Can you please post your system RAM specs, please? It is probably ddr5, what speed?
your benchmark results are indeed really poor with an i9-13900k
so i suggest that some hardware expert takes a deeper look into your system/bios settings.
it looks like the 13900k runs on one leg only, mabe the ram also
the cpu power limits and ram speed needs to be checked in bios asap.
saludos mike
I would be interested to know that is there much difference between 5800x3D and 5950x? Hard to find data on how much the processor affect the performance.
Not for most TVAI tasks. For example, I always limit my 7950 to 8 cores (all on CCD0) in TVAI tasks to achieve better performance in Proteus.
as mxrevolution pointed out, your system performs poorly in comparison with other system with this CPU+GPU.
Since the performance problems mostly affect 2X and 4X workloads in which the RAM plays an important role, it’s most likely a RAM speed problem. With faster RAM your 4X speeds could be up by 20%. If your TVAI tasks don’t involve rescaling (done on CPU+RAM) when u won’t see as much of an improvement by faster RAM.
For checking RAM performance, u can for example use
https://www.aida64.com/user-manual/benchmarks/cache-memory-benchmark
And for a quick and easy whole system performance comparison to similar systems:
https:://www.userbenchmark.com
Note: userbenchmark has some weird opinions/test results regarding AMD hardware but I find it still a quite useful tool because it compares your system with the same CPU/GPU combination which allows a quick analysis for some outliers in performance
Are you maybe running your bios in CSM (compatability) mode rather than UEFI-only mode?
Thanks! yeah, it looks like it is all about RAM almost certainly. So I’ll need both new ram and a mobo, dammit. I also start suspecting that X3D CPUs provide some notable advantage too.
No, if only it was this easy 
The cache of the X3D CPUs is not of much use for most TVAI tasks since TVAI constantly moves a lot of data (video images) through the RAM + CPU + GPU pipeline.
The cache is great if all or most of your important data which is constantly accessed fits into the cache, which for example, applies for most games.
Topaz Video AI v6.1.3
System Information
OS: Mac v15.04
CPU: Apple M2 Ultra 64 GB
GPU: Apple M2 Ultra 48 GB
Processing Settings
device: -2 vram: 1 instances: 1
Input Resolution: 1920x1080
Benchmark Results
Artemis 1X: 16.49 fps 2X: 07.32 fps 4X: 02.93 fps
Iris 1X: 10.22 fps 2X: 05.34 fps 4X: 02.30 fps
Proteus 1X: 17.39 fps 2X: 10.60 fps 4X: 03.05 fps
Gaia 1X: 05.01 fps 2X: 03.80 fps 4X: 02.80 fps
Nyx 1X: 05.97 fps 2X: 05.65 fps
Nyx Fast 1X: 17.21 fps
Rhea 4X: 01.67 fps
RXL 4X: 01.76 fps
Hyperion HDR 1X: 41.69 fps
4X Slowmo Apollo: 19.24 fps APFast: 63.10 fps Chronos: 06.25 fps CHFast: 07.76 fps
16X Slowmo Aion: 19.92 fps
Topaz Video AI Alpha v6.1.1.1.a.trt
System Information
OS: Windows v11.24
CPU: AMD Ryzen 9 9950X3D 16-Core Processor 61.613 GB
