Daz Studio Iray - Rendering Hardware Benchmarking

2456

Comments

  • DripDrip Posts: 410
    edited July 2019

    Missing a 2070 from the collection? Here's my bench to give people a crude estimate.

    System Configuration
    System/Motherboard: MSi B450-A Pro
    CPU: AMD Ryzen 5 2600x stock
    GPU: RTX2070 @ stock
    System Memory: Corsair LPX 8GB  (one bank died, currently running on only 1 bank) @ default (2166 mhz I think?)
    OS Drive: Samsung EVO 860 500GB
    Asset Drive: Seagate Barracuda 4TB @ 5400rpm
    Operating System: Win 10 Home
    Nvidia Drivers Version: 430.86
    Daz Studio Version: 4.11.0.383 64-bit
    Optix Prime Acceleration: Off

    Benchmark Results
    DAZ_STATS
    2019-07-26 00:04:38.901 Finished Rendering
    2019-07-26 00:04:38.951 Total Rendering Time: 8 minutes 23.21 seconds
    IRAY_STATS
    2019-07-26 00:04:52.779 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Device statistics:
    2019-07-26 00:04:52.779 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : CUDA device 0 (GeForce RTX 2070):      1800 iterations, 5.544s init, 493.208s render

    Iteration Rate: 1800/493.208=6.139  3.650 iterations per second  *FIXED, thank you*
    Loading Time: 503.21 - 493.21 = 10 seconds

    Post edited by Drip on
  • RayDAntRayDAnt Posts: 584

    Drip just an fyi that your Iteration Rate stat is actually 3.650 (that's a resounding "Yay!" from me for the wonders of built-in data collection redundancy.) Any chance you'd be up for benching with OptiX and/or in the current 4.12 Beta as well? You finally get the full benefits of raytracing there.

  • DripDrip Posts: 410
    edited July 2019
    RayDAnt said:

    Drip just an fyi that your Iteration Rate stat is actually 3.650 (that's a resounding "Yay!" from me for the wonders of built-in data collection redundancy.) Any chance you'd be up for benching with OptiX and/or in the current 4.12 Beta as well? You finally get the full benefits of raytracing there.

    Tick box, hit render.

    System Configuration
    System/Motherboard: MSi B450-A Pro
    CPU: AMD Ryzen 5 2600x stock
    GPU: RTX2070 @ stock
    System Memory: Corsair LPX 8GB  (one bank died, currently running on only 1 bank) @ default (2166 mhz I think?)
    OS Drive: Samsung EVO 860 500GB
    Asset Drive: Seagate Barracuda 4TB @ 5400rpm
    Operating System: Win 10 Home
    Nvidia Drivers Version: 430.86
    Daz Studio Version: 4.11.0.383 64-bit
    Optix Prime Acceleration: On

    Benchmark Results
    DAZ_STATS
    2019-07-26 08:52:43.305 Finished Rendering
    2019-07-26 08:52:44.210 Total Rendering Time: 8 minutes 37.38 seconds
    IRAY_STATS
    2019-07-26 08:52:54.770 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Device statistics:
    2019-07-26 08:52:54.770 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : CUDA device 0 (GeForce RTX 2070):      1800 iterations, 4.235s init, 507.814s render

    Iteration Rate: 1800/507.814=3.545 iterations per second
    Loading Time: 517.38 - 507.81 = 9.57 seconds

    I don't have the beta installed, so no results from that I'm afraid.
    Not entirely sure, but I *think* the Optix Prime version was actually a little bit faster to converge the first ~20%, though it still ended up slightly slower overall. So the initial iterations might be better in quality, the time per iteration is slightly longer.

    Post edited by Drip on
  • Jack TomalinJack Tomalin Posts: 9,495
    edited July 2019

    I don't want to muddy the waters, so consider this anecdotal.. but wanted to just test the new render server - so threw the benchmark on it and..

    Rendered snapshot in 69.6524s

    ETA that's 4x 2080ti's in there now, sent via DS 4.12 on Iray Server 2.53 (so Iray 317500.2436)

    Post edited by Jack Tomalin on
  • RayDAntRayDAnt Posts: 584
    edited July 2019

    Jack Tomalin said:

    I don't want to muddy the waters, so consider this anecdotal.. but wanted to just test the new render server - so threw the benchmark on it and..

