Vray 3.5 gpu and Fstorm gpu 2k scene demo

[Midiaeffects]

thanulee, But in your video Low Thread Priority is NOT enabled, and yet the interactivity looks pretty good. The progressive mode and CUDA should be default in active shade, and I think low thread priority too. The goal of Active Shade is to have fast feedback while working anyway.

Regards,
-Eugenio

[mitviz]

For Rt what does PPP stand for and why lower it?

[savage309]

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

It is doable, of course. Just to be sure, can you tell a use case for that ?

For Rt what does PPP stand for and why lower it?

Do you mean RPP (Rays Per Pixel) ?

Best,
Blago.

[thanulee]

thanulee, But in your video Low Thread Priority is NOT enabled, and yet the interactivity looks pretty good. The progressive mode and CUDA should be default in active shade, and I think low thread priority too. The goal of Active Shade is to have fast feedback while working anyway.

Regards,
-Eugenio

Yes u r right, but he probably uses 1 gpu for the system and 1 for the viewport? in any case low gpu priority means "do not use gpus at 100%" and it must be on when u wanna have interactivity, among other tips u saw in that video. cheers

[mitviz]

It is doable, of course. Just to be sure, can you tell a use case for that ?


Do you mean RPP (Rays Per Pixel) ?

Best,
Blago.

must be rpp it was in the first solution offered at the beginning of the thread so am here like a mad man looking for it, i did a comparison with a random scene using rt cuda and fstorm, all i did with fstorm was hit convert and changed the exposure which seems to be more of a post thing in fstorm, the noise in fstorm seems to be more acceptable than after 15 mins in vray and seems to keep clearing faster, am no expert in gpus especially with vray so would love to know what to change to speed this up, i used active shade model for vray also because i wanted to compare the interactive modes of each

[thanulee]

It is doable, of course. Just to be sure, can you tell a use case for that ?

For example, many times i have a heavy prop that i need to see its full detail to understand whats going on, like bump or sss or glasses, and i want for a moment more gpu power to render faster without re opening the engine. So as i am, i check my other gpu as well, or i change priority and when im done, i just change it again to low and continue working. I dont know if others work like this, but i find it very convinient to boost or release gpu on the fly. Thanks

[mitviz]

For example, many times i have a heavy prop that i need to see its full detail to understand whats going on, like bump or sss or glasses, and i want for a moment more gpu power to render faster without re opening the engine. So as i am, i check my other gpu as well, or i change priority and when im done, i just change it again to low and continue working. I dont know if others work like this, but i find it very convinient to boost or release gpu on the fly. Thanks

where is that low priority for gpu RT option?

[thanulee]

where is that low priority for gpu RT option?

Its in the beta not on the official release yet. Just in vray RT tab, u cant miss it

[AC5L4T3R]

Ah another anti-vray/pro fstorm post from Oren.

[Nicinus]

Ah another anti-vray/pro fstorm post from Oren.

Funny, my interpretation was the opposite, that he is doing this because he cares. I would also like to have maximum interactivity when working.

[thanulee]

Funny, my interpretation was the opposite, that he is doing this because he cares. I would also like to have maximum interactivity when working.

Yes its weird actually, i dont understand the post either. Comparing fstorm with vray is not the best example cause are 2 different approaches.
Vray and Redshift are closer as approach in gpu rendering. If he is posting about fstorm a lot though, is weird. I ve noticed similar threads about Fstorm from people that "cared" in Redshift forums and was actually for promoting the engine. In general i ve also seen alot of fake vray comparisons in fstorm group and i do not like their attitude. Similar situation in octane fb but not as much as fstorm.
Either way, is not bad to compare stuff of any short.

