The GPU should be able to run at full load as long as it’s not overheating. Your factiry thermal paste may have dried-out and caused elevated temps, but even then, the GPU should back off the clocks until temps come down. Topaz may have exposed a flaw or defect, but it shouldn’t damage properly’working hardware.
Wrong a 9950X easily fills a 4090
Its entirely possible the thermal paste dried out. I’m not a heavy user of this particular system, its only been used sporadically since 2018. I’m getting fairly low temps, anywhere from 57 at normal and 75 under load, I just started with the artifacting maybe a week or so ago after extended use. I left my computer on for several days last week (not in sleep mode), though not rendering or anything.
It was about a i9-7900X which might not be able to feed a 4070 when upscaling is involved.
Also, it could be useful to have a dedicated hardware thread (also for configuration recommendations) here on topazlabs for these discussions because they are a little offtopic from the benchmarks.
sorry about that, I figured the benchmarks were an indicator whether I was getting good framerates or not and whether my GPU was failing. That’s the only reason I posted it here. Its behaving erratically.
That’s what I’m afraid I did as well. What did you end up replacing it with?
4070 is where the 12gb vram buffer starts in Nvidia’s current linup. Below that are offerings with less vram than your 1080Ti. The 4070 Super is a better value in that the performance jump is bigger than price jump, but even a regular 4070 will do fine. Also, artifacting is usually associated with vram failure. The vram temps may be available in monitoring software (I don’t remember if it was available on my 1080ti), but the gpu temp doesn’t reflect how hot the vram is getting.
AMD 7950X 16 Core CPU comparison with using all 16 cores vs only 8 cores on CCD0 (faster of 2 CCDs) vs only 8 cores from CCD1
Most tasks are actually faster when only using 8 cores from the same CCD than using all 16 cores from both CCDs. Especially the tasks involving image scaling are about 10% faster with only 8 cores on same CCD which indicates that there is more RAM/CPU cache intensive work performed than the tassks without scaling.
(CCD = chiplet with 8 cores each which are both integrated to form the 16 core CPU)
16 cores (mostly slower than only 8 cores from a single CCD)
Topaz Video AI Alpha v5.3.0.1.a.hyp.rxl
System Information
OS: Windows v10.22
CPU: AMD Ryzen 9 7950X 16-Core Processor 31.118 GB
GPU: NVIDIA GeForce RTX 4090 23.576 GB
GPU: AMD Radeon(TM) Graphics 0.47445 GB
Processing Settings
device: 0 vram: 1 instances: 0
Input Resolution: 1920x1080
Benchmark Results
Artemis 1X: 41.97 fps 2X: 17.27 fps 4X: 05.02 fps
Iris 1X: 41.76 fps 2X: 20.55 fps 4X: 05.67 fps
Proteus 1X: 41.18 fps 2X: 20.79 fps 4X: 05.82 fps
Gaia 1X: 14.57 fps 2X: 09.63 fps 4X: 05.15 fps
Nyx 1X: 17.10 fps 2X: 15.10 fps
Nyx Fast 1X: 32.39 fps
Rhea 4X: 05.32 fps
4X Slowmo Apollo: 44.11 fps APFast: 90.15 fps Chronos: 32.32 fps CHFast: 40.06 fps
16X Slowmo Aion: 37.10 fps
