Daz Studio Pro BETA - version 4.11.0.335! (*UPDATED*)

1111214161732

Comments

  • shaneseymourstudioshaneseymourstudio Posts: 383
    edited October 2018

    Wow...I am very impressed with 4.11.0.236....

    Maybe it is just my own ignorance in past but the camera seems to behave more like expected with DoF...

    Also, nominal luminance seems to calculate correctly, allowing me to get much faster renders without the noise in a linear workflow...

    I still notice some limitations when using canvases with denoiser but I know this is still so new...with 30min or less renders on very complex scenes though, without denoiser, I am so happy:)

    Bravo team! From my point of view and daily usage bravo!!

     

    P.S....Meant to add that I don't know what has changed for hair but I truly love the results I am getting now!!

     

    Post edited by shaneseymourstudio on
  • MelanieLMelanieL Posts: 7,464
    edited October 2018

    I've found a problem with 3Delight render of a high-res figure (I don't necessarily mean HD, just the high res that for example base G8F loads) and clothing with smoothing applied (which many recent clothing items do out of the box). The skin goes semi-transparent in render. If I set the figure to Base res OR turn smoothing off OR set "Fit To" to none (obviously not very useful) then the skin renders just fine.

    This isn't specifically a DS4.11.0.126 problem  - I reinstalled 4.10.0.123 and it happens there too. It was definitely working in DS4.6 (which I still have installed alongside the beta) but stopped working somewhere along the line.

    Anyone else able to reproduce this?

    I raised a ticket about it, but I still have an open ticket about somewhat similar problems with shadows and reflections which I raised three years ago, so I don't have much optimism.

    ETA: By the way the skin shader type doesn't seem to be the problem - the G8F's are using AoA Subsurface, the G2F's are Ubersurface.

    G8F-FitTo-HighRes.jpg
    544 x 601 - 80K
    G8F-FitTo-BaseRes.jpg
    471 x 594 - 72K
    G8F-Not-Fit.jpg
    499 x 636 - 78K
    G2F-HighRes-SmoothingOn.jpg
    967 x 792 - 101K
    G2F-HighRes-SmoothingOff.jpg
    983 x 825 - 105K
    G2F-HighRes-SmoothingOn-DS4.6.jpg
    916 x 801 - 77K
    Post edited by MelanieL on
  • jestmartjestmart Posts: 4,449

    I thought that is do to having motion blur checked.

  • nonesuch00nonesuch00 Posts: 18,293

    Before I start submitting a support ticket, is anyone else experiencing complete system freezes during some renders in the beta with one of the new RTX cards? I got an RTX 2080 Ti this week which reliably freezes my system on certain renders, requiring a hard reboot. I've put the card through 3DMark and games without issues, so I suspect the problem is in the Daz beta. Since I haven't been following all the conversations about this topic, I'd just like to hear if I'm alone in this or not.

    Edit: It appears I have a faulty GPU. I'll be getting it replaced.

    In my exprience that is usually out of memory. If you open the problem scene, change the optimization to memory from speed and the problem goes away you know that was the problem with the caveat that optimized for speed or memory the disk that contains your swap space needs to have enough gigabytes (upwards of over a dozen gigabytes for some scenes) to expand available swap.

  • AalaAala Posts: 140

    Wow...I am very impressed with 4.11.0.236....

    Maybe it is just my own ignorance in past but the camera seems to behave more like expected with DoF...

    Also, nominal luminance seems to calculate correctly, allowing me to get much faster renders without the noise in a linear workflow...

    I still notice some limitations when using canvases with denoiser but I know this is still so new...with 30min or less renders on very complex scenes though, without denoiser, I am so happy:)

    Bravo team! From my point of view and daily usage bravo!!

     

    P.S....Meant to add that I don't know what has changed for hair but I truly love the results I am getting now!!

     

    Can you give us an example of what exactly has changed for you to make the renders faster? Personally, I'm getting slower results by quite a margin. I just tested out the scene quoted on this post:

    https://www.daz3d.com/forums/discussion/comment/4009596/#Comment_4009596

    With a 1080 Ti,

    On 4.10 I'm getting 5 mins 40 seconds 

    On 4.11 I'm getting 7 mins 40 seconds.

    That's almost 50% slower. Still haven't tested it out with more complex scenes, but it doesn't sound particularly good to me.

  • shaneseymourstudioshaneseymourstudio Posts: 383
    edited October 2018
    Aala said:

    Wow...I am very impressed with 4.11.0.236....

    Maybe it is just my own ignorance in past but the camera seems to behave more like expected with DoF...

    Also, nominal luminance seems to calculate correctly, allowing me to get much faster renders without the noise in a linear workflow...

    I still notice some limitations when using canvases with denoiser but I know this is still so new...with 30min or less renders on very complex scenes though, without denoiser, I am so happy:)

    Bravo team! From my point of view and daily usage bravo!!

     

    P.S....Meant to add that I don't know what has changed for hair but I truly love the results I am getting now!!

     

    Can you give us an example of what exactly has changed for you to make the renders faster? Personally, I'm getting slower results by quite a margin. I just tested out the scene quoted on this post:

    https://www.daz3d.com/forums/discussion/comment/4009596/#Comment_4009596

    With a 1080 Ti,

    On 4.10 I'm getting 5 mins 40 seconds 

    On 4.11 I'm getting 7 mins 40 seconds.

    That's almost 50% slower. Still haven't tested it out with more complex scenes, but it doesn't sound particularly good to me.

    I am not able to compare before or after regarding that test scene you are using but the difference for me is with the correctly calculated Nominal luminance. I put the nominal luminance value to be the same as what my exposure value is so if my exposure is at 13 I set my nominal luminance to match. Nominal luminance helps with the noise by telling iRay where to look with that value. Also, and for some reason it shows up sometimes and sometimes it is not an option, I combine that with using the noise filter set to 1, 1 .05, forget the individual setting names for those values (can check when home) but they appear when you turn on the noise filter which, again, does not always appear as an option in the Beta for some reason. I think they were taking out that option but sometimes I am able to get to it, I suspect it is when I open a scene saved from a previous version but not sure.

    There are other options that I use which may play a part in how fast my renders are but I cannot recall all the values but will try to upload some screenshots later for my typical setup. I work with linear values in the tone map pane for highlights and crush black set to 0 to that no clipping occurs with the lighting which may affect the render times. I have not looked at the test scene at all for what it is setup to use.

    I also render using a quadro p5000 card in a dell workstation laptop...this could be a big factor with the way the drivers work on that card vs others...so not sure with nothing else to run comparisons with but for me the speed is increased and I am getting much better results in most every area (In this version of the beta 4.11.0.236 vs current stable release 4.10 and even the prior beta I had installed) with exception that denoising seems to be a little more wonky now and totally broken when trying to render to canvases.

    Post edited by shaneseymourstudio on
  • btymoffersbtymoffers Posts: 22
    edited October 2018

    The Active Pose crashing Daz3d whenever using "Pin at Both" option or "Pin at End" is driving me insane. I have found that it appears to happen whenever you pin a few spots and attempt to move anything. However there has been a few times where I have pinned only one or two location and then it works aslong as I don't touch the Hip or the Head. But this really needs to be fixed more then anything. The Active Pose Tool is among my most used tools. Animating simply isn't do able in Daz3d without it.

     

    To any Daz people wanting to know how to replicate the issue.

    1. Add a Genesis Character to your scene

    2. Pin the feet, and the Head. Enable Pin At Both in Tool Settings.

    3. Move anything with Active Pose Tool.

    4. Daz Studio Crashes.

    Post edited by btymoffers on
  • ChoholeChohole Posts: 33,604
    edited October 2018

    btymoffers said:

    The Active Pose crashing Daz3d whenever using "Pin at Both" option or "Pin at End" is driving me insane. I have found that it appears to happen whenever you pin a few spots and attempt to move anything. However there has been a few times where I have pinned only one or two location and then it works aslong as I don't touch the Hip or the Head. But this really needs to be fixed more then anything. The Active Pose Tool is among my most used tools. Animating simply isn't do able in Daz3d without it.

     

    To any Daz people wanting to know how to replicate the issue.

    1. Add a Genesis Character to your scene

    2. Pin the feet, and the Head. Enable Pin At Both in Tool Settings.

    3. Move anything with Active Pose Tool.

    4. Daz Studio Crashes.

     

    Please report this to Daz, using a support ticket  citing Tech Support as the relevant dept.
    https://helpdaz.zendesk.com/hc/en-us/requests/new

    Thank you

    Post edited by Chohole on
  • AalaAala Posts: 140
    Aala said:

    Wow...I am very impressed with 4.11.0.236....

    Maybe it is just my own ignorance in past but the camera seems to behave more like expected with DoF...

    Also, nominal luminance seems to calculate correctly, allowing me to get much faster renders without the noise in a linear workflow...

    I still notice some limitations when using canvases with denoiser but I know this is still so new...with 30min or less renders on very complex scenes though, without denoiser, I am so happy:)

    Bravo team! From my point of view and daily usage bravo!!

     

    P.S....Meant to add that I don't know what has changed for hair but I truly love the results I am getting now!!

     

    Can you give us an example of what exactly has changed for you to make the renders faster? Personally, I'm getting slower results by quite a margin. I just tested out the scene quoted on this post:

    https://www.daz3d.com/forums/discussion/comment/4009596/#Comment_4009596

    With a 1080 Ti,

    On 4.10 I'm getting 5 mins 40 seconds 

    On 4.11 I'm getting 7 mins 40 seconds.

    That's almost 50% slower. Still haven't tested it out with more complex scenes, but it doesn't sound particularly good to me.

    I am not able to compare before or after regarding that test scene you are using but the difference for me is with the correctly calculated Nominal luminance. I put the nominal luminance value to be the same as what my exposure value is so if my exposure is at 13 I set my nominal luminance to match. Nominal luminance helps with the noise by telling iRay where to look with that value. Also, and for some reason it shows up sometimes and sometimes it is not an option, I combine that with using the noise filter set to 1, 1 .05, forget the individual setting names for those values (can check when home) but they appear when you turn on the noise filter which, again, does not always appear as an option in the Beta for some reason. I think they were taking out that option but sometimes I am able to get to it, I suspect it is when I open a scene saved from a previous version but not sure.

