# Topaz Video AI Performance Alpha 3.2.0.0.a **URL:** \ **Category:** Video AI Beta\ **Created:** [March 10, 2023, 2:33am UTC](https://community.topazlabs.com/t/topaz-video-ai-performance-alpha-3-2-0-0-a/41373 "2023-03-10T02:33:49Z")\ **Posts on this page:** 20\ **Page:** 1
**Author:** ![suraj](https://community.topazlabs.com/letter_avatar_proxy/v4/letter/s/ecccb3/32.png) [@suraj](https://community.topazlabs.com/u/suraj)\ **Post date:** [March 10, 2023, 2:33am UTC](https://community.topazlabs.com/t/topaz-video-ai-performance-alpha-3-2-0-0-a/41373/1 "2023-03-10T02:33:49Z")
Hi All, Here is an early version of app that should improve performance for all machines. The degree of performance increase will vary. This version should also allow for better support running multiple filters and multiple processes. Download: [Win](https://downloads.topazlabs.com/deploy/TopazVideoAIAlpha/3.2.0.0.a/TopazVideoAIAlpha-3.2.0.0.a.msi) | [Mac](https://downloads.topazlabs.com/deploy/TopazVideoAIAlpha/3.2.0.0.a/TopazVideoAIALPHA-3.2.0.0.a.dmg) | [Mac Update](https://downloads.topazlabs.com/deploy/TopazVideoAIAlpha/3.2.0.0.a/TopazVideoAIALPHA-3.2.0.0.a.pkg) Please upload problem videos and logs here: [Submit Files](https://www.dropbox.com/request/hxLNzpT1g3nsCHG4iM2E) **Known Issues** : - Proteus Auto will not work, use Proteus with manual settings - Running stabilization may cause failures - No way to boost performance if running single filter or process Thank you for testing PS: Since this is an alpha version, it will not interfere with Beta or Release versions on the machine. ---
**Author:** ![suraj](https://community.topazlabs.com/letter_avatar_proxy/v4/letter/s/ecccb3/32.png) [@suraj](https://community.topazlabs.com/u/suraj)\ **Post date:** [March 10, 2023, 2:41am UTC](https://community.topazlabs.com/t/topaz-video-ai-performance-alpha-3-2-0-0-a/41373/2 "2023-03-10T02:41:03Z")
---
**Author:** ![david.123](https://community.topazlabs.com/user_avatar/community.topazlabs.com/david.123/32/163156_2.png) [@david.123](https://community.topazlabs.com/u/david.123)\ **Post date:** [March 10, 2023, 3:34am UTC](https://community.topazlabs.com/t/topaz-video-ai-performance-alpha-3-2-0-0-a/41373/3 "2023-03-10T03:34:45Z")
Thanks for the new Alpha to test. **Testing results so far, (GTX1070, max memory set to 100%):** 1080p input, no resize, 2x Chronos Fast + Proteus Manual (no sliders moved from defaults), 0.78spf 1080p input, no resize, 2x Chronos Fast + Artemis High, 0.77spf **Gaia + Chronos testing, strange results:** 1080p input, no resize, 2x Chronos Fast + Gaia HQ, started out at a crawl - **8.11spf** - then when it hit 4% completion, it started speeding up, and by the time it hit 25% completion, it was running at **1.47spf** which is about the speed I expected based on the Artemis and Proteus numbers! Any ideas why? **Update 1:** 2nd test run Gaia, same settings as above except added grain, **~8spf** again, but not speeding up this time, stuck at low speed! Seems to be an issue here related to using Gaia. **Update 2:** Closed application, did another Gaia test run, same settings as update 1, this time it had the slow then fast behaviour as in the first Gaia test. Interestingly, during the times where speed was stuck at ~8spf, the task manager showed GPU usage sometimes dropping to zero, then spiking to 100, then dropping again: ![image](https://community-cdn.topazlabs.com/original/3X/b/6/b6fe24d0cfde43fe369dd2ef82c5fc5fd9857f17.png) **Update 3:** Good news with Chronos Fast on a 4k source, 2x from 30 to 60 fps conversion, has not crashed when memory is set to max on GTX 1070. Previously would crash every time unless max memory was set lower. Getting VERY good speed too at 1.12spf. ---
