- Mar 18, 2019
- 237
- 3,863
Interesting insights. Thanks for taking the time to test this.I have a 3070 + 1080ti, so I did a comparison:
You don't have permission to view the spoiler content. Log in or register now.
Render settings - bog standard, default Daz Iray settings (95% convergence, Render Quality = 1 etc etc).
The only changes were to render resolution size (1403 x 2000), Tone Mapping (ISO 80, gamma = 2.0) and going to Advanced Options and setting Medium & High Texture Compression thresholds to 8000 (overkill I guess, for no reason)
So the times, then:
1080ti solo:
CUDA device 1 (GeForce GTX 1080 Ti): 956 iterations, 4.269s init, 247.207s render
3070 solo:
CUDA device 0 (GeForce RTX 3070): 944 iterations, 4.220s init, 85.336s render
3070 + 1080ti:
CUDA device 0 (GeForce RTX 3070): 722 iterations, 4.664s init, 65.830s render
CUDA device 1 (GeForce GTX 1080 Ti): 239 iterations, 5.105s init, 65.861s render
Just be aware that both cards are in my rig at the same time, and Daz DOES share memory between them even when one card is not selected to render.
Total scene GPU memory load was 11,290MB (6696MB 3070 + 4594MB 1080ti)
ED: I accidentally the spoiler title
no problem, was interested in the results myselfInteresting insights. Thanks for taking the time to test this.
Honestly, I'm not sure I can answer that question.kazuma_the_guy what will pricing be on that 3080? I'm not looking for just a GPU upgrade. I have my eye on a fully built pc with that 3070 onboard. I can't exchange my current gtx660 with a current one since my mainboard and cpu aren't fully compatible with the newer hardware, it will work but I'll never be able to get full performance out of it.
It was a $ 480.00 USD upgrade for me on the R-11 Aroura Alienware. I cant find any available at any of the usual outlets and a loooooooooong wait time from Nvidia. A RTX 3090 is just out of the question.. The extra $480 was worth it to me to just get a pre-built machine. It was just over $1250.00 cheaper than building one out like I usually do, so a big savings and no real waite ( two weeks ) vs still waiting for parts to come in stock + build and test out. The 11 gen I-9 was and still is out of stock every where.Honestly, I'm not sure I can answer that question.
It depends on what country you live in, really. I did a quick search in my own country and now not even 3070s are available anymore, and 3060s are selling at 3080 prices
'Officially' from nVidia, a 3080 is around 700USD and a 3070 is 500USD (lmao a 3060 now sells for 698USD here)
But TBH if it's a full upgrade rig with a 3070 (from a 660), it's definitely not bad if the price is ok with you
I am having the same issue... Looks GreatI'm experimenting and trying to upgrade/port from G3F to G8.1F along with some changes. Not easy tbh... What do you guys think?
View attachment 1104797
It's hard to believe that this model needs 11GB with the hair and clothes alone. Do you also have 11GB with the solos? Isn't it more plausible that it only needs 4.5GB and that's on each card at the same time for them to work? 2GB extra on the one card should be from the system. Don't get me wrong, if shared memory is a thing, it would be a game changer.I have a 3070 + 1080ti, so I did a comparison:
You don't have permission to view the spoiler content. Log in or register now.
Render settings - bog standard, default Daz Iray settings (95% convergence, Render Quality = 1 etc etc).
The only changes were to render resolution size (1403 x 2000), Tone Mapping (ISO 80, gamma = 2.0) and going to Advanced Options and setting Medium & High Texture Compression thresholds to 8000 (overkill I guess, for no reason)
So the times, then:
1080ti solo:
CUDA device 1 (GeForce GTX 1080 Ti): 956 iterations, 4.269s init, 247.207s render
3070 solo:
CUDA device 0 (GeForce RTX 3070): 944 iterations, 4.220s init, 85.336s render
3070 + 1080ti:
CUDA device 0 (GeForce RTX 3070): 722 iterations, 4.664s init, 65.830s render
CUDA device 1 (GeForce GTX 1080 Ti): 239 iterations, 5.105s init, 65.861s render
Just be aware that both cards are in my rig at the same time, and Daz DOES share memory between them even when one card is not selected to render.
Total scene GPU memory load was 11,290MB (6696MB 3070 + 4594MB 1080ti)
ED: I accidentally the spoiler title
Yeah, man. It's hard to justify as well. But seeing so many things getting released for G8F is very tempting. Also, converter scripts for clothes and hair just don't work well - quite a lot. There's is lot of trial and error experimentations with this process, very time and nerve consuming.I am having the same issue... Looks Great
Can you show reference of this model in the other Genesis please? I'm also looking to port a character from G3 to G8 but I do want the character to stay true to the original form.I'm experimenting and trying to upgrade/port from G3F to G8.1F along with some changes. Not easy tbh... What do you guys think?
View attachment 1104797
Check my profile for my posts, you can see plenty of pics with the G3F model there. However, what I'm doing is not really 'porting'... I basically started the body with a G8.1F from scratch, using the texture jpg's and settings from the previous model and Checkertoo's Lara Croft For G8F.1 face morph, morphing/modifying EVERYTHING to my taste.Can you show reference of this model in the other Genesis please? I'm also looking to port a character from G3 to G8 but I do want the character to stay true to the original form.