    There are other options that I use which may play a part in how fast my renders are but I cannot recall all the values but will try to upload some screenshots later for my typical setup. I work with linear values in the tone map pane for highlights and crush black set to 0 to that no clipping occurs with the lighting which may affect the render times. I have not looked at the test scene at all for what it is setup to use.

    I also render using a quadro p5000 card in a dell workstation laptop...this could be a big factor with the way the drivers work on that card vs others...so not sure with nothing else to run comparisons with but for me the speed is increased and I am getting much better results in most every area (In this version of the beta 4.11.0.296 vs current stable release 4.10 and even the prior beta I had installed) with exception that denoising seems to be a little more wonky now and totally broken when trying to render to canvases.

    So if I'm understanding this correctly, you put the nominal luminance to 13 if your exposure is set to 13? Wouldn't that make the renders quite dark and the light quite flat? Or am I missing something here?

    Would be interesting and appreciated if you tested the scene I quoted with both DAZ versions (4.10 and the 4.11 latest beta) to see how it stacks up with your card. It's really hard to find out what the best and most optimal settings for Iray are. 

  • Aala said:
    Aala said:

    Wow...I am very impressed with 4.11.0.236....

    Maybe it is just my own ignorance in past but the camera seems to behave more like expected with DoF...

    Also, nominal luminance seems to calculate correctly, allowing me to get much faster renders without the noise in a linear workflow...

    I still notice some limitations when using canvases with denoiser but I know this is still so new...with 30min or less renders on very complex scenes though, without denoiser, I am so happy:)

    Bravo team! From my point of view and daily usage bravo!!

     

    P.S....Meant to add that I don't know what has changed for hair but I truly love the results I am getting now!!

     

    Can you give us an example of what exactly has changed for you to make the renders faster? Personally, I'm getting slower results by quite a margin. I just tested out the scene quoted on this post:

    https://www.daz3d.com/forums/discussion/comment/4009596/#Comment_4009596

    With a 1080 Ti,

    On 4.10 I'm getting 5 mins 40 seconds 

    On 4.11 I'm getting 7 mins 40 seconds.

    That's almost 50% slower. Still haven't tested it out with more complex scenes, but it doesn't sound particularly good to me.

    I am not able to compare before or after regarding that test scene you are using but the difference for me is with the correctly calculated Nominal luminance. I put the nominal luminance value to be the same as what my exposure value is so if my exposure is at 13 I set my nominal luminance to match. Nominal luminance helps with the noise by telling iRay where to look with that value. Also, and for some reason it shows up sometimes and sometimes it is not an option, I combine that with using the noise filter set to 1, 1 .05, forget the individual setting names for those values (can check when home) but they appear when you turn on the noise filter which, again, does not always appear as an option in the Beta for some reason. I think they were taking out that option but sometimes I am able to get to it, I suspect it is when I open a scene saved from a previous version but not sure.

    There are other options that I use which may play a part in how fast my renders are but I cannot recall all the values but will try to upload some screenshots later for my typical setup. I work with linear values in the tone map pane for highlights and crush black set to 0 to that no clipping occurs with the lighting which may affect the render times. I have not looked at the test scene at all for what it is setup to use.

    I also render using a quadro p5000 card in a dell workstation laptop...this could be a big factor with the way the drivers work on that card vs others...so not sure with nothing else to run comparisons with but for me the speed is increased and I am getting much better results in most every area (In this version of the beta 4.11.0.296 vs current stable release 4.10 and even the prior beta I had installed) with exception that denoising seems to be a little more wonky now and totally broken when trying to render to canvases.

    So if I'm understanding this correctly, you put the nominal luminance to 13 if your exposure is set to 13? Wouldn't that make the renders quite dark and the light quite flat? Or am I missing something here?

    Would be interesting and appreciated if you tested the scene I quoted with both DAZ versions (4.10 and the 4.11 latest beta) to see how it stacks up with your card. It's really hard to find out what the best and most optimal settings for Iray are. 

    Correct. It will make the lighting flatter working with linear space but allows you to get a larger range to work with in post without blowing out or losing detail and ultimately is how I prefer to work incorporating post work as I need after rendering. I will post something this evening (I will make sure to get to this tonight) with my settings and a result from the render and you can try yourself with a scene to see what you think about the result. Even if you do not use canvases the scene will have alot more detail rendered out if saved as tiff at least.

    Like I mentioned concerning the test scene, it may not prove to be any better than times you reported with those settings. But the scenes I setup and settings I use are not likely associated at all with that scene. My results are from my particular workflow and in my case with my current setup and drivers and the current BETA, my renders are completed quicker due to noise being filtered out more efficiently, I suspect due to the nominal luminance calculating correctly, or at least better than in previous versions. 

    One note is that I setup each of my scenes in this current beta fresh. So I am not rendering scenes setup from older versions typically which may make a difference as well. I am only comparing my results from the previous BETA I used to this one. I have not even touched 4.10 in months.

  • shaneseymourstudioshaneseymourstudio Posts: 383
    edited October 2018
    Aala said:
    Aala said:

    Wow...I am very impressed with 4.11.0.236....

    Maybe it is just my own ignorance in past but the camera seems to behave more like expected with DoF...

    Also, nominal luminance seems to calculate correctly, allowing me to get much faster renders without the noise in a linear workflow...

    I still notice some limitations when using canvases with denoiser but I know this is still so new...with 30min or less renders on very complex scenes though, without denoiser, I am so happy:)

    Bravo team! From my point of view and daily usage bravo!!

     

    P.S....Meant to add that I don't know what has changed for hair but I truly love the results I am getting now!!

     

    Can you give us an example of what exactly has changed for you to make the renders faster? Personally, I'm getting slower results by quite a margin. I just tested out the scene quoted on this post:

    https://www.daz3d.com/forums/discussion/comment/4009596/#Comment_4009596

    With a 1080 Ti,

    On 4.10 I'm getting 5 mins 40 seconds 

    On 4.11 I'm getting 7 mins 40 seconds.

    That's almost 50% slower. Still haven't tested it out with more complex scenes, but it doesn't sound particularly good to me.

    I am not able to compare before or after regarding that test scene you are using but the difference for me is with the correctly calculated Nominal luminance. I put the nominal luminance value to be the same as what my exposure value is so if my exposure is at 13 I set my nominal luminance to match. Nominal luminance helps with the noise by telling iRay where to look with that value. Also, and for some reason it shows up sometimes and sometimes it is not an option, I combine that with using the noise filter set to 1, 1 .05, forget the individual setting names for those values (can check when home) but they appear when you turn on the noise filter which, again, does not always appear as an option in the Beta for some reason. I think they were taking out that option but sometimes I am able to get to it, I suspect it is when I open a scene saved from a previous version but not sure.

    There are other options that I use which may play a part in how fast my renders are but I cannot recall all the values but will try to upload some screenshots later for my typical setup. I work with linear values in the tone map pane for highlights and crush black set to 0 to that no clipping occurs with the lighting which may affect the render times. I have not looked at the test scene at all for what it is setup to use.

    I also render using a quadro p5000 card in a dell workstation laptop...this could be a big factor with the way the drivers work on that card vs others...so not sure with nothing else to run comparisons with but for me the speed is increased and I am getting much better results in most every area (In this version of the beta 4.11.0.296 vs current stable release 4.10 and even the prior beta I had installed) with exception that denoising seems to be a little more wonky now and totally broken when trying to render to canvases.

    So if I'm understanding this correctly, you put the nominal luminance to 13 if your exposure is set to 13? Wouldn't that make the renders quite dark and the light quite flat? Or am I missing something here?

    Would be interesting and appreciated if you tested the scene I quoted with both DAZ versions (4.10 and the 4.11 latest beta) to see how it stacks up with your card. It's really hard to find out what the best and most optimal settings for Iray are. 

    Correct. It will make the lighting flatter working with linear space but allows you to get a larger range to work with in post without blowing out or losing detail and ultimately is how I prefer to work incorporating post work as I need after rendering. I will post something this evening (I will make sure to get to this tonight) with my settings and a result from the render and you can try yourself with a scene to see what you think about the result. Even if you do not use canvases the scene will have alot more detail rendered out if saved as tiff at least.

    Like I mentioned concerning the test scene, it may not prove to be any better than times you reported with those settings. But the scenes I setup and settings I use are not likely associated at all with that scene. My results are from my particular workflow and in my case with my current setup and drivers and the current BETA, my renders are completed quicker due to noise being filtered out more efficiently, I suspect due to the nominal luminance calculating correctly, or at least better than in previous versions. 

    One note is that I setup each of my scenes in this current beta fresh. So I am not rendering scenes setup from older versions typically which may make a difference as well. I am only comparing my results from the previous BETA I used to this one. I have not even touched 4.10 in months.

    Wow, I got a shock after testing the test scene...I didn't try doing the scene between this beta and 4.10, instead I just changed the nominal luminance and as I expected it cut my render time down from 12:18 seconds from the scene rendered with default from the .duf to 2:38 with nominal set to 13...however, what I didn't realize is just what it affected in the scene. The images below will show the drastic difference. I could not tell this in my renders until seeing it in this test scene but once I saw this I went back to some of my previous renders in this beta and loaded the scenes and then realized how dark, as you asked about, it was making the scene:

     

    DEFAULT:

    TestSceneDefault

    TestSceneDefaultRendered

     

    NOMINAL SET TO 13

    TestSceneNominal

    TestSceneNominalRendered

     

    You can see that the nominal version seemed to remove all the brightness. I am going to test some things on my prior scenes and see what the time difference is but I suspect that the reason I was getting such fast renders was that the nominal is indeed taking away from my scenes, though I didn't notice because the result was ok to my eyes. I will be back within next couple of days with anything further I find.

