Announcement
Collapse
No announcement yet.
Vray 3.5 gpu and Fstorm gpu 2k scene demo
Collapse
X
-
I don't see qny reason why fstorm is better here on the contrary.
Vray uses the gpu in a much more efficient way, and that's exactly why your system is stuck, because it uses the gpu at its maximum.
Fstorm doesn't and that's why you have more smooth windows interactions.
Take a cpu who gas 8 cores. Do one render with the 8 cores on, everything will lag. Start a render and set the affinity to use only 6 cores, well you will be able to surf and listen to spotify. Is that really better? Not in my vision, because your render with 6 cores will take more time.
If you want more interactiveness with vray, lower down the PPP and you will force vray to run slower and you will gain some interactiveness in windows. Again, the same way as fstorm, but for me really, that's not a feature, it's just fstorm that can't utilise the gpu at it's maximum speed, and I really don't understand how this can be a good thing for the end user...Stan
-
Well I think he has a point here. GPU rendering is still unattractive to many, and many like to use their computer while rendering. So entry barrier of needing another GPU to even be able to use your computer with V-Ray GPU comfortably adds to already strong skepticism and discouragement from using GPU solution. I don't think that people are happy to buy another graphics card and possibly needing to upgrade also PSU just to give GPU renderer a shot.
Furthermore, I think that V-Ray GPU is one of the very few GPU renderers that do not allow limiting of GPU usage to leave at least 5-10% of GPU performance for OS. I find it hard to believe there is some "special" way of utilizing GPU that is "much more efficient" but requires GPU to be utilized at 100%, as Sbrusse has said, especially given the fact that V-Ray GPU does not seem to be any more efficient when it comes to raw path tracing power than its competitors. Often it even feels the other way around.Last edited by LudvikKoutny; 23-01-2017, 05:58 AM.
Comment
-
If you use the V-Ray GPU in active shade, there is "Low GPU Thread priority option", OS UI will be much more smooth. Some people want GPUs fully utilized, others don't. We have presets for both, literally 2 clicks away. If we didn't people would complain for the opposite.
If you have actual scene where V-Ray GPU renders slow, please send them over and we will fix them. It is hard to debug feelings and testing it here makes me feel pretty confident about how solid the V-Ray GPU performance is, especially 3.5 .
Best,
Blago.V-Ray fan.
Looking busy around GPUs ...
RTX ON
Comment
-
Yes, but you can't really use activeshade for final frame rendering, and it's mostly final frame rendering where you want to use your PC to do something else in the meanwhile. During activeshade, you are doing look development, and you are engaged with 3ds Max, but when rendering final frame, you usually switch to some other program, unless you like to watch noise slowly going away. So absence of GPU utilization limit in production rendering mode is what's an issue here.
If I see any scene that performs significantly slower in V-Ray GPU compared to other GPU renders, then I will definitely send it your way.
Comment
-
Originally posted by Recon442 View PostYes, but you can't really use activeshade for final frame rendering
So absence of GPU utilization limit in production rendering mode is what's an issue here.
If I see any scene that performs significantly slower in V-Ray GPU compared to other GPU renders, then I will definitely send it your way.
Best,
Blago.V-Ray fan.
Looking busy around GPUs ...
RTX ON
Comment
-
I always thought ActiveShade can not properly lock down 3ds Max during rendering to make it uninterruptible, so there's a reason why people use production to render final frames and/or sequences. You may be right though, I never looked into that.
As for the setting GPU priority lower in production mode, I had absolutely no idea about that. I guess this comes down to my lack of knowledge about V-Ray GPU.
Maybe it would be good idea to enable GPU low thread priority by default? So people who do not know about it don't get scared away... ? I mean if you don't use your PC much during rendering, then GPU will be utilized almost entirely anyway, won't it?
Comment
-
Originally posted by Recon442 View PostAs for the setting GPU priority lower in production mode, I had absolutely no idea about that. I guess this comes down to my lack of knowledge about V-Ray GPU.
Anyway, it is written in this (really) short FAQ here , and is part of almost any GPU guide and tutorial we make. There are links to those in the sticky topic in the RT GPU forum as well.
Best,
Blago.Last edited by savage309; 23-01-2017, 06:45 AM.V-Ray fan.
Looking busy around GPUs ...
RTX ON
Comment
-
Hmm... I just looked at that guide, and this is what I see:
Not a single mention of GPU thread priority checkbox in paragraph about sluggish UI
But then again, most of the experienced people will figure out what it does right away. It's just that I, personally, have very little experience with V-Ray GPU. Everytime I gave it a try, something scared me away very quickly
Comment
-
Originally posted by Recon442 View PostHmm... I just looked at that guide, and this is what I see:
[ATTACH=CONFIG]35512[/ATTACH]
Not a single mention of GPU thread priority checkbox in paragraph about sluggish UI
But then again, most of the experienced people will figure out what it does right away. It's just that I, personally, have very little experience with V-Ray GPU. Everytime I gave it a try, something scared me away very quickly
The tips for the OS UI in the 3.4 FAQ are still valid, of course.
Best,
Blago.Last edited by savage309; 23-01-2017, 06:56 AM.V-Ray fan.
Looking busy around GPUs ...
RTX ON
Comment
-
Originally posted by savage309 View PostThe GPU thread priority is option from 3.5, which is not yet released officially, therefore it is not in the 3.4 docs.
The tips for the OS UI in the 3.4 FAQ are still valid, of course.
Best,
Blago.
Comment
-
Low thread priority guys. And this https://www.youtube.com/watch?v=0CjnkDk481E
I use octane with c4d a lot too, and i believe that RT is better for production. A bit less interactive (though hugely improved in 3.5) but a lot faster than octane, therefore than fstorm too. Try interiors or bright dof/rough highlights an u ll see fstorm bend comparing to RT.
Previewing 80% of the image in seconds and the rest 20 in great time inconsistency is what u get with those engines. RT on the other hand for actual clean image can be really faster than both.
Btw Blago, is there the possibility that switching among gpus and priorities to be done live without reopening RT? Its really convinient this way. Cheers
Comment
Comment