CCD0 (first 8 cores which are a little bit faster on AMD CPUs)
Topaz Video AI Alpha v5.3.0.1.a.hyp.rxl
System Information
OS: Windows v10.22
CPU: AMD Ryzen 9 7950X 16-Core Processor 31.118 GB
GPU: NVIDIA GeForce RTX 4090 23.576 GB
GPU: AMD Radeon(TM) Graphics 0.47445 GB
Processing Settings
device: 0 vram: 1 instances: 0
Input Resolution: 1920x1080
Benchmark Results
Artemis 1X: 41.01 fps 2X: 19.24 fps 4X: 05.19 fps
Iris 1X: 42.58 fps 2X: 22.68 fps 4X: 06.24 fps
Proteus 1X: 40.98 fps 2X: 22.54 fps 4X: 06.27 fps
Gaia 1X: 14.47 fps 2X: 09.65 fps 4X: 05.75 fps
Nyx 1X: 17.15 fps 2X: 15.41 fps
Nyx Fast 1X: 32.41 fps
Rhea 4X: 05.63 fps
4X Slowmo Apollo: 44.09 fps APFast: 91.22 fps Chronos: 32.26 fps CHFast: 39.83 fps
16X Slowmo Aion: 36.79 fps
CCD1 (second 8 cores which are a little bit slower on AMD CPUs)
Topaz Video AI Alpha v5.3.0.1.a.hyp.rxl
System Information
OS: Windows v10.22
CPU: AMD Ryzen 9 7950X 16-Core Processor 31.118 GB
GPU: NVIDIA GeForce RTX 4090 23.576 GB
GPU: AMD Radeon(TM) Graphics 0.47445 GB
Processing Settings
device: 0 vram: 1 instances: 0
Input Resolution: 1920x1080
Benchmark Results
Artemis 1X: 42.01 fps 2X: 20.59 fps 4X: 04.93 fps
Iris 1X: 41.79 fps 2X: 22.07 fps 4X: 05.87 fps
Proteus 1X: 41.19 fps 2X: 22.55 fps 4X: 05.95 fps
Gaia 1X: 14.46 fps 2X: 09.55 fps 4X: 05.41 fps
Nyx 1X: 17.14 fps 2X: 15.82 fps
Nyx Fast 1X: 32.27 fps
Rhea 4X: 05.62 fps
4X Slowmo Apollo: 43.40 fps APFast: 90.50 fps Chronos: 32.17 fps CHFast: 38.83 fps
16X Slowmo Aion: 36.25 fps
1 Like
perfect that’s what I needed to know, thanks everyone for indulging this tangent.
1 Like
Hello.
I don’t have to remind you of this, Michael… but time… TIME… is money right?
Your rig is your personal money-maker, right, Michael?
Most of us just love holding on to our legacy… uhm… stuff… BUuTtT… what does that mean for you again, Michael for your PROFESSIONAL workloads again…
I think the benchmark only runs one instance of the model it’s benchmarking. I suspect that rendering 2 or 3 videos at once could benefit from increase core count. -Assuming the GPU isn’t allready fully utilized.
Topaz Video AI v5.3.2
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: 18.21 fps 2X: 12.71 fps 4X: 04.01 fps
Iris 1X: 19.23 fps 2X: 11.66 fps 4X: 03.28 fps
Proteus 1X: 17.92 fps 2X: 11.32 fps 4X: 04.49 fps
Gaia 1X: 06.18 fps 2X: 04.26 fps 4X: 02.89 fps
Nyx 1X: 07.30 fps 2X: 06.24 fps
Nyx Fast 1X: 13.55 fps
Rhea 4X: 02.36 fps
4X Slowmo Apollo: 24.29 fps APFast: 55.55 fps Chronos: 13.27 fps CHFast: 20.49 fps
16X Slowmo Aion: ERR fps
No it doesn’t at all multi GPU only performs well in Single video. With multiple files, spreading the stuff to all files tiles/job/write turns into a high IDLE GPU, so this bench is at least unaccurate with single video, drama in multi
Sure, processing multiple videos in parallel will make better use of the CPU if there is free GPU processing power as u mentioned.
I primarily wanted to show that u can gain about 10% performance for jobs involving image scaling on 16 core AMD CPUs by limiting ffmpeg exe to the first 8 cores (via Windows task manager or Process Lasso).