    TestScene4.11Default.JPG
    3437 x 1440 - 491K
    TestScene4.11Nominal.JPG
    3433 x 1440 - 381K
    TestScene4.11Default.png
    1000 x 1300 - 2M
    TestScene4.11Nominal.png
    1000 x 1300 - 2M
    Post edited by shaneseymourstudio on
  • olesehenolesehen Posts: 15
    edited October 2018

    I have encountred an issue with the beta , not sure if this has something to do with my setup or something else .  I currently run the newest drivers ( tried older drivers also)  416.34  and when I run a render in the beta with different cards it will not release the VRAM anymore . I have tried both my old GTX 1080 and my new 2080 Ti .  one render then it will only render on CPU until I have exit the application .   this is highly annoying ..  Same thing on both cards .  when I try the same in 4.10 its fine .. obvisouley I Can not try the 2080 as its not supported in 4.10 but the 1080 works fine . 

    I am prettys sure I tried the beta earlier without these issues with the 1080 so perhaps something setting wise has changed but I am unsure what ..  all the usal stuff like optix on/off and denoiser on / off  has been tested without any result .  if anyone has an idé I could try that would work would make me a happy camper .. also the 2080 is working fine in any other application like benchmarks and games so it does not seem to be harware related .

     

     

    *EDIT*

    log file indicates something fishy with de deallocation of the memory 

    2018-10-26 08:42:52.335 Iray INFO - module:category(IRAY:RENDER):   1.7   IRAY   rend info : Allocating 1-layer frame buffer
    2018-10-26 08:42:52.523 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00001 iterations after 7.542s.
    2018-10-26 08:42:53.243 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00002 iterations after 8.266s.
    2018-10-26 08:42:53.992 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00003 iterations after 9.021s.
    2018-10-26 08:42:54.742 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00004 iterations after 9.766s.
    2018-10-26 08:42:55.492 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00005 iterations after 10.514s.
    2018-10-26 08:42:56.258 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00006 iterations after 11.285s.
    2018-10-26 08:42:57.070 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00007 iterations after 12.098s.
    2018-10-26 08:42:57.945 Iray INFO - module:category(POST:RENDER):   1.0   POST   rend info : using cuda device 0, "GeForce RTX 2080 Ti" (7.5), buffers: fp16, cuDNN v7101, rt v9000
    2018-10-26 08:43:04.685 Iray INFO - module:category(POST:RENDER):   1.0   POST   rend info : layers created for resolution 2058 1254, inp 6, outp 3, cuDNN memory 16.0 MB, total 900.1 MB
    2018-10-26 08:43:04.716 WARNING: ..\..\..\..\..\src\pluginsource\DzIrayRender\dzneuraymgr.cpp(304): Iray ERROR - module:category(IRAY:RENDER):   1.15  IRAY   rend error: CUDA device 0 (GeForce RTX 2080 Ti): an illegal memory access was encountered (while launching CUDA renderer in core_renderer_wf.cpp:796)
    2018-10-26 08:43:04.732 WARNING: ..\..\..\..\..\src\pluginsource\DzIrayRender\dzneuraymgr.cpp(304): Iray ERROR - module:category(POST:RENDER):   1.0   POST   rend error: CUDNN_STATUS_EXECUTION_FAILED
    2018-10-26 08:43:04.732 WARNING: ..\..\..\..\..\src\pluginsource\DzIrayRender\dzneuraymgr.cpp(304): Iray ERROR - module:category(IRAY:RENDER):   1.15  IRAY   rend error: CUDA device 0 (GeForce RTX 2080 Ti): Failed to launch renderer
    2018-10-26 08:43:04.732 WARNING: ..\..\..\..\..\src\pluginsource\DzIrayRender\dzneuraymgr.cpp(304): Iray ERROR - module:category(POST:RENDER):   1.0   POST   rend error: CUDNN_STATUS_EXECUTION_FAILED
    2018-10-26 08:43:04.763 WARNING: ..\..\..\..\..\src\pluginsource\DzIrayRender\dzneuraymgr.cpp(304): Iray ERROR - module:category(IRAY:RENDER):   1.7   IRAY   rend error: CUDA device 0 (GeForce RTX 2080 Ti): Device failed while rendering
    2018-10-26 08:43:04.763 WARNING: ..\..\..\..\..\src\pluginsource\DzIrayRender\dzneuraymgr.cpp(304): Iray ERROR - module:category(POST:RENDER):   1.0   POST   rend error: CUDNN_STATUS_EXECUTION_FAILED
    2018-10-26 08:43:04.763 WARNING: ..\..\..\..\..\src\pluginsource\DzIrayRender\dzneuraymgr.cpp(304): Iray ERROR - module:category(IRAY:RENDER):   1.7   IRAY   rend error: CUDA device 0 (GeForce RTX 2080 Ti): an illegal memory access was encountered (while initializing memory buffer)
    2018-10-26 08:43:04.826 WARNING: ..\..\..\..\..\src\pluginsource\DzIrayRender\dzneuraymgr.cpp(304): Iray ERROR - module:category(POST:RENDER):   1.0   POST   rend error: CUDNN_STATUS_EXECUTION_FAILED
    2018-10-26 08:43:04.826 WARNING: ..\..\..\..\..\src\pluginsource\DzIrayRender\dzneuraymgr.cpp(304): Iray ERROR - module:category(IRAY:RENDER):   1.7   IRAY   rend error: CUDA device 0 (GeForce RTX 2080 Ti): an illegal memory access was encountered (while de-allocating memory)

     

     

    **EDIT 2 **

    Seems the main issue is the denoiser as its set to kick in at iteration 8 .  I turned it off again and ran a series of tests .. and even if the VRAM is set to around 10 GB it does not jump to CPU .. then again a bit fishy that the same scene uses around 6 GB of memory for the same scen in 4.10 with the GTX 1080 .. :| 

     

     

    Post edited by olesehen on
  • AalaAala Posts: 140
    Aala said:
    Aala said:

    Wow...I am very impressed with 4.11.0.236....

    Maybe it is just my own ignorance in past but the camera seems to behave more like expected with DoF...

    Also, nominal luminance seems to calculate correctly, allowing me to get much faster renders without the noise in a linear workflow...

    I still notice some limitations when using canvases with denoiser but I know this is still so new...with 30min or less renders on very complex scenes though, without denoiser, I am so happy:)

    Bravo team! From my point of view and daily usage bravo!!

     

    P.S....Meant to add that I don't know what has changed for hair but I truly love the results I am getting now!!

     

    Can you give us an example of what exactly has changed for you to make the renders faster? Personally, I'm getting slower results by quite a margin. I just tested out the scene quoted on this post:

    https://www.daz3d.com/forums/discussion/comment/4009596/#Comment_4009596

    With a 1080 Ti,

    On 4.10 I'm getting 5 mins 40 seconds 

    On 4.11 I'm getting 7 mins 40 seconds.

    That's almost 50% slower. Still haven't tested it out with more complex scenes, but it doesn't sound particularly good to me.

    I am not able to compare before or after regarding that test scene you are using but the difference for me is with the correctly calculated Nominal luminance. I put the nominal luminance value to be the same as what my exposure value is so if my exposure is at 13 I set my nominal luminance to match. Nominal luminance helps with the noise by telling iRay where to look with that value. Also, and for some reason it shows up sometimes and sometimes it is not an option, I combine that with using the noise filter set to 1, 1 .05, forget the individual setting names for those values (can check when home) but they appear when you turn on the noise filter which, again, does not always appear as an option in the Beta for some reason. I think they were taking out that option but sometimes I am able to get to it, I suspect it is when I open a scene saved from a previous version but not sure.

    There are other options that I use which may play a part in how fast my renders are but I cannot recall all the values but will try to upload some screenshots later for my typical setup. I work with linear values in the tone map pane for highlights and crush black set to 0 to that no clipping occurs with the lighting which may affect the render times. I have not looked at the test scene at all for what it is setup to use.

    I also render using a quadro p5000 card in a dell workstation laptop...this could be a big factor with the way the drivers work on that card vs others...so not sure with nothing else to run comparisons with but for me the speed is increased and I am getting much better results in most every area (In this version of the beta 4.11.0.296 vs current stable release 4.10 and even the prior beta I had installed) with exception that denoising seems to be a little more wonky now and totally broken when trying to render to canvases.

    So if I'm understanding this correctly, you put the nominal luminance to 13 if your exposure is set to 13? Wouldn't that make the renders quite dark and the light quite flat? Or am I missing something here?

    Would be interesting and appreciated if you tested the scene I quoted with both DAZ versions (4.10 and the 4.11 latest beta) to see how it stacks up with your card. It's really hard to find out what the best and most optimal settings for Iray are. 

    Correct. It will make the lighting flatter working with linear space but allows you to get a larger range to work with in post without blowing out or losing detail and ultimately is how I prefer to work incorporating post work as I need after rendering. I will post something this evening (I will make sure to get to this tonight) with my settings and a result from the render and you can try yourself with a scene to see what you think about the result. Even if you do not use canvases the scene will have alot more detail rendered out if saved as tiff at least.

    Like I mentioned concerning the test scene, it may not prove to be any better than times you reported with those settings. But the scenes I setup and settings I use are not likely associated at all with that scene. My results are from my particular workflow and in my case with my current setup and drivers and the current BETA, my renders are completed quicker due to noise being filtered out more efficiently, I suspect due to the nominal luminance calculating correctly, or at least better than in previous versions. 

    One note is that I setup each of my scenes in this current beta fresh. So I am not rendering scenes setup from older versions typically which may make a difference as well. I am only comparing my results from the previous BETA I used to this one. I have not even touched 4.10 in months.

