thanks, what’s the difference with 1.6.0 ? i saw no annoncement anywhere about this 1.6.1 !
from what Matt Told, it has not been released yet… so no idea what it is…
there’s no RTX3000 laptops yet
Hi, thanks for the update! It’s great to see Theia model had been improved! Gonna try it now! thanks!!
I think it’s probably referring to the Quadro RTX 3000.
If it’s a benchmark for the GeForce RTX 3000 series, I’m interested in it.
I have a 3080 GPU but it happens to sit in 2nd PCI-E, therefore it’s a CUDA device #1, while 1st slot is occupied by 1080ti which is CUDA device #0.
Since i can’t actually select a GPU to use (saw that in known bugs) it’s either running on 1080ti or endlessly initializing on 3080.
I’m not entirely sure if it’s a GPU selection bug that i need to blame or if Ampere GPUs are not working yet…
UP: Never mind, using VEAI it actually managed to initialize on 3080 10 minutes later, but it’s not working properly, resulting images are full of artifacts.
veai.exe -i "input.avi" -f png -s 2.0 -m ahq-1.0.1
I’ll try other models when i have time.
already installed , launched and testing :). bu thanks 
then it’s the same GPU as the RTX2070 (vanilla), with waaaay lower clocks (very close to the RTX2070 max-Q) and less memory, no need for benchmarks
uncool and low priority! there are more important features to be added first.
Hi mistdragon, you’re one of the lucky guy who has RTX3080 in hand. Appreciate if you could post some benchmark on 352x240 / 352x288 (VCD), 720x480 / 720x756 (DVD) Upscale 200%. On RTX2060 VCDs are 0.10s / frame, DVDs are 0.30s / frame. I wonder if RTX3080 can process 4 times faster, because the CUDA counts 4 times over RTX2060. If not, how fast it really can go? Is it worth buying to speed up the process? Thanks.
anyone could tell me is it not working on RTX3080 ? i just change my display card and keep loading on perpareing in 5mins screen , thanks all guy 
I will test the speed if i’ll get it to work properly. For now, it only managed to run once and produced corrupted frames. I’ll check other models and if it will work on any of them, i’ll run speed benchmarks.
I’ve had the same problem which makes me think it’s not my mGPU setup that’s at fault. I suggest you try running the proccess using VEAI. It runs after ~10 minutes of initializing, but produces corrupted frames on my machine. I’m curious to see if it’s my problem or it you can repeat it on your 3080 too.
Thank you for the information.
GeForce RTX 3080 costs about $1000 in Japan at the shop price, so I can’t afford it for a poor person like me.
For your reference, I would appreciate it if you could tell me how many seconds per frame can be processed in the following settings.
SIZE: 1920×1080, SCALE: 100%, Gaia-CG
Please take your time.
0.41 per frame
Thank you for the information.
That’s great!
On my PC, it’s 13-15 sec./frame.
(Ryzen 5 3600)
I don’t play PC games, so I haven’t needed the luxury of a GPU in the past, but I would consider replacing it with the GeForce RTX 3000 series.
Anyone has benchmark table for various GPUs? Are CUDA cores count matters for speed? I do not use GPU for playing games, but rather for scaling videos. Please advice.
a benchmark table is available on the facebook beta group for VEAI
I just almost (got) a heart attack when I saw the upgrade yearly price for VEAI tonight for the first time…
I tought at first it was a joke. i’m very sad, i was not aware of such a price after the topic of the topaz “boss” about the change in the upgrade plan, telling that all software will have upgrade price between 49 and 99" per year… but didn’t know it would be 200 for VEAI…
it’s something i can’t afford… and will certainly have to leave… i hope some discount on the upgrade will maybe happen one day…
paying the upgrade more expensive than what i bought is just impossible…
1 Like