**Author:** ![alanhusband-677420](https://community.topazlabs.com/letter_avatar_proxy/v4/letter/a/7ba0ec/32.png) [@alanhusband-677420](https://community.topazlabs.com/u/alanhusband-677420)\ **Post date:** [March 10, 2023, 3:37am UTC](https://community.topazlabs.com/t/topaz-video-ai-performance-alpha-3-2-0-0-a/41373/4 "2023-03-10T03:37:04Z")
On my GTX 1060 it’s perhaps slightly faster than the previous alpha, but this is close to margin of error level of improvement. 720p to 1080p Manual Proteus upscale, 2.4 - 2.5fps ---
**Author:** ![joachim.ehrhardt](https://community.topazlabs.com/user_avatar/community.topazlabs.com/joachim.ehrhardt/32/76638_2.png) [@joachim.ehrhardt](https://community.topazlabs.com/u/joachim.ehrhardt)\ **Post date:** [March 10, 2023, 8:53am UTC](https://community.topazlabs.com/t/topaz-video-ai-performance-alpha-3-2-0-0-a/41373/6 "2023-03-10T08:53:48Z")
On my RTX 3060 using Artemis HQ for 1920 to 4k I get about 10% speed improvement (5.7 fps vs. 5.2 fps with 3.1.8). GAIA, THEIA settings are operational, but no PROTEUS. ---
**Author:** ![alanhusband-677420](https://community.topazlabs.com/letter_avatar_proxy/v4/letter/a/7ba0ec/32.png) [@alanhusband-677420](https://community.topazlabs.com/u/alanhusband-677420)\ **Post date:** [March 10, 2023, 9:21am UTC](https://community.topazlabs.com/t/topaz-video-ai-performance-alpha-3-2-0-0-a/41373/7 "2023-03-10T09:21:31Z")
For more detail I’d consider using GPU-Z for a broader look as to what is happening on the gpu. Certainly monitor vram usage across a test run. ---
**Author:** ![TPX](https://community.topazlabs.com/user_avatar/community.topazlabs.com/tpx/32/194727_2.png) [@TPX](https://community.topazlabs.com/u/TPX)\ **Post date:** [March 10, 2023, 10:34am UTC](https://community.topazlabs.com/t/topaz-video-ai-performance-alpha-3-2-0-0-a/41373/8 "2023-03-10T10:34:43Z")
Sys: TR 3960X - W6800 Artemis Low, 10 sec preview: 4K (2160p), 25 fps. 3.2.0.0 = 2 min 10 sec. 3.1.8 = 3 Min 6 sec. * * * Artemis Low, 10 sec preview: 4K (2160p), 25 fps. **+Themis** 3.2.0.0 = 6 min 52 sec. ← strong flicker in some areas. 3.1.8 = 7 Min 6 sec. ← light flicker in some areas. * * * Artemis Low, 10 sec preview: 4K (2160p), 25 fps. **+Themis** + **60fps Chronos Fast.** 3.2.0.0 = 13 min 11 sec. ← light flicker in some areas. 3.1.8 = 13 Min 45 sec. ← light flicker in some areas. **Uploaded Nightshot mp4 ← pre-denoised with DaVinci Resolve.** (ask Suraj) * * * Proteus, 10 sec preview: 4K (2160p), 25 fps. 3.2.0.0 = 2 min 18 sec. 3.1.8 = 3 Min 43 sec. ---
**Author:** ![david.123](https://community.topazlabs.com/user_avatar/community.topazlabs.com/david.123/32/163156_2.png) [@david.123](https://community.topazlabs.com/u/david.123)\ **Post date:** [March 10, 2023, 3:49pm UTC](https://community.topazlabs.com/t/topaz-video-ai-performance-alpha-3-2-0-0-a/41373/9 "2023-03-10T15:49:34Z")
I think the W11 task manager shows accurate GPU usage as opposed to W10 (I could be wrong though). ---