    Wow, I got a shock after testing the test scene...I didn't try doing the scene between this beta and 4.10, instead I just changed the nominal luminance and as I expected it cut my render time down from 12:18 seconds from the scene rendered with default from the .duf to 2:38 with nominal set to 13...however, what I didn't realize is just what it affected in the scene. The images below will show the drastic difference. I could not tell this in my renders until seeing it in this test scene but once I saw this I went back to some of my previous renders in this beta and loaded the scenes and then realized how dark, as you asked about, it was making the scene:

     

    DEFAULT:

    TestSceneDefault

    TestSceneDefaultRendered

     

    NOMINAL SET TO 13

    TestSceneNominal

    TestSceneNominalRendered

     

    You can see that the nominal version seemed to remove all the brightness. I am going to test some things on my prior scenes and see what the time difference is but I suspect that the reason I was getting such fast renders was that the nominal is indeed taking away from my scenes, though I didn't notice because the result was ok to my eyes. I will be back within next couple of days with anything further I find.

    That's pretty interesting though! Thanks for the tests! I do use nominal luminance too in my renders, though my values are in the range of around 300-500. I didn't expect the drastic difference in render time though. 12 minutes to 2 minutes is quite a lot.

  • Aala said:
    Aala said:
    Aala said:

    Wow...I am very impressed with 4.11.0.236....

    Maybe it is just my own ignorance in past but the camera seems to behave more like expected with DoF...

    Also, nominal luminance seems to calculate correctly, allowing me to get much faster renders without the noise in a linear workflow...

    I still notice some limitations when using canvases with denoiser but I know this is still so new...with 30min or less renders on very complex scenes though, without denoiser, I am so happy:)

    Bravo team! From my point of view and daily usage bravo!!

     

    P.S....Meant to add that I don't know what has changed for hair but I truly love the results I am getting now!!

     

    Can you give us an example of what exactly has changed for you to make the renders faster? Personally, I'm getting slower results by quite a margin. I just tested out the scene quoted on this post:

    https://www.daz3d.com/forums/discussion/comment/4009596/#Comment_4009596

    With a 1080 Ti,

    On 4.10 I'm getting 5 mins 40 seconds 

    On 4.11 I'm getting 7 mins 40 seconds.

    That's almost 50% slower. Still haven't tested it out with more complex scenes, but it doesn't sound particularly good to me.

    I am not able to compare before or after regarding that test scene you are using but the difference for me is with the correctly calculated Nominal luminance. I put the nominal luminance value to be the same as what my exposure value is so if my exposure is at 13 I set my nominal luminance to match. Nominal luminance helps with the noise by telling iRay where to look with that value. Also, and for some reason it shows up sometimes and sometimes it is not an option, I combine that with using the noise filter set to 1, 1 .05, forget the individual setting names for those values (can check when home) but they appear when you turn on the noise filter which, again, does not always appear as an option in the Beta for some reason. I think they were taking out that option but sometimes I am able to get to it, I suspect it is when I open a scene saved from a previous version but not sure.

    There are other options that I use which may play a part in how fast my renders are but I cannot recall all the values but will try to upload some screenshots later for my typical setup. I work with linear values in the tone map pane for highlights and crush black set to 0 to that no clipping occurs with the lighting which may affect the render times. I have not looked at the test scene at all for what it is setup to use.

    I also render using a quadro p5000 card in a dell workstation laptop...this could be a big factor with the way the drivers work on that card vs others...so not sure with nothing else to run comparisons with but for me the speed is increased and I am getting much better results in most every area (In this version of the beta 4.11.0.296 vs current stable release 4.10 and even the prior beta I had installed) with exception that denoising seems to be a little more wonky now and totally broken when trying to render to canvases.

    So if I'm understanding this correctly, you put the nominal luminance to 13 if your exposure is set to 13? Wouldn't that make the renders quite dark and the light quite flat? Or am I missing something here?

    Would be interesting and appreciated if you tested the scene I quoted with both DAZ versions (4.10 and the 4.11 latest beta) to see how it stacks up with your card. It's really hard to find out what the best and most optimal settings for Iray are. 

    Correct. It will make the lighting flatter working with linear space but allows you to get a larger range to work with in post without blowing out or losing detail and ultimately is how I prefer to work incorporating post work as I need after rendering. I will post something this evening (I will make sure to get to this tonight) with my settings and a result from the render and you can try yourself with a scene to see what you think about the result. Even if you do not use canvases the scene will have alot more detail rendered out if saved as tiff at least.

    Like I mentioned concerning the test scene, it may not prove to be any better than times you reported with those settings. But the scenes I setup and settings I use are not likely associated at all with that scene. My results are from my particular workflow and in my case with my current setup and drivers and the current BETA, my renders are completed quicker due to noise being filtered out more efficiently, I suspect due to the nominal luminance calculating correctly, or at least better than in previous versions. 

    One note is that I setup each of my scenes in this current beta fresh. So I am not rendering scenes setup from older versions typically which may make a difference as well. I am only comparing my results from the previous BETA I used to this one. I have not even touched 4.10 in months.

    Wow, I got a shock after testing the test scene...I didn't try doing the scene between this beta and 4.10, instead I just changed the nominal luminance and as I expected it cut my render time down from 12:18 seconds from the scene rendered with default from the .duf to 2:38 with nominal set to 13...however, what I didn't realize is just what it affected in the scene. The images below will show the drastic difference. I could not tell this in my renders until seeing it in this test scene but once I saw this I went back to some of my previous renders in this beta and loaded the scenes and then realized how dark, as you asked about, it was making the scene:

     

    DEFAULT:

    TestSceneDefault

    TestSceneDefaultRendered

     

    NOMINAL SET TO 13

    TestSceneNominal

    TestSceneNominalRendered

     

    You can see that the nominal version seemed to remove all the brightness. I am going to test some things on my prior scenes and see what the time difference is but I suspect that the reason I was getting such fast renders was that the nominal is indeed taking away from my scenes, though I didn't notice because the result was ok to my eyes. I will be back within next couple of days with anything further I find.

    That's pretty interesting though! Thanks for the tests! I do use nominal luminance too in my renders, though my values are in the range of around 300-500. I didn't expect the drastic difference in render time though. 12 minutes to 2 minutes is quite a lot.

    Ya I tested with some higher values as well as negative values and at 160 I was able to achieve the same look in that scene as the default but render time was same as default at 12ish min. Thought I had a magic setting but nope, not if I want all the light details.

  • gederixgederix Posts: 390

    The Active Pose crashing Daz3d whenever using "Pin at Both" option or "Pin at End" is driving me insane. I have found that it appears to happen whenever you pin a few spots and attempt to move anything. However there has been a few times where I have pinned only one or two location and then it works aslong as I don't touch the Hip or the Head. But this really needs to be fixed more then anything. The Active Pose Tool is among my most used tools. Animating simply isn't do able in Daz3d without it.

     

    To any Daz people wanting to know how to replicate the issue.

    1. Add a Genesis Character to your scene

    2. Pin the feet, and the Head. Enable Pin At Both in Tool Settings.

    3. Move anything with Active Pose Tool.

    4. Daz Studio Crashes.

    Yep, still seeing this too. I cannot tell if it has anything to do with which pin setting is active, and previously the crash occured while posing figures, but the latest was posing a length of rope. And seems to happen randomly, as in, everything seems to be working fine, much posing is done, then during posing daz just disappears, no crash report, just 'hello desktop'.

  • olesehenolesehen Posts: 15
    edited October 2018

    I have encountred an issue with the beta , not sure if this has something to do with my setup or something else .  I currently run the newest drivers ( tried older drivers also)  416.34  and when I run a render in the beta with different cards it will not release the VRAM anymore . I have tried both my old GTX 1080 and my new 2080 Ti .  one render then it will only render on CPU until I have exit the application .   this is highly annoying ..  Same thing on both cards .  when I try the same in 4.10 its fine .. obvisouley I Can not try the 2080 as its not supported in 4.10 but the 1080 works fine . 

    I am prettys sure I tried the beta earlier without these issues with the 1080 so perhaps something setting wise has changed but I am unsure what ..  all the usal stuff like optix on/off and denoiser on / off  has been tested without any result .  if anyone has an idé I could try that would work would make me a happy camper .. also the 2080 is working fine in any other application like benchmarks and games so it does not seem to be harware related .

     

     