[mitviz]

so what are the settings to get rt running more smoothly? i tried lowering rpp and renders are just much slower, what the heck are you guys doing to get VRAY rt to run so fast on gpu? so much talk from people yet i have not seen faster results using vray RT compared to fstorm RT, i would love to use it more often so if someone could let me know, please let me know

[thanulee]

so what are the settings to get rt running more smoothly? i tried lowering rpp and renders are just much slower, what the heck are you guys doing to get VRAY rt to run so fast on gpu? so much talk from people yet i have not seen faster results using vray RT compared to fstorm RT, i would love to use it more often so if someone could let me know, please let me know

a. Use RT with low priority for previews. Stop comparing Fstorm preview with RT. Its apples vs oranges for many reasons. As u see in the video (its not talk) RT responds pretty same with fstorm/octane. In interactive mode RT is bf/bf and yes no reason to use it like this. If u go production though after u finish previewing ur stuff...

b. Use LC as secondary engine in production mode and render ur final image there. As u will immediately see it will be faster than fstorm cause engine is not unbiased anymore, though its physically correct. Throw as many as render elements as u want in there and u will see 0 time increase. Change whatever filter u like and u will see 0 time increase. I dont know about fstorm but octane doubles the time when u use AOVs. I guess same on Fstorm? (i think Fstorm doesnt even have passes yet, or has some really basic ones, no way near the flexibility of RT).
Want even faster? Use higher noise threshold and add denoiser for ur stills. Check proof of concept here http://dabarti.com/vfx/interior-rend...rmotion-scene/

c. If u wanna use fstorm, use fstorm :P if it fits ur needs, do it. At the end of the day, comparing engines all the time, making posts about it and switching back and forth is a waste of time and unproductive. I dont know why u are in search of vray since u seem happy with fstorm.

ps. Dabarti is showing through tutorials a lot of stuff about RT, Blago is following up on every post, i think there is no such thing as secret settings or "talk". Here is another example of a company that utilises RT for all their jobs http://moxels.com/

[mitviz]

a. Use RT with low priority for previews. Stop comparing Fstorm preview with RT. Its apples vs oranges for many reasons. As u see in the video (its not talk) RT responds pretty same with fstorm/octane. In interactive mode RT is bf/bf and yes no reason to use it like this. If u go production though after u finish previewing ur stuff...

b. Use LC as secondary engine in production mode and render ur final image there. As u will immediately see it will be faster than fstorm cause engine is not unbiased anymore, though its physically correct. Throw as many as render elements as u want in there and u will see 0 time increase. Change whatever filter u like and u will see 0 time increase. I dont know about fstorm but octane doubles the time when u use AOVs. I guess same on Fstorm? (i think Fstorm doesnt even have passes yet, or has some really basic ones, no way near the flexibility of RT).
Want even faster? Use higher noise threshold and add denoiser for ur stills. Check proof of concept here http://dabarti.com/vfx/interior-rend...rmotion-scene/

c. If u wanna use fstorm, use fstorm :P if it fits ur needs, do it. At the end of the day, comparing engines all the time, making posts about it and switching back and forth is a waste of time and unproductive. I dont know why u are in search of vray since u seem happy with fstorm.

ps. Dabarti is showing through tutorials a lot of stuff about RT, Blago is following up on every post, i think there is no such thing as secret settings or "talk". Here is another example of a company that utilises RT for all their jobs http://moxels.com/

actually i use all of them, throw in corona as well, you don't need to know why, just that i do and so do many others, its just how the world works now, when i need vray i use vray, when i need corona i use corona or fstorm, whichever i chose, what's wrong with having choices? personally all the people i know are using more than one engine, i know some of you are really focusing on one for company reasons or just love of it but i like them all personally and i don't see a reason not to have them all, thanks for these tips i will test them

[thanulee]

I didnt ask why. I use different engines myself depending the 3d app or my licences.
But u obviously from ur post u seem happy with fstorm. This aint bad, but insisting about how better it is in vray forums and that u only read about "people talk" while we give u examples and links on how to do stuff, is just idk.. What do u expect from others to say?
We already tried answering ur questions i think. And i guess if u got any more problems support will gladly help u with ur speed issues. Cheers.