GPU: NVIDIA GeForce RTX 5090 31.349 GB
Processing Settings
device: 0 vram: 1 instances: 1
Input Resolution: 1920x1080
Benchmark Results
Artemis 1X: 46.52 fps 2X: 16.78 fps 4X: 04.64 fps
Iris 1X: 44.33 fps 2X: 20.00 fps 4X: 05.45 fps
Proteus 1X: 49.75 fps 2X: 22.45 fps 4X: 05.24 fps
Gaia 1X: 18.18 fps 2X: 13.18 fps 4X: 05.20 fps
Nyx 1X: 16.97 fps 2X: 12.50 fps
Nyx Fast 1X: 38.43 fps
Rhea 4X: 05.06 fps
RXL 4X: 04.92 fps
Hyperion HDR 1X: 31.72 fps
4X Slowmo Apollo: 44.79 fps APFast: 71.35 fps Chronos: 43.57 fps CHFast: 47.97 fps
16X Slowmo Aion: 35.02 fps
Topaz Video AI v6.1.3
System Information
OS: Windows v11.24
CPU: Intel(R) Core(TM) i9-14900K 63.691 GB
GPU: NVIDIA GeForce RTX 5080 15.517 GB
Processing Settings
device: 0 vram: 1 instances: 1
Input Resolution: 1920x1080
Benchmark Results
Artemis 1X: 24.58 fps 2X: 16.09 fps 4X: 04.28 fps
Iris 1X: 25.50 fps 2X: 14.97 fps 4X: 04.76 fps
Proteus 1X: 26.10 fps 2X: 18.50 fps 4X: 05.27 fps
Gaia 1X: 11.76 fps 2X: 07.79 fps 4X: 04.96 fps
Nyx 1X: 10.15 fps 2X: 08.01 fps
Nyx Fast 1X: 17.37 fps
Rhea 4X: 03.61 fps
RXL 4X: 03.35 fps
Hyperion HDR 1X: 27.24 fps
4X Slowmo Apollo: 27.25 fps APFast: 63.77 fps Chronos: 15.43 fps CHFast: 27.25 fps
16X Slowmo Aion: ERR fps
Would be great if someone can share the M4 Max and the M3 Ultra benchmarks.
Thank you!
1 Like
I’m waiting for that too! haha
Topaz Video AI v6.1.3
System Information
OS: Windows v10.22
CPU: AMD Ryzen Threadripper 3970X 32-Core Processor 127.87 GB
GPU: NVIDIA GeForce RTX 3090 23.756 GB
Processing Settings
device: -2 vram: 1 instances: 0
Input Resolution: 1920x1080
Benchmark Results
Artemis 1X: 24.44 fps 2X: 12.62 fps 4X: 03.72 fps
Iris 1X: 22.32 fps 2X: 14.22 fps 4X: 03.76 fps
Proteus 1X: 23.63 fps 2X: 14.72 fps 4X: 03.80 fps
Gaia 1X: 08.29 fps 2X: 05.77 fps 4X: 03.36 fps
Nyx 1X: 10.30 fps 2X: 08.12 fps
Nyx Fast 1X: 19.88 fps
Rhea 4X: 03.03 fps
RXL 4X: 02.93 fps
Hyperion HDR 1X: 21.60 fps
4X Slowmo Apollo: 33.33 fps APFast: 58.22 fps Chronos: 19.03 fps CHFast: 26.65 fps
16X Slowmo Aion: 25.07 fps
1 Like
Topaz Video AI v6.1.3
System Information
OS: Mac v14.03
CPU: Apple M3 Max 128 GB
GPU: Apple M3 Max 96 GB
Processing Settings
device: 0 vram: 1 instances: 1
Input Resolution: 3840x2160
Benchmark Results
Artemis 1X: 02.41 fps 2X: 01.45 fps 4X: 00.52 fps
Iris 1X: 02.09 fps 2X: 01.11 fps 4X: 00.30 fps
Proteus 1X: 01.95 fps 2X: 01.36 fps 4X: 00.44 fps
Gaia 1X: 00.64 fps 2X: 00.47 fps 4X: 00.37 fps
Nyx 1X: 00.87 fps 2X: 00.82 fps
Nyx Fast 1X: 01.96 fps
Rhea 4X: 00.19 fps
RXL 4X: 00.20 fps
Hyperion HDR 1X: 20.67 fps
4X Slowmo Apollo: 02.19 fps APFast: 10.22 fps Chronos: 00.96 fps CHFast: 01.47 fps
16X Slowmo Aion: 04.74 fps
Why do you have RAM set to 10%?
While this still can speed up things with SD material on AppleSilicon, HD will be faster with RAM at 100%.
Nice catch! Had it set at 10% to test something for a user yesterday and forgot to change it back! Rerunning and will edit post!