    *EDIT*

    log file indicates something fishy with de deallocation of the memory 

    2018-10-26 08:42:52.335 Iray INFO - module:category(IRAY:RENDER):   1.7   IRAY   rend info : Allocating 1-layer frame buffer
    2018-10-26 08:42:52.523 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00001 iterations after 7.542s.
    2018-10-26 08:42:53.243 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00002 iterations after 8.266s.
    2018-10-26 08:42:53.992 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00003 iterations after 9.021s.
    2018-10-26 08:42:54.742 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00004 iterations after 9.766s.
    2018-10-26 08:42:55.492 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00005 iterations after 10.514s.
    2018-10-26 08:42:56.258 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00006 iterations after 11.285s.
    2018-10-26 08:42:57.070 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00007 iterations after 12.098s.
    2018-10-26 08:42:57.945 Iray INFO - module:category(POST:RENDER):   1.0   POST   rend info : using cuda device 0, "GeForce RTX 2080 Ti" (7.5), buffers: fp16, cuDNN v7101, rt v9000
    2018-10-26 08:43:04.685 Iray INFO - module:category(POST:RENDER):   1.0   POST   rend info : layers created for resolution 2058 1254, inp 6, outp 3, cuDNN memory 16.0 MB, total 900.1 MB
    2018-10-26 08:43:04.716 WARNING: ..\..\..\..\..\src\pluginsource\DzIrayRender\dzneuraymgr.cpp(304): Iray ERROR - module:category(IRAY:RENDER):   1.15  IRAY   rend error: CUDA device 0 (GeForce RTX 2080 Ti): an illegal memory access was encountered (while launching CUDA renderer in core_renderer_wf.cpp:796)
    2018-10-26 08:43:04.732 WARNING: ..\..\..\..\..\src\pluginsource\DzIrayRender\dzneuraymgr.cpp(304): Iray ERROR - module:category(POST:RENDER):   1.0   POST   rend error: CUDNN_STATUS_EXECUTION_FAILED
    2018-10-26 08:43:04.732 WARNING: ..\..\..\..\..\src\pluginsource\DzIrayRender\dzneuraymgr.cpp(304): Iray ERROR - module:category(IRAY:RENDER):   1.15  IRAY   rend error: CUDA device 0 (GeForce RTX 2080 Ti): Failed to launch renderer
    2018-10-26 08:43:04.732 WARNING: ..\..\..\..\..\src\pluginsource\DzIrayRender\dzneuraymgr.cpp(304): Iray ERROR - module:category(POST:RENDER):   1.0   POST   rend error: CUDNN_STATUS_EXECUTION_FAILED
    2018-10-26 08:43:04.763 WARNING: ..\..\..\..\..\src\pluginsource\DzIrayRender\dzneuraymgr.cpp(304): Iray ERROR - module:category(IRAY:RENDER):   1.7   IRAY   rend error: CUDA device 0 (GeForce RTX 2080 Ti): Device failed while rendering
    2018-10-26 08:43:04.763 WARNING: ..\..\..\..\..\src\pluginsource\DzIrayRender\dzneuraymgr.cpp(304): Iray ERROR - module:category(POST:RENDER):   1.0   POST   rend error: CUDNN_STATUS_EXECUTION_FAILED
    2018-10-26 08:43:04.763 WARNING: ..\..\..\..\..\src\pluginsource\DzIrayRender\dzneuraymgr.cpp(304): Iray ERROR - module:category(IRAY:RENDER):   1.7   IRAY   rend error: CUDA device 0 (GeForce RTX 2080 Ti): an illegal memory access was encountered (while initializing memory buffer)
    2018-10-26 08:43:04.826 WARNING: ..\..\..\..\..\src\pluginsource\DzIrayRender\dzneuraymgr.cpp(304): Iray ERROR - module:category(POST:RENDER):   1.0   POST   rend error: CUDNN_STATUS_EXECUTION_FAILED
    2018-10-26 08:43:04.826 WARNING: ..\..\..\..\..\src\pluginsource\DzIrayRender\dzneuraymgr.cpp(304): Iray ERROR - module:category(IRAY:RENDER):   1.7   IRAY   rend error: CUDA device 0 (GeForce RTX 2080 Ti): an illegal memory access was encountered (while de-allocating memory)

     

     

    **EDIT 2 **

    Seems the main issue is the denoiser as its set to kick in at iteration 8 .  I turned it off again and ran a series of tests .. and even if the VRAM is set to around 10 GB it does not jump to CPU .. then again a bit fishy that the same scene uses around 6 GB of memory for the same scen in 4.10 with the GTX 1080 .. :| 

     

     

    Seems I found the root cause after many MANY hours of testing , apperently the Beta handles textures differently or is not tuned enough  .. so the page size got to large .. I solved it by dedicating 64GB to the pagefile , a set size so studio always has a big chunk of memory to access .. not had any issues since then . will try and get a extremly fast USB 3 and dedicate that to a "ready boost"  setup to simulate extra write cache  . 

    Post edited by olesehen on
  • I have encountred an issue with the beta , not sure if this has something to do with my setup or something else .  I currently run the newest drivers ( tried older drivers also)  416.34  and when I run a render in the beta with different cards it will not release the VRAM anymore . I have tried both my old GTX 1080 and my new 2080 Ti .  one render then it will only render on CPU until I have exit the application .   this is highly annoying ..  Same thing on both cards .  when I try the same in 4.10 its fine .. obvisouley I Can not try the 2080 as its not supported in 4.10 but the 1080 works fine . 

    I am prettys sure I tried the beta earlier without these issues with the 1080 so perhaps something setting wise has changed but I am unsure what ..  all the usal stuff like optix on/off and denoiser on / off  has been tested without any result .  if anyone has an idé I could try that would work would make me a happy camper .. also the 2080 is working fine in any other application like benchmarks and games so it does not seem to be harware related .

     

     

    *EDIT*

    log file indicates something fishy with de deallocation of the memory 

    2018-10-26 08:42:52.335 Iray INFO - module:category(IRAY:RENDER):   1.7   IRAY   rend info : Allocating 1-layer frame buffer
    2018-10-26 08:42:52.523 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00001 iterations after 7.542s.
    2018-10-26 08:42:53.243 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00002 iterations after 8.266s.
    2018-10-26 08:42:53.992 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00003 iterations after 9.021s.
    2018-10-26 08:42:54.742 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00004 iterations after 9.766s.
    2018-10-26 08:42:55.492 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00005 iterations after 10.514s.
    2018-10-26 08:42:56.258 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00006 iterations after 11.285s.
    2018-10-26 08:42:57.070 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00007 iterations after 12.098s.
    2018-10-26 08:42:57.945 Iray INFO - module:category(POST:RENDER):   1.0   POST   rend info : using cuda device 0, "GeForce RTX 2080 Ti" (7.5), buffers: fp16, cuDNN v7101, rt v9000
    2018-10-26 08:43:04.685 Iray INFO - module:category(POST:RENDER):   1.0   POST   rend info : layers created for resolution 2058 1254, inp 6, outp 3, cuDNN memory 16.0 MB, total 900.1 MB
    2018-10-26 08:43:04.716 WARNING: ..\..\..\..\..\src\pluginsource\DzIrayRender\dzneuraymgr.cpp(304): Iray ERROR - module:category(IRAY:RENDER):   1.15  IRAY   rend error: CUDA device 0 (GeForce RTX 2080 Ti): an illegal memory access was encountered (while launching CUDA renderer in core_renderer_wf.cpp:796)
    2018-10-26 08:43:04.732 WARNING: ..\..\..\..\..\src\pluginsource\DzIrayRender\dzneuraymgr.cpp(304): Iray ERROR - module:category(POST:RENDER):   1.0   POST   rend error: CUDNN_STATUS_EXECUTION_FAILED
    2018-10-26 08:43:04.732 WARNING: ..\..\..\..\..\src\pluginsource\DzIrayRender\dzneuraymgr.cpp(304): Iray ERROR - module:category(IRAY:RENDER):   1.15  IRAY   rend error: CUDA device 0 (GeForce RTX 2080 Ti): Failed to launch renderer
    2018-10-26 08:43:04.732 WARNING: ..\..\..\..\..\src\pluginsource\DzIrayRender\dzneuraymgr.cpp(304): Iray ERROR - module:category(POST:RENDER):   1.0   POST   rend error: CUDNN_STATUS_EXECUTION_FAILED
    2018-10-26 08:43:04.763 WARNING: ..\..\..\..\..\src\pluginsource\DzIrayRender\dzneuraymgr.cpp(304): Iray ERROR - module:category(IRAY:RENDER):   1.7   IRAY   rend error: CUDA device 0 (GeForce RTX 2080 Ti): Device failed while rendering
    2018-10-26 08:43:04.763 WARNING: ..\..\..\..\..\src\pluginsource\DzIrayRender\dzneuraymgr.cpp(304): Iray ERROR - module:category(POST:RENDER):   1.0   POST   rend error: CUDNN_STATUS_EXECUTION_FAILED
    2018-10-26 08:43:04.763 WARNING: ..\..\..\..\..\src\pluginsource\DzIrayRender\dzneuraymgr.cpp(304): Iray ERROR - module:category(IRAY:RENDER):   1.7   IRAY   rend error: CUDA device 0 (GeForce RTX 2080 Ti): an illegal memory access was encountered (while initializing memory buffer)
    2018-10-26 08:43:04.826 WARNING: ..\..\..\..\..\src\pluginsource\DzIrayRender\dzneuraymgr.cpp(304): Iray ERROR - module:category(POST:RENDER):   1.0   POST   rend error: CUDNN_STATUS_EXECUTION_FAILED
    2018-10-26 08:43:04.826 WARNING: ..\..\..\..\..\src\pluginsource\DzIrayRender\dzneuraymgr.cpp(304): Iray ERROR - module:category(IRAY:RENDER):   1.7   IRAY   rend error: CUDA device 0 (GeForce RTX 2080 Ti): an illegal memory access was encountered (while de-allocating memory)

     

     

    **EDIT 2 **

    Seems the main issue is the denoiser as its set to kick in at iteration 8 .  I turned it off again and ran a series of tests .. and even if the VRAM is set to around 10 GB it does not jump to CPU .. then again a bit fishy that the same scene uses around 6 GB of memory for the same scen in 4.10 with the GTX 1080 .. :| 

     

     

    Seems I found the root cause after many MANY hours of testing , apperently the Beta handles textures differently or is not tuned enough  .. so the page size got to large .. I solved it by dedicating 64GB to the pagefile , a set size so studio always has a big chunk of memory to access .. not had any issues since then . will try and get a extremly fast USB 3 and dedicate that to a "ready boost"  setup to simulate extra write cache  . 

    Very good to know. Thanks for sharing that!

  • AalaAala Posts: 140
    Aala said:
    Aala said:
    Aala said:

    Wow...I am very impressed with 4.11.0.236....

    Maybe it is just my own ignorance in past but the camera seems to behave more like expected with DoF...

    Also, nominal luminance seems to calculate correctly, allowing me to get much faster renders without the noise in a linear workflow...

    I still notice some limitations when using canvases with denoiser but I know this is still so new...with 30min or less renders on very complex scenes though, without denoiser, I am so happy:)

    Bravo team! From my point of view and daily usage bravo!!

     

    P.S....Meant to add that I don't know what has changed for hair but I truly love the results I am getting now!!

     

    Can you give us an example of what exactly has changed for you to make the renders faster? Personally, I'm getting slower results by quite a margin. I just tested out the scene quoted on this post:

    https://www.daz3d.com/forums/discussion/comment/4009596/#Comment_4009596

    With a 1080 Ti,

    On 4.10 I'm getting 5 mins 40 seconds 

    On 4.11 I'm getting 7 mins 40 seconds.

    That's almost 50% slower. Still haven't tested it out with more complex scenes, but it doesn't sound particularly good to me.