**Author:** ![david.123](https://community.topazlabs.com/user_avatar/community.topazlabs.com/david.123/32/163156_2.png) [@david.123](https://community.topazlabs.com/u/david.123)\ **Post date:** [March 10, 2023, 3:54pm UTC](https://community.topazlabs.com/t/topaz-video-ai-performance-alpha-3-2-0-0-a/41373/10 "2023-03-10T15:54:58Z")
> [@joachim.ehrhardt](#): > > GAIA, THEIA settings are operational, but no PROTEUS. Did you catch the part of the release notes that Proteus must be used in ‘manual’ mode? ---
**Author:** ![JesusSheep](https://community.topazlabs.com/letter_avatar_proxy/v4/letter/j/ec9cab/32.png) [@JesusSheep](https://community.topazlabs.com/u/JesusSheep)\ **Post date:** [March 10, 2023, 4:01pm UTC](https://community.topazlabs.com/t/topaz-video-ai-performance-alpha-3-2-0-0-a/41373/11 "2023-03-10T16:01:26Z")
@suraj Uploaded Test Videos .🙂 ---
**Author:** ![TPX](https://community.topazlabs.com/user_avatar/community.topazlabs.com/tpx/32/194727_2.png) [@TPX](https://community.topazlabs.com/u/TPX)\ **Post date:** [March 10, 2023, 4:04pm UTC](https://community.topazlabs.com/t/topaz-video-ai-performance-alpha-3-2-0-0-a/41373/12 "2023-03-10T16:04:12Z")
Does the Taskmanager readout every 100 milisecond? With HW info i did set the readout to 25 milisecond and then you can see when TVAI is writing and reading to the drive where your video is located. A HDD can slow down thw software by a little bit but not by much when you process 4K files. PCI-E 3 vs 4 is also no problem, PCI-E 3 is fast enough. TVAI does react like a game with DirectX. The higher the Resolution of a Video and the more filter you have the more plays the GPU a role, if its the oposite the CPU plays a bigger role. ---
**Author:** ![david.123](https://community.topazlabs.com/user_avatar/community.topazlabs.com/david.123/32/163156_2.png) [@david.123](https://community.topazlabs.com/u/david.123)\ **Post date:** [March 10, 2023, 4:10pm UTC](https://community.topazlabs.com/t/topaz-video-ai-performance-alpha-3-2-0-0-a/41373/13 "2023-03-10T16:10:17Z")
> [@TPX](#): > > Does the Taskmanager readout every 100 milisecond? > > With HW info i did set the readout to 25 milisecond and then you can see when TVAI is writing and reading to the drive where your video is located. Good info thanks, I’m not sure of the refresh rate of task manager. > [@TPX](#): > > A HDD can slow down thw software by a little bit but not by much when you process 4K files. All SSD on my system so should not be a factor. Performance across all models was good with this Alpha, the only outstanding issue is the strange issue with Gaia starting at (and sometimes remaining at) extremely slow speed. On the occasions when it gets ‘unstuck’ it speeds up by quite a lot. I will continue to try and reproduce the issue and provide logs if devs ask. ---
**Author:** ![JesusSheep](https://community.topazlabs.com/letter_avatar_proxy/v4/letter/j/ec9cab/32.png) [@JesusSheep](https://community.topazlabs.com/u/JesusSheep)\ **Post date:** [March 10, 2023, 4:14pm UTC](https://community.topazlabs.com/t/topaz-video-ai-performance-alpha-3-2-0-0-a/41373/14 "2023-03-10T16:14:49Z")
@suraj Uploaded Video is Screen Shot. With Settings 2 hrs+ ![Screenshot 2023-03-10 at 9.11.07 AM](https://community-cdn.topazlabs.com/original/3X/6/1/6183acd14f9e5498e1f5a51ce224effa66b80a1f.jpeg) ![Screenshot 2023-03-10 at 9.12.28 AM](https://community-cdn.topazlabs.com/original/3X/d/4/d40207fe8c42947e83419012060b0a79a46a73d6.jpeg) ---
