: 0 vram: 1 instances: 0
Benchmark Results
Artemis		1X: 2.43 fps 	2X: 1.64 fps 	4X: 0.65 fps 	
Proteus		1X: 2.31 fps 	2X: 1.59 fps 	4X: 0.64 fps 	
Gaia		1X: 0.9 fps 	2X: 0.61 fps 	4X: 0.06 fps 	
4X Slowmo		Apollo: 3.37 fps 	Chronos: 1.34 fps 	Chronos Fast: 2.24 fps

Martyprod · March 11, 2023, 4:01am

trying to not hate anybody here but it’s not easy…

Topaz Video AI Alpha  v3.2.0.1.a
System Information
OS: Windows v11.2009
CPU: Intel(R) Core(TM) i7-7700 CPU @ 3.60GHz  31.895 GB
GPU: NVIDIA GeForce GTX 1050 Ti  3.9229 GB
GPU: Intel(R) HD Graphics 630  0.125 GB
Settings: device: 0 vram: 1 instances: 0
Benchmark Results
Artemis		1X: 0.88 fps 	2X: 0.61 fps 	4X: 0.24 fps 	
Proteus		1X: 0.86 fps 	2X: 0.6 fps 	4X: 0.24 fps 	
Gaia		1X: 0.3 fps 	2X: 0.04 fps 	4X: 0.03 fps 	
4X Slowmo		Apollo: 0.38 fps 	Chronos: 0.49 fps 	Chronos Fast: 0.36 fps

lhkjacky · March 11, 2023, 4:22am

michael.carter-7687:

Topaz Video AI Alpha v3.2.0.1.a
System Information
OS: Windows v11.2009
CPU: 13th Gen Intel(R) Core™ i9-13900K 31.747 GB
GPU: NVIDIA GeForce RTX 4090 23.59 GB
Settings: device: 0 vram: 1 instances: 0
Benchmark Results
Artemis 	1X: 32.48 fps 	2X: 15.92 fps 	4X: 4.74 fps
Proteus 	1X: 26.99 fps 	2X: 14.46 fps 	4X: 4.52 fps
Gaia 		1X: 13.86 fps 	2X: 9.11 fps 	4X: 4.82 fps
4X Slowmo Apollo: 26.5 fps 	Chronos: 28.23 fps 	Chronos Fast: 36.35 fps

20rushtonj:

Topaz Video AI Alpha v3.2.0.1.a
System Information
OS: Windows v11.2009
CPU: AMD Ryzen 9 5950X 16-Core Processor 63.928 GB
GPU: NVIDIA GeForce RTX 4090  23.59 GB
Settings: device: 0 vram: 1 instances: 0
Benchmark Results
Artemis		1X: 25.73 fps 	2X: 11 fps 		4X: 2.74 fps 	
Proteus		1X: 21.37 fps 	2X: 9.23 fps 	4X: 2.6 fps 	
Gaia		1X: 13.34 fps 	2X: 8.49 fps 	4X: 3.36 fps 	
4X Slowmo Apollo: 17.69 fps 	Chronos: 28.18 fps 	Chronos Fast: 30.6 fps

Thanks all for the data.
It seem that CPU make a huge difference in Artemis and Proteus.

michael.carter-7687 · March 11, 2023, 5:47am

Yeah, seems like. For what it’s worth, I also specifically have DDR5-7200 memory, so a lot of memory bandwidth available.

david.123 · March 11, 2023, 6:57am

So let’s see your results!

david.123 · March 11, 2023, 6:58am

You: last, me: second last.

karol.guzik · March 11, 2023, 8:32am

Topaz Video AI Alpha v3.2.0.1.a
System Information
OS: Windows v11.2009
CPU: AMD Ryzen 9 7950X 16-Core Processor (eco mode 65W) 31.118 GB
GPU: NVIDIA GeForce RTX 4090 23.59 GB (60% power limit)
Settings: device: 0 vram: 0.9 instances: 0
Benchmark Results
Artemis 1X: 32.76 fps 2X: 15.98 fps 4X: 4.44 fps
Proteus 1X: 28.56 fps 2X: 14.52 fps 4X: 4.35 fps
Gaia 1X: 15.15 fps 2X: 9.86 fps 4X: 5.06 fps
4X Slowmo Apollo: 23.22 fps Chronos: 29.95 fps Chronos Fast: 37.26 fps

TPX · March 11, 2023, 8:57am

How the 4090 erodes through here.

@suraj i had to force it to download the models by run them before benchmarking.

I try now a uninstaller software to wipe out everything from TL, last try.

tony11 · March 11, 2023, 9:10am

Topaz Video AI Alpha v3.2.0.1.a
System Information
OS: Windows v10.2009
CPU: AMD Ryzen 5 2600X Six-Core Processor 31.927 GB
GPU: NVIDIA GeForce RTX 3080 Ti 11.816 GB
Settings: device: 0 vram: 1 instances: 0
Benchmark Results
Artemis 1X: 6.75 fps 2X: 3.53 fps 4X: 1.02 fps
Proteus 1X: 5.96 fps 2X: 3.05 fps 4X: 0.97 fps
Gaia 1X: 2.96 fps 2X: 2.24 fps 4X: 1.13 fps
4X Slowmo Apollo: 7.02 fps Chronos: 7.76 fps Chronos Fast: 8.38 fps

Topaz AI running behind v3.1.8