    I am not able to compare before or after regarding that test scene you are using but the difference for me is with the correctly calculated Nominal luminance. I put the nominal luminance value to be the same as what my exposure value is so if my exposure is at 13 I set my nominal luminance to match. Nominal luminance helps with the noise by telling iRay where to look with that value. Also, and for some reason it shows up sometimes and sometimes it is not an option, I combine that with using the noise filter set to 1, 1 .05, forget the individual setting names for those values (can check when home) but they appear when you turn on the noise filter which, again, does not always appear as an option in the Beta for some reason. I think they were taking out that option but sometimes I am able to get to it, I suspect it is when I open a scene saved from a previous version but not sure.

    There are other options that I use which may play a part in how fast my renders are but I cannot recall all the values but will try to upload some screenshots later for my typical setup. I work with linear values in the tone map pane for highlights and crush black set to 0 to that no clipping occurs with the lighting which may affect the render times. I have not looked at the test scene at all for what it is setup to use.

    I also render using a quadro p5000 card in a dell workstation laptop...this could be a big factor with the way the drivers work on that card vs others...so not sure with nothing else to run comparisons with but for me the speed is increased and I am getting much better results in most every area (In this version of the beta 4.11.0.296 vs current stable release 4.10 and even the prior beta I had installed) with exception that denoising seems to be a little more wonky now and totally broken when trying to render to canvases.

    So if I'm understanding this correctly, you put the nominal luminance to 13 if your exposure is set to 13? Wouldn't that make the renders quite dark and the light quite flat? Or am I missing something here?

    Would be interesting and appreciated if you tested the scene I quoted with both DAZ versions (4.10 and the 4.11 latest beta) to see how it stacks up with your card. It's really hard to find out what the best and most optimal settings for Iray are. 

    Correct. It will make the lighting flatter working with linear space but allows you to get a larger range to work with in post without blowing out or losing detail and ultimately is how I prefer to work incorporating post work as I need after rendering. I will post something this evening (I will make sure to get to this tonight) with my settings and a result from the render and you can try yourself with a scene to see what you think about the result. Even if you do not use canvases the scene will have alot more detail rendered out if saved as tiff at least.

    Like I mentioned concerning the test scene, it may not prove to be any better than times you reported with those settings. But the scenes I setup and settings I use are not likely associated at all with that scene. My results are from my particular workflow and in my case with my current setup and drivers and the current BETA, my renders are completed quicker due to noise being filtered out more efficiently, I suspect due to the nominal luminance calculating correctly, or at least better than in previous versions. 

    One note is that I setup each of my scenes in this current beta fresh. So I am not rendering scenes setup from older versions typically which may make a difference as well. I am only comparing my results from the previous BETA I used to this one. I have not even touched 4.10 in months.

    Wow, I got a shock after testing the test scene...I didn't try doing the scene between this beta and 4.10, instead I just changed the nominal luminance and as I expected it cut my render time down from 12:18 seconds from the scene rendered with default from the .duf to 2:38 with nominal set to 13...however, what I didn't realize is just what it affected in the scene. The images below will show the drastic difference. I could not tell this in my renders until seeing it in this test scene but once I saw this I went back to some of my previous renders in this beta and loaded the scenes and then realized how dark, as you asked about, it was making the scene:

     

    DEFAULT:

    TestSceneDefault

    TestSceneDefaultRendered

     

    NOMINAL SET TO 13

    TestSceneNominal

    TestSceneNominalRendered

     

    You can see that the nominal version seemed to remove all the brightness. I am going to test some things on my prior scenes and see what the time difference is but I suspect that the reason I was getting such fast renders was that the nominal is indeed taking away from my scenes, though I didn't notice because the result was ok to my eyes. I will be back within next couple of days with anything further I find.

    That's pretty interesting though! Thanks for the tests! I do use nominal luminance too in my renders, though my values are in the range of around 300-500. I didn't expect the drastic difference in render time though. 12 minutes to 2 minutes is quite a lot.

    Ya I tested with some higher values as well as negative values and at 160 I was able to achieve the same look in that scene as the default but render time was same as default at 12ish min. Thought I had a magic setting but nope, not if I want all the light details.

    Yeah. The new cards, 2080 ti specifically, is quite faster than the 1080 Ti, but you can only use it on the beta, on which it is just a little bit faster for twice the price. Hard choices to make in the near future.

  • nicsttnicstt Posts: 11,715
    Aala said:
    Aala said:
    Aala said:
    Aala said:

    Wow...I am very impressed with 4.11.0.236....

    Maybe it is just my own ignorance in past but the camera seems to behave more like expected with DoF...

    Also, nominal luminance seems to calculate correctly, allowing me to get much faster renders without the noise in a linear workflow...

    I still notice some limitations when using canvases with denoiser but I know this is still so new...with 30min or less renders on very complex scenes though, without denoiser, I am so happy:)

    Bravo team! From my point of view and daily usage bravo!!

     

    P.S....Meant to add that I don't know what has changed for hair but I truly love the results I am getting now!!

     

    Can you give us an example of what exactly has changed for you to make the renders faster? Personally, I'm getting slower results by quite a margin. I just tested out the scene quoted on this post:

    https://www.daz3d.com/forums/discussion/comment/4009596/#Comment_4009596

    With a 1080 Ti,

    On 4.10 I'm getting 5 mins 40 seconds 

    On 4.11 I'm getting 7 mins 40 seconds.

    That's almost 50% slower. Still haven't tested it out with more complex scenes, but it doesn't sound particularly good to me.

    I am not able to compare before or after regarding that test scene you are using but the difference for me is with the correctly calculated Nominal luminance. I put the nominal luminance value to be the same as what my exposure value is so if my exposure is at 13 I set my nominal luminance to match. Nominal luminance helps with the noise by telling iRay where to look with that value. Also, and for some reason it shows up sometimes and sometimes it is not an option, I combine that with using the noise filter set to 1, 1 .05, forget the individual setting names for those values (can check when home) but they appear when you turn on the noise filter which, again, does not always appear as an option in the Beta for some reason. I think they were taking out that option but sometimes I am able to get to it, I suspect it is when I open a scene saved from a previous version but not sure.

    There are other options that I use which may play a part in how fast my renders are but I cannot recall all the values but will try to upload some screenshots later for my typical setup. I work with linear values in the tone map pane for highlights and crush black set to 0 to that no clipping occurs with the lighting which may affect the render times. I have not looked at the test scene at all for what it is setup to use.

    I also render using a quadro p5000 card in a dell workstation laptop...this could be a big factor with the way the drivers work on that card vs others...so not sure with nothing else to run comparisons with but for me the speed is increased and I am getting much better results in most every area (In this version of the beta 4.11.0.296 vs current stable release 4.10 and even the prior beta I had installed) with exception that denoising seems to be a little more wonky now and totally broken when trying to render to canvases.

    So if I'm understanding this correctly, you put the nominal luminance to 13 if your exposure is set to 13? Wouldn't that make the renders quite dark and the light quite flat? Or am I missing something here?

    Would be interesting and appreciated if you tested the scene I quoted with both DAZ versions (4.10 and the 4.11 latest beta) to see how it stacks up with your card. It's really hard to find out what the best and most optimal settings for Iray are. 

    Correct. It will make the lighting flatter working with linear space but allows you to get a larger range to work with in post without blowing out or losing detail and ultimately is how I prefer to work incorporating post work as I need after rendering. I will post something this evening (I will make sure to get to this tonight) with my settings and a result from the render and you can try yourself with a scene to see what you think about the result. Even if you do not use canvases the scene will have alot more detail rendered out if saved as tiff at least.

    Like I mentioned concerning the test scene, it may not prove to be any better than times you reported with those settings. But the scenes I setup and settings I use are not likely associated at all with that scene. My results are from my particular workflow and in my case with my current setup and drivers and the current BETA, my renders are completed quicker due to noise being filtered out more efficiently, I suspect due to the nominal luminance calculating correctly, or at least better than in previous versions. 

    One note is that I setup each of my scenes in this current beta fresh. So I am not rendering scenes setup from older versions typically which may make a difference as well. I am only comparing my results from the previous BETA I used to this one. I have not even touched 4.10 in months.

    Wow, I got a shock after testing the test scene...I didn't try doing the scene between this beta and 4.10, instead I just changed the nominal luminance and as I expected it cut my render time down from 12:18 seconds from the scene rendered with default from the .duf to 2:38 with nominal set to 13...however, what I didn't realize is just what it affected in the scene. The images below will show the drastic difference. I could not tell this in my renders until seeing it in this test scene but once I saw this I went back to some of my previous renders in this beta and loaded the scenes and then realized how dark, as you asked about, it was making the scene:

     

    DEFAULT:

    TestSceneDefault

    TestSceneDefaultRendered

     

    NOMINAL SET TO 13

    TestSceneNominal

    TestSceneNominalRendered

     

    You can see that the nominal version seemed to remove all the brightness. I am going to test some things on my prior scenes and see what the time difference is but I suspect that the reason I was getting such fast renders was that the nominal is indeed taking away from my scenes, though I didn't notice because the result was ok to my eyes. I will be back within next couple of days with anything further I find.

    That's pretty interesting though! Thanks for the tests! I do use nominal luminance too in my renders, though my values are in the range of around 300-500. I didn't expect the drastic difference in render time though. 12 minutes to 2 minutes is quite a lot.

    Ya I tested with some higher values as well as negative values and at 160 I was able to achieve the same look in that scene as the default but render time was same as default at 12ish min. Thought I had a magic setting but nope, not if I want all the light details.

    Yeah. The new cards, 2080 ti specifically, is quite faster than the 1080 Ti, but you can only use it on the beta, on which it is just a little bit faster for twice the price. Hard choices to make in the near future.

    It is a simple choice; I can wait. I may even splash out on the Threadripper 2990, as I'd just need to drop it in my system.

    I'm waiting to see what Octane do, and Blender - then may be looking at an alternative to Nvidia.

  • DAZ Studio 4.11.0.236 is wonderful for Iray rendering.

    I installed it yesterday and immediately noticed *two* great features:

    1. Noise: Post Denoiser (I have a GEFORCE GTX 960) is incredibly efficient... Images look a little bit plasticky but the speed is stunning and results are much better than using Photoshop/Lightroom filters!