    Rendered snapshot in 69.6524s

    ETA that's 4x 2080ti's in there now, sent via DS 4.12 on Iray Server 2.53 (so Iray 317500.2436)

    By all means, muddy the waters (post as much benchmarking data/hardware details as you can stand.) 'Tis the reason I created this benchmark/thread in the first place. I can always just not include whatever other data you have if it proves too convoluted to get into the graphs/tables above.

    Btw do you run Iray Server on Windows or Linux? Because in case you don't already know, running it under Linux will give you some MAJOR surprise perks.

    Post edited by RayDAnt on
  • Jack TomalinJack Tomalin Posts: 9,495
    RayDAnt said:

    I don't want to muddy the waters

    By all means, muddy the waters (post as much benchmarking data/hardware details as you can stand.) 'Tis the reason I created this benchmark/thread in the first place. I can always just not include whatever other data you have if it proves too convoluted to get into the graphs/tables above.

    Btw do you run Iray Server on Windows or Linux? Because in case you don't already know, running it under Linux will give you some MAJOR surprise perks.

    Currently Windows, since clustering only works between the same OS's.. and I wanted to try and get that working with my work rig.  Initial attempts saw some bugs, but I didn't know if it was due to the fact I was mixing the cards.  Now everything is 2080ti's between the two machines, I'll try again.

    But yea, it would be nice to try a Linux distro at some point.

  • RayDAntRayDAnt Posts: 584
    edited July 2019
    RayDAnt said:

    Jack Tomalin said:

    I don't want to muddy the waters, so consider this anecdotal.. but wanted to just test the new render server - so threw the benchmark on it and..

    Rendered snapshot in 69.6524s

    ETA that's 4x 2080ti's in there now, sent via DS 4.12 on Iray Server 2.53 (so Iray 317500.2436)

    By all means, muddy the waters (post as much benchmarking data/hardware details as you can stand.) 'Tis the reason I created this benchmark/thread in the first place. I can always just not include whatever other data you have if it proves too convoluted to get into the graphs/tables above.

    Btw do you run Iray Server on Windows or Linux? Because in case you don't already know, running it under Linux will give you some MAJOR surprise perks.

    Currently Windows, since clustering only works between the same OS's.. and I wanted to try and get that working with my work rig.  Initial attempts saw some bugs, but I didn't know if it was due to the fact I was mixing the cards.  Now everything is 2080ti's between the two machines, I'll try again.

    But yea, it would be nice to try a Linux distro at some point.

    There's been LOTS of hay made about whether or not Iray is currently capable of implementing VRAM pooling across multiple GPUs. The full truth of the matter (and this is information I was only able to finally tease out properly as recently as two weeks ago after months of exhaustive research) is that it DOES (and has been capable of doing so for years) - but only in the case of Iray Server being run on a Unix-based system with the addition of NVLink bridges with capable Turing cards (like your 2080Ti's or my Titan RTX.)

    The key sticking point for compatibility under other operating systems is apparently in the way graphics cards and their indivial hardware components are enumerated as devices by the operating system. Both modern versions of Windows and Mac OS handle io operations with onboard graphics memory directly, whereas Unix systems farm the task off to the manufacturer's driver (assuming one exists, of course.) This is key because Nvidia's implementation of memory pooling (what they loosely refer to as enhanced Unified Memory) is enacted at the basic io level. Making it currently a no-go on OS X or Windows unless Apple/Microsoft start tailoring there own low level graphics card support systems (eg. WDDM) to fit Nvidia's own proprietary implementation needs...