It could be that TVAI on Intel systems benefit from running exclusively on the P-Cores (Last generations of Intel CPU have a mix of Performance= and Economy Cores)
Topaz Video AI v5.3.2
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: 10.96 fps 2X: 07.47 fps 4X: 02.31 fps
Iris 1X: 10.93 fps 2X: 06.43 fps 4X: 01.80 fps
Proteus 1X: 10.35 fps 2X: 07.23 fps 4X: 02.45 fps
Gaia 1X: 03.28 fps 2X: 02.20 fps 4X: 01.54 fps
Nyx 1X: 04.13 fps 2X: 03.55 fps
Nyx Fast 1X: 07.59 fps
Rhea 4X: 01.35 fps
4X Slowmo Apollo: 14.65 fps APFast: 40.37 fps Chronos: 07.94 fps CHFast: 12.54 fps
16X Slowmo Aion: 05.74 fps
1 Like
Topaz Video AI v5.3.2
System Information
OS: Windows v10.22
CPU: Intel(R) Core(TM) i7-8700K CPU @ 3.70GHz 63.936 GB
GPU: NVIDIA GeForce RTX 3060 11.826 GB
Processing Settings
device: 0 vram: 1 instances: 1
Input Resolution: 1920x1080
Benchmark Results
Artemis 1X: 09.14 fps 2X: 06.61 fps 4X: 02.23 fps
Iris 1X: 09.39 fps 2X: 05.90 fps 4X: 01.72 fps
Proteus 1X: 09.25 fps 2X: 06.55 fps 4X: 02.24 fps
Gaia 1X: 03.10 fps 2X: 02.15 fps 4X: 01.42 fps
Nyx 1X: 03.84 fps 2X: 03.27 fps
Nyx Fast 1X: 07.54 fps
Rhea 4X: 01.15 fps
4X Slowmo Apollo: 13.40 fps APFast: 33.02 fps Chronos: 07.48 fps CHFast: 12.26 fps
16X Slowmo Aion: 22.84 fps
1 Like
70 Watt limited 7950X + 4090 425W max, RAM 6200CL30
Topaz Video AI Alpha v5.3.0.1.a.hyp.rxl
System Information
OS: Windows v10.22
CPU: AMD Ryzen 9 7950X 16-Core Processor 31.118 GB
GPU: NVIDIA GeForce RTX 4090 23.576 GB
GPU: AMD Radeon(TM) Graphics 0.47445 GB
Processing Settings
device: 0 vram: 1 instances: 0
Input Resolution: 1920x1080
Benchmark Results
Artemis 1X: 44.05 fps 2X: 18.96 fps 4X: 05.19 fps
Iris 1X: 42.82 fps 2X: 22.31 fps 4X: 05.99 fps
Proteus 1X: 41.98 fps 2X: 22.13 fps 4X: 06.17 fps
Gaia 1X: 15.66 fps 2X: 10.96 fps 4X: 05.60 fps
Nyx 1X: 17.64 fps 2X: 15.97 fps
Nyx Fast 1X: 31.61 fps
Rhea 4X: 05.50 fps
4X Slowmo Apollo: 44.50 fps APFast: 91.33 fps Chronos: 33.24 fps CHFast: 39.70 fps
16X Slowmo Aion: 37.31 fps
7950X@70W max + 4090@300W max + RAM 6200CL30
Topaz Video AI Alpha v5.3.0.1.a.hyp.rxl
System Information
OS: Windows v10.22
CPU: AMD Ryzen 9 7950X 16-Core Processor 31.118 GB
GPU: NVIDIA GeForce RTX 4090 23.576 GB
GPU: AMD Radeon(TM) Graphics 0.47445 GB
Processing Settings
device: 0 vram: 1 instances: 0
Input Resolution: 1920x1080
Benchmark Results
Artemis 1X: 39.44 fps 2X: 18.12 fps 4X: 05.15 fps
Iris 1X: 37.83 fps 2X: 22.54 fps 4X: 06.06 fps
Proteus 1X: 38.67 fps 2X: 22.19 fps 4X: 06.21 fps
Gaia 1X: 14.48 fps 2X: 09.84 fps 4X: 05.58 fps
Nyx 1X: 16.11 fps 2X: 14.18 fps
Nyx Fast 1X: 29.27 fps