    2. Speed: Daz Studio now seems to use ALL the CPU cores (not just one as before) to create the Iray version of the scene and send it to the GPU. Much much faster than before!

    And I'm sure I'll find other great features...

    Thany you, guys: great work!

     

     

     

  • 3dOutlaw3dOutlaw Posts: 2,480

    I saw your post, and since I have a GTX960 as well, decided to download and give it a shot.  Wow...ur right, the denoiser really cleans it up quick, and then you gain detail the longer you let it run.  Love it.

  • vaaernvaaern Posts: 80

    This might've been asked before, but can you install both 4.10 and the 4.11 beta? Will I lose anything in the process? Can I just keep using 4.10 if I get errors in the beta?

  • vaaern said:

    This might've been asked before, but can you install both 4.10 and the 4.11 beta? Will I lose anything in the process? Can I just keep using 4.10 if I get errors in the beta?

    https://www.daz3d.com/forums/discussion/comment/990684/#Comment_990684

  • nicsttnicstt Posts: 11,715
    3dOutlaw said:

    I saw your post, and since I have a GTX960 as well, decided to download and give it a shot.  Wow...ur right, the denoiser really cleans it up quick, and then you gain detail the longer you let it run.  Love it.

    Wish it worked well on skin; it looks fake due to the 'clean-up'.

  • barbultbarbult Posts: 24,755
    nicstt said:
    3dOutlaw said:

    I saw your post, and since I have a GTX960 as well, decided to download and give it a shot.  Wow...ur right, the denoiser really cleans it up quick, and then you gain detail the longer you let it run.  Love it.

    Wish it worked well on skin; it looks fake due to the 'clean-up'.

    Perhaps you have the Post Denoiser Start Iteration set too low. The default of 8 seems way to low for anything, to me. 

  • barbult said:
    nicstt said:
    3dOutlaw said:

    I saw your post, and since I have a GTX960 as well, decided to download and give it a shot.  Wow...ur right, the denoiser really cleans it up quick, and then you gain detail the longer you let it run.  Love it.

    Wish it worked well on skin; it looks fake due to the 'clean-up'.

    Perhaps you have the Post Denoiser Start Iteration set too low. The default of 8 seems way to low for anything, to me. 

    Just watched a good video from the iray dev blog that explains what they are working on with regards to convergence and interfacing with the denoiser:

    http://on-demand.gputechconf.com/gtc/2018/video/S8788/

     

  • 3dOutlaw3dOutlaw Posts: 2,480
    barbult said:
    nicstt said:
    3dOutlaw said:

    I saw your post, and since I have a GTX960 as well, decided to download and give it a shot.  Wow...ur right, the denoiser really cleans it up quick, and then you gain detail the longer you let it run.  Love it.

    Wish it worked well on skin; it looks fake due to the 'clean-up'.

    Perhaps you have the Post Denoiser Start Iteration set too low. The default of 8 seems way to low for anything, to me. 

    Agreed, I set mine to 40 for big scenes and like 200 or so for portrait style.

    For a scifi scene recently, I let it run so that the scenery all looked good with the denoiser (but the models face was still very plastic-like), then quit and saved that.  Then I just spot-rendered the models face to a new window, and disabled the denoiser.  The face rendered in good details very fast by itself.  Then layered the two in post.  Worked great!  Details on the skin, and quick render for scenery.

  • nicsttnicstt Posts: 11,715
    edited November 2018
    barbult said:
    nicstt said:
    3dOutlaw said:

    I saw your post, and since I have a GTX960 as well, decided to download and give it a shot.  Wow...ur right, the denoiser really cleans it up quick, and then you gain detail the longer you let it run.  Love it.

    Wish it worked well on skin; it looks fake due to the 'clean-up'.

    Perhaps you have the Post Denoiser Start Iteration set too low. The default of 8 seems way to low for anything, to me. 

    I've tried it at various settings up to 500; it gets to the stage where it might as well not be used.

    Post edited by nicstt on
  • Dolce SaitoDolce Saito Posts: 195
    edited November 2018

    Hello again;

    Still with the latest beta (and latest drivers), Denoiser generates broken image on Titan V. And this is all the log I have:

    Edit: I also attached the screenshot of the window I am looking at (if I disable denoiser, picture renders fine)

