OctaneRender™ 2019.2 XB1

Forums: OctaneRender™ 2019.2 XB1
A forum where development builds are posted for testing by the community.
Forum rules
NOTE: The software in this forum is not %100 reliable, they are development builds and are meant for testing by experienced octane users. If you are a new octane user, we recommend to use the current stable release from the 'Commercial Product News & Releases' forum.

Re: OctaneRender™ 2019.2 XB1

Postby Notiusweb » Tue Jul 30, 2019 5:39 pm

Notiusweb Tue Jul 30, 2019 5:39 pm
Wallace - can we get an Opacity channel for the Hair shader?
I like it a lot, but I would like to use it for Transparency based hair meshes!

THX!

Also - Normal/Bump?
Win 10 Pro 64, Xeon E5-2687W v2 (8x 3.40GHz), G.Skill 64 GB DDR3-2400, ASRock X79 Extreme 11
Mobo: 1 Titan Xp, 1 Titan X Pascal
External: 6 Titan X Pascal - 1600W EVGA T2
Plugs: 3DSMax, C4D, Daz, Houdini, Lightwave, Sketchup
User avatar
Notiusweb
Licensed Customer
Licensed Customer
 
Posts: 1162
Joined: Mon Nov 10, 2014 4:51 am

Re: OctaneRender™ 2019.2 XB1

Postby Zay » Tue Jul 30, 2019 5:51 pm

Zay Tue Jul 30, 2019 5:51 pm
Comparing to 2018.1.3, 2019.2 XB1 it is about 8% slower.
Adds 400 MB more of Vram per scene.
Still crashes when using Up-sampling:
Started logging on 30.07.19 19:29:40

OctaneRender Enterprise 2019.2 XB1 (7000000)

Up-sampling error: CUDA_ERROR_ILLEGAL_ADDRESS: an illegal memory access was encountered

CUDA error 700 on device 0: an illegal memory access was encountered
-> failed to copy pitched memory from device

Also reported by another user here viewtopic.php?f=33&t=72219&p=366378#p366378 who is also on Win 7.
Win 7 64bit | 3.4GHz I7 | 64GB ram | 2x GTX 1080Ti - driver 431.36| Modo/ZBrush/Poser
Zay
Licensed Customer
Licensed Customer
 
Posts: 1004
Joined: Sun Jan 17, 2010 2:53 am

Re: OctaneRender™ 2019.2 XB1

Postby mojave » Tue Jul 30, 2019 8:58 pm

mojave Tue Jul 30, 2019 8:58 pm
divasoft wrote:Something breaks when I turn on RTX


Hi,

Could you give more details about what doesn't look right in your scene or share it via PM so we can look into it?

Thank you.
User avatar
mojave
OctaneRender Team
OctaneRender Team
 
Posts: 883
Joined: Sun Feb 08, 2015 10:35 pm

Re: OctaneRender™ 2019.2 XB1

Postby mojave » Tue Jul 30, 2019 9:01 pm

mojave Tue Jul 30, 2019 9:01 pm
divasoft wrote:Emmm...
In this case I do not see the benefits of RTX. No acceleration.


There's a small overhead by using RTX so for very small scenes or certain setups this may be larger than the speedup the HW will give you. We might add some mechanism to enable it just for certain scenes and this case would be one in which it will not.
User avatar
mojave
OctaneRender Team
OctaneRender Team
 
Posts: 883
Joined: Sun Feb 08, 2015 10:35 pm

Re: OctaneRender™ 2019.2 XB1

Postby xevious2501 » Tue Jul 30, 2019 10:39 pm

xevious2501 Tue Jul 30, 2019 10:39 pm
We are all expecting the great boost in speed via RTX, and i dont doubt its presence, but also i think speed boosts will vary on systems setups as well.

I wouldnt doubt cards like the Quadro's having peak benefits over say Geforce rtx cards simply due to the Quadro's increased memory bandwidth.
Also, couple that with NVlink. which as of now only exists in 3 and 4 way support for geforce cards.
I dont understand why NVIDIA are not providing 2way nvlinks for geforce, though they have 2way sli support. most likely to do with sales strategy.

Anyways, you also have to consider the current build being experimental.
Yeah i know. big promises for great speed, but lets all just be a bit patient. after all we're still getting great improvements in features, some long overdue.