Rhea 4X: 05.43 fps
4X Slowmo Apollo: 43.65 fps APFast: 91.17 fps Chronos: 29.59 fps CHFast: 38.77 fps
16X Slowmo Aion: 35.92 fps
Windows 11 Pro 22h2 debloated
No Antivirus or cleaner programs
ASUS-PRIME Z790-A WIFI (Bios 1663)
i9-13900K + Kraken AIO
4090 drivers + 560.81 studio drivers
128 Gig DDR5-4800 UDIMM 1.1V CL40
Motherboard bios all auto no OC
Running Quick CPU
Core parking 100%
Frequency scaling 100%
Turbo boost 100%
Corsair HX1200 PS
4 onboard Nvme 4th gen drives
- Boot drive 2 TB
- Cache drive 2 TB
- Media drive 4 TB
- Mixdown drive 2 TB
Topaz Video AI v5.3.2
System Information
OS: Windows v11.22
CPU: 13th Gen Intel(R) Core(TM) i9-13900K 127.75 GB
GPU: NVIDIA GeForce RTX 4090 23.576 GB
Processing Settings
device: 0 vram: 1 instances: 1
Input Resolution: 720x480
Benchmark Results
Artemis 1X: 192.30 fps 2X: 80.69 fps 4X: 22.86 fps
Iris 1X: 175.45 fps 2X: 109.36 fps 4X: 30.77 fps
Proteus 1X: 230.52 fps 2X: 111.57 fps 4X: 30.55 fps
Gaia 1X: 87.63 fps 2X: 60.03 fps 4X: 27.32 fps
Nyx 1X: 52.41 fps 2X: 56.16 fps
Nyx Fast 1X: 94.03 fps
Rhea 4X: 25.22 fps
4X Slowmo Apollo: 223.39 fps APFast: 355.97 fps Chronos: 168.30 fps CHFast: 204.10 fps
16X Slowmo Aion: 139.89 fps
Topaz Video AI v5.3.2
System Information
OS: Windows v11.22
CPU: 13th Gen Intel(R) Core(TM) i9-13900K 127.75 GB
GPU: NVIDIA GeForce RTX 4090 23.576 GB
Processing Settings
device: 0 vram: 1 instances: 1
Input Resolution: 1920x1080
Benchmark Results
Artemis 1X: 41.18 fps 2X: 18.24 fps 4X: 04.61 fps
Iris 1X: 38.79 fps 2X: 18.74 fps 4X: 04.89 fps
Proteus 1X: 41.48 fps 2X: 20.10 fps 4X: 05.06 fps
Gaia 1X: 15.86 fps 2X: 11.17 fps 4X: 04.41 fps
Nyx 1X: 17.71 fps 2X: 14.52 fps
Nyx Fast 1X: 31.16 fps
Rhea 4X: 04.27 fps
4X Slowmo Apollo: 41.57 fps APFast: 72.48 fps Chronos: 33.47 fps CHFast: 34.32 fps
16X Slowmo Aion: 31.34 fps
Topaz Video AI v5.3.2
System Information
OS: Windows v11.22
CPU: 13th Gen Intel(R) Core(TM) i9-13900K 127.75 GB
GPU: NVIDIA GeForce RTX 4090 23.576 GB
Processing Settings
device: 0 vram: 1 instances: 1
Input Resolution: 3840x2160
Benchmark Results
Artemis 1X: 08.70 fps 2X: 04.10 fps 4X: 01.14 fps
Iris 1X: 08.37 fps 2X: 04.59 fps 4X: 01.15 fps
Proteus 1X: 08.93 fps 2X: 04.79 fps 4X: 01.25 fps
Gaia 1X: 03.38 fps 2X: 02.28 fps 4X: 01.03 fps
Nyx 1X: 02.93 fps 2X: 03.58 fps
Nyx Fast 1X: 05.20 fps
Rhea 4X: 00.96 fps
4X Slowmo Apollo: 14.28 fps APFast: 21.53 fps Chronos: 07.20 fps CHFast: 12.58 fps
16X Slowmo Aion: 15.29 fps
What you mean by “cache” drive ?
Interesting … CCD1 has always been the thorn in the foot of all Ryzens
Still get it with Granit Ridge 
