TVAI 3.1.x slower than 3.0.x

[suraj]

Anyone having TVAI 3.1.x running slower than 3.0.x or 2.6.4 please post the following information below.
CPU, RAM, GPU, VRAM, input resolution, model used, output resolution, fps or slow motion amount, spf/total time taken in both versions.
If the amount is significant please try reducing the Max Memory setting to around 75% and try again.

[adrjet22]

Tried with a low end PC.

CPU: I7 7700
RAM: 16gb
GPU: NVIDIA GTX 1060
VRAM: 6gb
OS: Windows 10

TVAI version: 3.1.0
Model: Proteus v3
Input Resolution: 1080p
Output Resolution: 1080p (only enhancing with relative to auto)
100%: 0.9fps
75%: 0.5fps

[ssbroly]

Hi, I made a comparison with the same video 720 x 576 50 fps, with artemis medium, a comparison in my original size, and a comparison in x2 upscale, 2.6.4 is much faster, I don’t understand nothing … I tried to lower the vram to 75%, even lower, nothing changes. I have a gtx 1660, a ryzen 7 2700 and 16gb of ram.

[david.123]

I will do some tests tonight with 3.1.1 vs 3.0.12 and post results.

[david.123]

@suraj

Good news, and some bad news!

Testing complete, in summary, v3.1.1 is faster with Artemis and Proteus Auto (didn’t test other enhancement models) than v3.0.12. However, Chronos Fast performance in v3.1.1 is about 25 to 30% slower, or in the case of a 4k input source, will still crash Windows with a BSOD. Also, I did not get the 4x slowdowns on Chronos Fast that were present on v3.1.0, so performance has improved in v3.1.1 but still lags v3.0.12.

OS: Windows 11, Nvidia Studio Driver 528.02
GPU: GTX 1070 8GB VRAM
CPU: Intel 6700K
System RAM: 16GB

Max memory set to 100% in Video AI preferences

3.0.12:

Input source: 3840x2160 30 FPS:

2x Chronos Fast to 60 FPS conversion:
1.30 spf
Artemis High, no FPS conversion:
2.40 spf
Proteus Auto, no FPS conversion:
2.88 spf

Input source: 1920x1080 29.97 FPS:

2x Chronos Fast to 59.94 FPS conversion:
0.33 spf
Artemis High, no FPS conversion:
0.57 spf
Proteus Auto, no FPS conversion:
0.60 spf

3.1.1:

Input source: 3840x2160 30 FPS:

2x Chronos Fast to 60 FPS conversion:
Max memory set to 100% in Video AI preferences, Windows BSOD/hard crash!
Max memory set to 80%:
1.65 spf

Max memory set back to 100% in Video AI preferences for following testing as only Chronos Fast 2x on a 4k source seems to crash Windows when max memory usage is set to 100%.

Artemis High, no FPS conversion:
2.14 spf
Proteus Auto, no FPS conversion:
2.47 spf

Input source: 1920x1080 29.97 FPS:

2x Chronos Fast to 59.94 FPS conversion
0.50 spf
Artemis High, no FPS conversion:
0.45 spf
Proteus Auto, no FPS conversion:
0.49 spf

[meimeiriver]

@suraj

If the amount is significant please try reducing the Max Memory setting to around 75% and try again.

Fortunately, v3.1 is vastly superior here, speed-wise, to the v3.0 series. But why would assigning more memory decrease speed in the first place? Paging?

[TPX (Thomas D.)]

Change the size of memory does nothing.

Model:	V2.6.4	RAM	VRAM	CPU-Load	GPU-Load
	Sec/frame	MB	MB
Proteus	0.12	15300	2629	10-20%	2-94%
Artemis LQ	0.09	15300	2690	8-30%	5-90%
Chronos v3 f	0.03	14500	1718	19-38%	6-8%
Stabilisation
Gaia HQ	0.12	15500	3847	5-20%	20-90%

---

Model:	V3.0.12	RAM	VRAM	CPU-Load	GPU-Load
	Sec/frame	MB	MB
Proteus	0.15	16300	3911	9-15%	5-45%
Artemis LQ	0.12	15900	3802	13-20%	8-70%
Chronos v3 f	0.12	16300	4054	10-18%	6-60%
Stabilisation	0.41	16500	5421	5-15%	9-60%
Gaia HQ	0.07	16000	4817	12-20%	6-100%

---

Model:	V3.1.1	RAM	VRAM	CPU-Load	GPU-Load
	Sec/frame	MB	MB
Proteus	0.12	15400	4726	6-13%	6-80%
Artemis LQ	0.08	15300	4581	3-10%	30-70%
Chronos v3 f	0.07	15600	5643	8-15%	30-70%
Stabilisation	0.11	16300	6727	5-15%	30-80%
Gaia HQ	0.09	15300	5956	5-10%	50-98%

---

---

Closed no active APP	Closed no active APP
RAM	VRAM
13424	1215

Input	File:
Res:	720x576
Format:	.mov

---

output:	File:
Res:	1080p
Format:	.mp4
188% scale	H.264 (AMD)

---

Sys specs:
CPU	AMD	Threadripper	3960X	128 GB	ECC on	SMT off 24 cores only
GPU	AMD	Radeon PRO	W6800	32GB	ECC on
OS	Microsoft Windows 11 Pro for Workstations
Version	10.0.22621 Build 22621
MainBoard	PRIME TRX40-PRO
Radeon PRO	GPU Driver	V22.11.2	Date 30.11.22

---

---

GPU Memory Visualisation while Proteus is processing.