**Author:** ![david.123](https://community.topazlabs.com/user_avatar/community.topazlabs.com/david.123/32/163156_2.png) [@david.123](https://community.topazlabs.com/u/david.123)\ **Post date:** [March 10, 2023, 4:31pm UTC](https://community.topazlabs.com/t/topaz-video-ai-performance-alpha-3-2-0-0-a/41373/15 "2023-03-10T16:31:42Z")
Instead of showing your entire desktop in all your screen shots, please try this to capture just the Video AI window: ![image](https://community-cdn.topazlabs.com/original/3X/2/1/21587ac6edd8dd3f6b14cff7e0416f400e9bed6c.jpeg) ---
**Author:** ![david.123](https://community.topazlabs.com/user_avatar/community.topazlabs.com/david.123/32/163156_2.png) [@david.123](https://community.topazlabs.com/u/david.123)\ **Post date:** [March 10, 2023, 4:39pm UTC](https://community.topazlabs.com/t/topaz-video-ai-performance-alpha-3-2-0-0-a/41373/16 "2023-03-10T16:39:50Z")
> [@TPX](#): > > Does the Taskmanager readout every 100 milisecond? Looks like default refresh rate is 1 second but you can change it to 500ms under settings. ---
**Author:** ![JesusSheep](https://community.topazlabs.com/letter_avatar_proxy/v4/letter/j/ec9cab/32.png) [@JesusSheep](https://community.topazlabs.com/u/JesusSheep)\ **Post date:** [March 10, 2023, 4:47pm UTC](https://community.topazlabs.com/t/topaz-video-ai-performance-alpha-3-2-0-0-a/41373/17 "2023-03-10T16:47:22Z")
ok, will do. ---
**Author:** ![joachim.ehrhardt](https://community.topazlabs.com/user_avatar/community.topazlabs.com/joachim.ehrhardt/32/76638_2.png) [@joachim.ehrhardt](https://community.topazlabs.com/u/joachim.ehrhardt)\ **Post date:** [March 10, 2023, 4:51pm UTC](https://community.topazlabs.com/t/topaz-video-ai-performance-alpha-3-2-0-0-a/41373/18 "2023-03-10T16:51:40Z")
Yes, I’ve read that, and did use it that way successfully. But when I changed scaling for that video (after I stoped it!) same Proteus manual settings refused to function any longer (unkwown error), and any other video also could no longer use PROTEUS. When I restarted 3.2. I could use Proteus manual again, but not with all scaling factors. ---
**Author:** ![suraj](https://community.topazlabs.com/letter_avatar_proxy/v4/letter/s/ecccb3/32.png) [@suraj](https://community.topazlabs.com/u/suraj)\ **Post date:** [March 10, 2023, 4:53pm UTC](https://community.topazlabs.com/t/topaz-video-ai-performance-alpha-3-2-0-0-a/41373/19 "2023-03-10T16:53:31Z")
Can you send me the logs for situations when proteus errored out. ---
**Author:** ![joachim.ehrhardt](https://community.topazlabs.com/user_avatar/community.topazlabs.com/joachim.ehrhardt/32/76638_2.png) [@joachim.ehrhardt](https://community.topazlabs.com/u/joachim.ehrhardt)\ **Post date:** [March 10, 2023, 5:12pm UTC](https://community.topazlabs.com/t/topaz-video-ai-performance-alpha-3-2-0-0-a/41373/20 "2023-03-10T17:12:24Z")
[2023-03-10-09-23-14-Main.tzlog](https://community.topazlabs.com/uploads/short-url/fdrmGfzCKMJxAE2GsVpeeQwVUPc.tzlog) (1.4 MB) [2023-03-10-11-48-4-Main.tzlog](https://community.topazlabs.com/uploads/short-url/ilHIymbqB8FTiLWHqVRDPkqi7Sv.tzlog) (2.0 MB) [2023-03-10-12-04-55-Main.tzlog](https://community.topazlabs.com/uploads/short-url/kJiWB9kLhM0Ao14E2FXQL8unkvM.tzlog) (382.9 KB) I can use proteus manual, but certain combinations produce unknown error ---
**Author:** ![suraj](https://community.topazlabs.com/letter_avatar_proxy/v4/letter/s/ecccb3/32.png) [@suraj](https://community.topazlabs.com/u/suraj)\ **Post date:** [March 10, 2023, 10:57pm UTC](https://community.topazlabs.com/t/topaz-video-ai-performance-alpha-3-2-0-0-a/41373/22 "2023-03-10T22:57:03Z")
[Next page](https://community.topazlabs.com/t/topaz-video-ai-performance-alpha-3-2-0-0-a/41373.md?page=2)