But i hear ya. especially when youve purchased big with RTX cards specifically for the benefits of RTX. but no worries its going to pay off.
xevious2501
Licensed Customer
Licensed Customer
 
Posts: 16
Joined: Mon Sep 02, 2013 4:23 am

Re: OctaneRender™ 2019.2 XB1

Postby Notiusweb » Wed Jul 31, 2019 1:56 am

Notiusweb Wed Jul 31, 2019 1:56 am
xevious2501 wrote:We are all expecting the great boost in speed via RTX, and i dont doubt its presence, but also i think speed boosts will vary on systems setups as well.

I wouldnt doubt cards like the Quadro's having peak benefits over say Geforce rtx cards simply due to the Quadro's increased memory bandwidth.
Also, couple that with NVlink. which as of now only exists in 3 and 4 way support for geforce cards.
I dont understand why NVIDIA are not providing 2way nvlinks for geforce, though they have 2way sli support. most likely to do with sales strategy.

Anyways, you also have to consider the current build being experimental.
Yeah i know. big promises for great speed, but lets all just be a bit patient. after all we're still getting great improvements in features, some long overdue.

But i hear ya. especially when youve purchased big with RTX cards specifically for the benefits of RTX. but no worries its going to pay off.


I can conceptually agree in sentiment, but it's kind of like as a user you feel it either works for you or it doesn't. There really needn't be this fancy middle ground of mystery, it is a computation device after all.
I think we just may not know entirely how and in what context we will see a speed boost, so it would be good to know these in a non-vague sense.
Which is why I ask about things like the 3-6 second time scales, does it make sense to expect these time scales to be cut in half, or not.
And, like Divasoft is asking, why am I not seeing a speed boost, yeah well he's right....Why Not? What is 'the scene' where he would see it working, what is it that that scene has...specifically.
Because, if I use a newly released denoiser and there's noise in my render, I would report it. Not out of impatience, but because it is a new release item.
If I up-sample and then it doesn't up-sample, I ask for clarification, maybe I am doing something wrong. Not because I am failing to look at the bigger long game, but because I am testing the Beta and want to see it function as developers articulate it could be working. Just good things to do in a development process.

So, maybe dev team can layout some fundamentals of which scenes will be benefiting most from a RTX Boost?


PS - this next sequence of sentences is an example of impatience:

WHERE IS ANIME KERNEL!!!!!!????
HUH!!!???
WHERE IS IT!!!???
STOP TEASING!!!
ME WANT NOW!!!!!
NOW!!!!
:twisted:
Win 10 Pro 64, Xeon E5-2687W v2 (8x 3.40GHz), G.Skill 64 GB DDR3-2400, ASRock X79 Extreme 11
Mobo: 1 Titan Xp, 1 Titan X Pascal
External: 6 Titan X Pascal - 1600W EVGA T2
Plugs: 3DSMax, C4D, Daz, Houdini, Lightwave, Sketchup
User avatar
Notiusweb
Licensed Customer
Licensed Customer
 
Posts: 1162
Joined: Mon Nov 10, 2014 4:51 am

Re: OctaneRender™ 2019.2 XB1

Postby Iceman9 » Wed Jul 31, 2019 2:23 am

Iceman9 Wed Jul 31, 2019 2:23 am
it's kind of like as a user you feel it either works for you or it doesn't. There really needn't be this fancy middle ground of mystery, it is a computation device after all.
I think we just may not know entirely how and in what context we will see a speed boost, so it would be good to know these in a non-vague sense.


I don't agree with this sentiment at all. Not for most anything in 3d. And not here.

The lighting in one scene doesn't work for the lighting in another. In some scenes you can dial the needed samples down...in other times you must go way up. For premium results materials must often be tweaked per scene. Etc. Our 3d world is a contextual world where operator talent means something (thank God). The devs have been quite consistent and clear in stating that RTX works best in heavy poly scenes. That isn't mysterious at all in terms of what you really need to know.

No one is going to give you a formula to say: your current scene will be 50% faster with RTX enabled...or 180% or -5%. But you do know polygon count/scene complexity yields greater gains.

Maybe at some point when RTX rendering is more mature it will *always* be faster in all scenes. Not now. Adjust accordingly.
Iceman9
Licensed Customer
Licensed Customer
 
Posts: 179
Joined: Sun Jun 09, 2013 1:22 am

Re: OctaneRender™ 2019.2 XB1

Postby Notiusweb » Wed Jul 31, 2019 3:05 am

Notiusweb Wed Jul 31, 2019 3:05 am
Iceman9 wrote:
it's kind of like as a user you feel it either works for you or it doesn't. There really needn't be this fancy middle ground of mystery, it is a computation device after all.
I think we just may not know entirely how and in what context we will see a speed boost, so it would be good to know these in a non-vague sense.


I don't agree with this sentiment at all. Not for most anything in 3d. And not here.