---

Visualisation of Proteus while export

[meimeiriver]

TVAI uses about 1.3G (an upscale to 4K), ffmpeg ca. 5G.

[suraj]

Found the issue here and fixed it. Performance should be 5-25% better than 2.6.4 with todays beta. Looking into others.

[suraj]

In some cases especially for cards with lower vram, swapping from ram to vram happens a lot more leading to slow downs.

[meimeiriver]

@suraj

In some cases especially for cards with lower vram, swapping from ram to vram happens a lot more leading to slow downs

Thanks for getting back to me on this one. It’s appreciated.

[TPX (Thomas D.)]

That means boards + GPUs with both pci-e 4 will be faster.

[ssbroly]

that’s great! thank you very much!

[TPX (Thomas D.)]

Updated Numbers with a Intel Nvidia system, just for comparison.
I’ve to retest the AMD sys with the beta, i did turn off Freq-enhance, so tested with lower freq.

---

AMD Threadripper 3960X|128 GB ECC on SMT off 24 cores only
AMD Radeon PRO W6800 32GB ECC on
Microsoft Windows 11 Pro for Workstations
Version 10.0.22621 Build 22621
Main Board PRIME TRX40-PRO
Radeon PRO GPU Driver V22.11.2 Date 30.11.22

Model:	V3.1.2.0.b	RAM	VRAM	CPU-Load	GPU-Load	10 sec preview	AI-Processor
	Time:
							Mem = 50%
	Sec/frame	MB	MB
Proteus	0.09	15700	2657	20-35%	2-94%	56 sec	Auto
Artemis LQ	0.08	15700	2625	8-30%	5-90%	49 sec	Auto
Chronos v3 f	0.08	15800	4945	19-38%	2-94%	47 sec	Auto
Stabilisation	0.22	16123	3861	12-30%	60-80%	1m45sec	Auto
Gaia HQ	0.08	15700	6204	5-20%	60-90%	43 sec	Auto

---

---

Just for Comparison
Nvidia Quadro RTX 5000 / 16GB VRAM (like 2080S), Driver: 528.24
Intel 7820X, 8C/16T, 64 GB Ram
Win 11, PRO

---

Model:	V3.1.2.0.b	RAM	VRAM	CPU-Load	GPU-Load	10 sec preview	AI-Processor
	Time:
							Mem = 50%
	Sec/frame	MB	MB
Proteus	0.05	11715	2584	12-20%	40-50%	30 sec	Auto
Artemis LQ	0.04	11234	2461	12-20%	60-66%	22 sec	Auto
Chronos v3 f	0.04	11584	3075	12-24%	55-62%	37 sec	Auto
Stabilisation	0.12	12200	4816	16-30%	30-45%	1m40sec	Auto
Gaia HQ	0.08	10899	3763	9-16%	95-97%	45 sec	Auto

---

---

---

Model:	V3.1.1 AMD sys	RAM	VRAM	CPU-Load	GPU-Load	10 sec preview	AI-Processor
							Mem = 50%
	Sec/frame	MB	MB
Proteus	0.11	16100	5183	6-13%	6-80%	58 sec	Auto
Artemis LQ	0.08	15900	5071	13-30%	30-70%	46 sec	Auto
Chronos v3 f	0.07	16200	6100	20-30%	30-70%	58 sec	Auto
Stabilisation	0.11	17200	7200	20-30%	30-80%	1m45sec	Auto
Gaia HQ	0.08	15400	6420	10-20%	50-98%	48 sec	Auto

---

---

output:	File:
Res:	1080p
Format:	.mp4
188% scale	H.264 (AMD/Nvidia)

---

---

[suraj]

Here is a version for anyone with the issue to test

Windows Alpha for AMD & Nvidia GTX GPUs only

Hi All, Here is an early version of app that should improve performance for AMD GPUs and hopefully older Nvidia GPUs(non RTX cards) that use onnx models. Download: Win Known Issues: Proteus Auto will not work, use Proteus with manual settings Models cannot be run on CPU Models cannot be run on intel GPU Models run slower on RTX GPUs Thank you for testing PS: Since this is an alpha version, it will not interfere with Beta or Release versions on the machine.