    2018-11-06 22:56:57.254 Rendering image
    2018-11-06 22:56:57.276 Iray INFO - module:category(IMAGE:IO):   1.0   IMAGE  io   info : Loading image "D:\DAZ\Content\Runtime\Textures\Cista_San\cs_floor.jpg", pixel type "Rgb", 3584x3584x1 pixels, 1 miplevel.
    2018-11-06 22:56:57.278 Iray INFO - module:category(IMAGE:IO):   1.0   IMAGE  io   info : Loading image "D:\DAZ\Content\Runtime\Textures\Cista_San\cs_floor_B.jpg", pixel type "Rgb", 3584x3584x1 pixels, 1 miplevel.
    2018-11-06 22:56:57.282 Iray INFO - module:category(IMAGE:IO):   1.0   IMAGE  io   info : Loading image "D:\DAZ\Content\Runtime\Textures\Cista_San\cs_wall_fl.jpg", pixel type "Rgb", 3584x3584x1 pixels, 1 miplevel.
    2018-11-06 22:56:57.286 Iray INFO - module:category(IMAGE:IO):   1.0   IMAGE  io   info : Loading image "D:\DAZ\Content\Runtime\Textures\Cista_San\cs_wall_fl_B.jpg", pixel type "Rgb", 3584x3584x1 pixels, 1 miplevel.
    2018-11-06 22:56:57.294 Iray INFO - module:category(IMAGE:IO):   1.0   IMAGE  io   info : Loading image "D:\DAZ\Content\Runtime\Textures\Cista_San\cs_ceiling_vault.jpg", pixel type "Rgb", 2048x2048x1 pixels, 1 miplevel.
    2018-11-06 22:56:57.296 Iray INFO - module:category(IMAGE:IO):   1.0   IMAGE  io   info : Loading image "D:\DAZ\Content\Runtime\Textures\Cista_San\cs_ceiling_vault_B.jpg", pixel type "Rgb", 2048x2048x1 pixels, 1 miplevel.
    2018-11-06 22:56:57.300 Iray INFO - module:category(IMAGE:IO):   1.0   IMAGE  io   info : Loading image "D:\DAZ\Content\Runtime\Textures\Cista_San\cs_ceiling_arch.jpg", pixel type "Rgb", 3072x3072x1 pixels, 1 miplevel.
    2018-11-06 22:56:57.302 Iray INFO - module:category(IMAGE:IO):   1.0   IMAGE  io   info : Loading image "D:\DAZ\Content\Runtime\Textures\Cista_San\cs_ceiling_arch_B.jpg", pixel type "Rgb", 3072x3072x1 pixels, 1 miplevel.
    2018-11-06 22:56:57.326 Iray INFO - module:category(IMAGE:IO):   1.0   IMAGE  io   info : Loading image "D:\DAZ\Content\Runtime\Textures\Cista_San\cs_cols.jpg", pixel type "Rgb", 3072x3072x1 pixels, 1 miplevel.
    2018-11-06 22:56:57.328 Iray INFO - module:category(IMAGE:IO):   1.0   IMAGE  io   info : Loading image "D:\DAZ\Content\Runtime\Textures\Cista_San\cs_cols_B.jpg", pixel type "Rgb", 3072x3072x1 pixels, 1 miplevel.
    2018-11-06 22:56:57.346 Iray INFO - module:category(IMAGE:IO):   1.0   IMAGE  io   info : Loading image "D:\DAZ\Content\Runtime\Textures\Cista_San\cs_gate.jpg", pixel type "Rgb", 3072x3072x1 pixels, 1 miplevel.
    2018-11-06 22:56:57.348 Iray INFO - module:category(IMAGE:IO):   1.0   IMAGE  io   info : Loading image "D:\DAZ\Content\Runtime\Textures\Cista_San\cs_gate_B.jpg", pixel type "Rgb", 3072x3072x1 pixels, 1 miplevel.
    2018-11-06 22:56:57.354 Iray INFO - module:category(IMAGE:IO):   1.0   IMAGE  io   info : Loading image "D:\DAZ\Content\Runtime\Textures\Cista_San\cs_steps.jpg", pixel type "Rgb", 3072x3072x1 pixels, 1 miplevel.
    2018-11-06 22:56:57.356 Iray INFO - module:category(IMAGE:IO):   1.0   IMAGE  io   info : Loading image "D:\DAZ\Content\Runtime\Textures\Cista_San\cs_steps_B.jpg", pixel type "Rgb", 3072x3072x1 pixels, 1 miplevel.
    2018-11-06 22:56:57.358 Iray INFO - module:category(IMAGE:IO):   1.0   IMAGE  io   info : Loading image "D:\DAZ\Content\Runtime\Textures\Cista_San\cs_steps_railing.jpg", pixel type "Rgb", 3072x3072x1 pixels, 1 miplevel.
    2018-11-06 22:56:57.362 Iray INFO - module:category(IMAGE:IO):   1.0   IMAGE  io   info : Loading image "D:\DAZ\Content\Runtime\Textures\Cista_San\cs_steps_railing_B.jpg", pixel type "Rgb", 3072x3072x1 pixels, 1 miplevel.
    2018-11-06 22:56:57.364 Iray INFO - module:category(IMAGE:IO):   1.0   IMAGE  io   info : Loading image "D:\DAZ\Content\Runtime\Textures\Cista_San\cs_wall_arch.jpg", pixel type "Rgb", 3584x3584x1 pixels, 1 miplevel.
    2018-11-06 22:56:57.366 Iray INFO - module:category(IMAGE:IO):   1.0   IMAGE  io   info : Loading image "D:\DAZ\Content\Runtime\Textures\Cista_San\cs_wall_arch_B.jpg", pixel type "Rgb", 3584x3584x1 pixels, 1 miplevel.
    2018-11-06 22:56:57.428 Iray INFO - module:category(IMAGE:IO):   1.0   IMAGE  io   info : Loading image "D:\DAZ\Content\Runtime\Textures\Cista_San\cs_wall_inset.jpg", pixel type "Rgb", 3584x3584x1 pixels, 1 miplevel.
    2018-11-06 22:56:57.430 Iray INFO - module:category(IMAGE:IO):   1.0   IMAGE  io   info : Loading image "D:\DAZ\Content\Runtime\Textures\Cista_San\cs_wall_inset_B.jpg", pixel type "Rgb", 3584x3584x1 pixels, 1 miplevel.
    2018-11-06 22:56:57.436 Iray INFO - module:category(IMAGE:IO):   1.0   IMAGE  io   info : Loading image "D:\DAZ\Content\Runtime\Textures\Cista_San\cs_bust.jpg", pixel type "Rgb", 2560x2560x1 pixels, 1 miplevel.
    2018-11-06 22:56:57.440 Iray INFO - module:category(IMAGE:IO):   1.0   IMAGE  io   info : Loading image "D:\DAZ\Content\Runtime\Textures\Cista_San\cs_bust_B.jpg", pixel type "Rgb", 2560x2560x1 pixels, 1 miplevel.
    2018-11-06 22:56:57.444 Iray INFO - module:category(IMAGE:IO):   1.0   IMAGE  io   info : Loading image "D:\DAZ\Content\Runtime\Textures\Cista_San\cs_pedestal.jpg", pixel type "Rgb", 2560x2560x1 pixels, 1 miplevel.
    2018-11-06 22:56:57.446 Iray INFO - module:category(IMAGE:IO):   1.0   IMAGE  io   info : Loading image "D:\DAZ\Content\Runtime\Textures\Cista_San\cs_pedestal_B.jpg", pixel type "Rgb", 2560x2560x1 pixels, 1 miplevel.
    2018-11-06 22:56:57.456 Iray INFO - module:category(IMAGE:IO):   1.0   IMAGE  io   info : Loading image "D:\DAZ\Content\Runtime\Textures\Cista_San\cs_water.jpg", pixel type "Rgb", 3584x3584x1 pixels, 1 miplevel.
    2018-11-06 22:56:57.648 Iray INFO - module:category(POST:RENDER):   1.0   POST   rend info : Postprocessing denoiser is enabled
    2018-11-06 22:56:57.648 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Updating geometry.
    2018-11-06 22:56:57.688 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Using OptiX Prime ray tracing (5.0.1).
    2018-11-06 22:56:57.690 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Importing scene graph.
    2018-11-06 22:56:57.690 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Importing geometry for motion time 0
    2018-11-06 22:56:57.744 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Geometry import (1 object with 213k triangles, 1 instance yielding 213k triangles) took 0.052618
    2018-11-06 22:56:57.744 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Updating materials.
    2018-11-06 22:56:57.750 Iray INFO - module:category(MATCNV:RENDER):   1.0   MATCNV rend info : found 63 textures, 0 lambdas (0 unique)
    2018-11-06 22:56:57.764 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Emitter geometry import (1 light source with 1 triangle, 1 instance) took 0.00s
    2018-11-06 22:56:57.764 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Updating environment.
    2018-11-06 22:56:57.764 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Updating backplate.
    2018-11-06 22:56:57.766 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Updating lens.
    2018-11-06 22:56:57.766 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Updating lights.
    2018-11-06 22:56:57.766 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Updating object flags.
    2018-11-06 22:56:57.768 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Updating caustic portals.
    2018-11-06 22:56:57.768 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Updating decals.
    2018-11-06 22:56:57.784 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Allocating 1-layer frame buffer
    2018-11-06 22:56:57.796 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Using batch scheduling, caustic sampler disabled
    2018-11-06 22:56:57.796 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Initializing local rendering.
    2018-11-06 22:56:57.802 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Rendering with 1 device(s):
    2018-11-06 22:56:57.802 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info :     CUDA device 0 (TITAN V)
    2018-11-06 22:56:57.804 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Rendering...
    2018-11-06 22:56:57.806 Iray VERBOSE - module:category(IRAY:RENDER):   1.12  IRAY   rend progr: CUDA device 0 (TITAN V): Processing scene...
    2018-11-06 22:56:57.818 Iray VERBOSE - module:category(IRAY:RENDER):   1.5   IRAY   rend stat : Geometry memory consumption: 6.55595 MiB (device 0), 0 B (host)
    2018-11-06 22:57:00.506 Iray VERBOSE - module:category(IRAY:RENDER):   1.5   IRAY   rend stat : Texture memory consumption: 149.609 MiB for 25 bitmaps (device 0)
    2018-11-06 22:57:00.508 Iray INFO - module:category(IRAY:RENDER):   1.5   IRAY   rend info : Importing lights for motion time 0
    2018-11-06 22:57:00.508 Iray INFO - module:category(IRAY:RENDER):   1.5   IRAY   rend info : Initializing light hierarchy.
    2018-11-06 22:57:00.510 Iray INFO - module:category(IRAY:RENDER):   1.5   IRAY   rend info : Light hierarchy initialization took 0.00s
    2018-11-06 22:57:00.512 Iray VERBOSE - module:category(IRAY:RENDER):   1.5   IRAY   rend stat : Lights memory consumption: 772 B (device 0)
    2018-11-06 22:57:00.518 Iray VERBOSE - module:category(IRAY:RENDER):   1.5   IRAY   rend stat : Material measurement memory consumption: 0 B (GPU)
    2018-11-06 22:57:00.520 Iray VERBOSE - module:category(IRAY:RENDER):   1.5   IRAY   rend stat : Materials memory consumption: 124.637 KiB (GPU)
    2018-11-06 22:57:00.590 Iray INFO - module:category(IRAY:RENDER):   1.12  IRAY   rend info : CUDA device 0 (TITAN V): Scene processed in 2.785s
    2018-11-06 22:57:00.594 Iray INFO - module:category(IRAY:RENDER):   1.12  IRAY   rend info : CUDA device 0 (TITAN V): Allocated 9.37531 MiB for frame buffer
    2018-11-06 22:57:00.596 Iray INFO - module:category(IRAY:RENDER):   1.12  IRAY   rend info : CUDA device 0 (TITAN V): Allocated 1.6875 GiB of work space (2048k active samples in 0.000s)
    2018-11-06 22:57:00.596 Iray INFO - module:category(IRAY:RENDER):   1.12  IRAY   rend info : CUDA device 0 (TITAN V): Used for display, optimizing for interactive usage (performance could be sacrificed)
    2018-11-06 22:57:00.630 Iray INFO - module:category(IRAY:RENDER):   1.12  IRAY   rend info : Allocating 1-layer frame buffer
    2018-11-06 22:57:00.650 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00001 iterations after 2.845s.
    2018-11-06 22:57:00.688 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00002 iterations after 2.883s.
    2018-11-06 22:57:00.740 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00004 iterations after 2.935s.
    2018-11-06 22:57:00.820 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00008 iterations after 3.016s.
    2018-11-06 22:57:00.954 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00016 iterations after 3.150s.
    2018-11-06 22:57:01.132 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00027 iterations after 3.327s.
    2018-11-06 22:57:01.308 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00038 iterations after 3.504s.
    2018-11-06 22:57:01.486 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00049 iterations after 3.681s.
    2018-11-06 22:57:01.676 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00061 iterations after 3.872s.
    2018-11-06 22:57:01.740 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 00064 iterations after 3.936s.
    2018-11-06 22:57:01.760 Iray INFO - module:category(POST:RENDER):   1.0   POST   rend info : using cuda device 0, "TITAN V" (7.0), buffers: fp16, cuDNN v7101, rt v9000
    2018-11-06 22:57:01.772 Iray INFO - module:category(POST:RENDER):   1.0   POST   rend info : layers created for resolution 640 480, inp 8, outp 8, cuDNN memory 1.8 MB, total 107.7 MB
    2018-11-06 22:57:01.808 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Maximum number of samples reached.
    2018-11-06 22:57:01.902 Saved image: C:\Users\unfortunateTitanVuser\AppData\Roaming\DAZ 3D\Studio4 Public Build\temp\render\r.png
    2018-11-06 22:57:01.908 Finished Rendering
    2018-11-06 22:57:01.964 Total Rendering Time: 4.81 seconds
    2018-11-06 22:57:01.976 Loaded image r.png
    2018-11-06 22:57:01.994 Saved image: C:\Users\unfortunateTitanVuser\AppData\Roaming\DAZ 3D\Studio4 Public Build\temp\RenderAlbumTmp\Render 32.jpg
    2018-11-06 22:57:05.462 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Device statistics:
    2018-11-06 22:57:05.462 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : CUDA device 0 (TITAN V):      64 iterations, 2.802s init, 1.135s render
    2018-11-06 22:57:08.376 Saved image: D:\Renders\castle.duf.png
    2018-11-06 22:57:09.378 WARNING: ..\..\..\..\..\src\sdksource\cloud\dzcloudtasknotifier.cpp(178): Tcp shutdown called on a Tcp connection without a shutdown
    2018-11-06 22:57:10.618 Iray INFO - module:category(IRT:RENDER):   1.1   IRT    rend info : Shutting down irt render plugin.
    2018-11-06 22:57:10.622 Iray INFO - module:category(IRT:RENDER):   1.0   IRT    rend info : Destroying context.
    2018-11-06 22:57:11.192 *** Scene Cleared ***
    2018-11-06 22:57:11.416 DEBUG: End DAZ Studio to Hexagon Bridge log...
    2018-11-06 22:57:11.666 --------------- DAZ Studio 4.11.0.236 exited ------------------
    2018-11-06 22:57:11.666 ~

    broken denoiser.png
    1919 x 1084 - 2M
    Post edited by Dolce Saito on
  • VEGAVEGA Posts: 86

    I've noticed there OptiX Prime 5.0.1 does it work on Titan V ? Because when I use Optix Prime in Beta on RTX 2080 ti it takes longer than without it. I'm only waiting for this to be fixed.

This discussion has been closed.