The lighting in one scene doesn't work for the lighting in another. In some scenes you can dial the needed samples down...in other times you must go way up. For premium results materials must often be tweaked per scene. Etc. Our 3d world is a contextual world where operator talent means something (thank God). The devs have been quite consistent and clear in stating that RTX works best in heavy poly scenes. That isn't mysterious at all in terms of what you really need to know.

No one is going to give you a formula to say: your current scene will be 50% faster with RTX enabled...or 180% or -5%. But you do know polygon count/scene complexity yields greater gains.

Maybe at some point when RTX rendering is more mature it will *always* be faster in all scenes. Not now. Adjust accordingly.


You bring up a great technical citation, how many polygons is a heavy poly scene? Like, exactly how many, is there some sort of tier we could see?....
And what does it mean to say "works best" - x4, vs x2? Because if there is a (1) "works best" there would have to be a (2) "works worse", even a possible known (3) "doesn't work at all" scenario.

Like, it's a RTX-Spectrum of some sort, where there must be some parameter metric involved.

So-
Poly situation A : renders x2 RTX BOOST = Poly situation B : Renders x4 RTX BOOST....

does this mean Poly situation A had 1/2 the polys Situation B does?
is that what RTX Speed Boost does, scale-by-Polys?

Because Iceman, think of this...they had this whole Benchmark to show off and trial rendering speed with RTX vs non-RTX.
Well, what was it that was speeding up the render scene when RTX was used...a heavier Poly situation?
So if a scene has more polys, it makes it go faster, but at lower polys, there is less gain, like that?

So a figure with hair would render better under RTX, vs non-rtx, whereas a bald head would render more or less the same in both situations, because of less polys to the scene to begin with, like that right?

See, I don't know if this is the case or not, but this would explain why Divasoft wouldn't see a speed gain in the case of his mat render, as the scene wasn't high poly. So this type of explaining would be very valuable to know!
Win 10 Pro 64, Xeon E5-2687W v2 (8x 3.40GHz), G.Skill 64 GB DDR3-2400, ASRock X79 Extreme 11
Mobo: 1 Titan Xp, 1 Titan X Pascal
External: 6 Titan X Pascal - 1600W EVGA T2
Plugs: 3DSMax, C4D, Daz, Houdini, Lightwave, Sketchup
User avatar
Notiusweb
Licensed Customer
Licensed Customer
 
Posts: 1162
Joined: Mon Nov 10, 2014 4:51 am

Re: OctaneRender™ 2019.2 XB1

Postby Jolbertoquini » Wed Jul 31, 2019 8:34 am

Jolbertoquini Wed Jul 31, 2019 8:34 am
divasoft wrote:Emmm...
In this case I do not see the benefits of RTX. No acceleration.


Hi Guys...

Same here! :o I tried more the 10 scenes and always I have slow renders like 1 or 2 seconds slower, really strange I update drivers to the latest nothing I guess is not working.

Best,
JO
Octane Render for Maya.
https://vimeo.com/jocg/videos
https://www.linkedin.com/in/jocgtd
http://www.hmxmedia.com/
--------------------
Join MAYA OCTANE USERS Skype discussion here :
https://join.skype.com/LXEQaqqfN15w
User avatar
Jolbertoquini
Licensed Customer
Licensed Customer
 
Posts: 962
Joined: Sun Aug 31, 2014 7:08 am
Location: London

Re: OctaneRender™ 2019.2 XB1

Postby Jolbertoquini » Wed Jul 31, 2019 8:37 am

Jolbertoquini Wed Jul 31, 2019 8:37 am
mojave wrote:
divasoft wrote:Something breaks when I turn on RTX


Hi,

Could you give more details about what doesn't look right in your scene or share it via PM so we can look into it?

Thank you.


I tried on the Groot scene the one I send to the siggraph I can send you again the scene without I have 25 seconds 2 RTX 2080ti and with RTX on I have 27seconds.

I will send to you the files.

Cheers,
JO
Octane Render for Maya.
https://vimeo.com/jocg/videos
https://www.linkedin.com/in/jocgtd
http://www.hmxmedia.com/
--------------------
Join MAYA OCTANE USERS Skype discussion here :
https://join.skype.com/LXEQaqqfN15w
User avatar
Jolbertoquini
Licensed Customer
Licensed Customer
 
Posts: 962
Joined: Sun Aug 31, 2014 7:08 am
Location: London
PreviousNext

Return to Development Build Releases


Who is online

Users browsing this forum: No registered users and 1 guest

Thu Nov 21, 2019 1:02 am [ UTC ]