    Post edited by RayDAnt on
  • System Configuration
    System/Motherboard: ASRock Z170 OC Formula
    CPU: Intel Core i7-6700 @ 3.40GHz (stock)
    GPU: Nvidia GTX 1070 Founders Edition @ stock Nvidia RTX 2080 Founders Edition @ stock
    System Memory: Kingston Hynix 32GB DDR4 @ 2133MHz
    OS Drive: Samsung 850 EVO 500GB
    Asset Drive: Western Digital Mainstream (WD30EZRZ) 3TB
    Operating System: Windows 8.1 6.3.9600 Build 9600
    Nvidia Drivers Version: 430.86 Standard

     

    Benchmark Results - GTX 1070 Only
    Daz Studio Version: 4.11.0.383 64-bit
    Optrix Prime Acceleration: Yes
    CUDA device 1 (GeForce GTX 1070): 1800 iterations, 4.233s init, 917.842s render
    Total Rendering Time: 15 minutes 25.59 seconds
    Iteration Rate: (1800/917.842) = 1.961 iterations per second
    Loading Time: ((0 + 900 + 25.59) 925.59) = (925.59 - 917.842) = 7.748 seconds

    Benchmark Results - RTX 2080 Only
    Daz Studio Version: 4.11.0.383 64-bit
    Optrix Prime Acceleration: Yes
    CUDA device 0 (GeForce RTX 2080): 1800 iterations, 4.161s init, 344.199s render
    Total Rendering Time: 5 minutes 51.33 seconds
    Iteration Rate: (1800/344.199) = 5.229 iterations per second
    Loading Time: ((0 + 300 + 51.33) 351.33) = (351.33 - 344.199) = 7.131 seconds

    Benchmark Results - GTX 1070 + RTX 2080
    Daz Studio Version: 4.11.0.383 64-bit
    Optrix Prime Acceleration: Yes
    CUDA device 1 (GeForce GTX 1070): 475 iterations, 4.603s init, 254.723s render
    CUDA device 0 (GeForce RTX 2080): 1325 iterations, 4.605s init, 255.111s render
    Total Rendering Time: 4 minutes 22.59 seconds
    Iteration Rate: (1800/255.111) = 7.055 iterations per second
    Loading Time: ((0 + 240 + 22.59) 262.59) = (262.59 - 255.111) = 7.479 seconds

    Benchmark Results - GTX 1070 Only
    Daz Studio Version: 4.12.0.33 Beta 64-bit

    Optrix Prime Acceleration: Yes
    CUDA device 1 (GeForce GTX 1070): 1800 iterations, 4.548s init, 762.203s render
    Total Rendering Time: 12 minutes 49.99 seconds
    Iteration Rate: (1800/762.203) = 2.361 iterations per second
    Loading Time: ((0 + 720 + 49.99) 769.99) = (769.99 - 762.203) = 7.787 seconds

    Benchmark Results - RTX 2080 Only
    Daz Studio Version: 4.12.0.33 Beta 64-bit

    Optrix Prime Acceleration: Yes
    CUDA device 0 (GeForce RTX 2080): 1800 iterations, 4.234s init, 313.011s render
    Total Rendering Time: 5 minutes 20.48 seconds
    Iteration Rate: (1800/313.011) = 5.750 iterations per second
    Loading Time: ((0 + 300 + 20.48) 320.48) = (320.48 - 313.011) = 7.469

    Benchmark Results - GTX 1070 + RTX 2080
    Daz Studio Version: 4.12.0.33 Beta 64-bit
    Optrix Prime Acceleration: Yes
    CUDA device 1 (GeForce GTX 1070): 511 iterations, 4.805s init, 226.737s render
    CUDA device 0 (GeForce RTX 2080): 1289 iterations, 4.326s init, 227.118s rende
    Total Rendering Time: 3 minutes 54.87 seconds
    Iteration Rate: (1800/227.118) = 7.925 iterations per second
    Loading Time: ((0 + 180 + 54.87) 234.87) = (234.87 - 227.118) = 7.752

     

     

     

     

  • AelfinMaegikAelfinMaegik Posts: 47
    edited July 2019

    2x2080ti
    Wasn't really sure what to do with the maths w/2 cards? Maybe take the highest? (1800/151.359 = 11.892? Seems too high)

    System Configuration
    System/Motherboard
    : Gigabyte ATX DDR3 LGA 1150 SATA DIMM 6Gb/s Motherboard (GA-Z97X-UD3H)
    CPU: Intel Core i7-4790 Processor @ stock
    GPU: 2 x EVGA GeForce RTX 2080 Ti Black Edition Gaming, 11GB GDDR6 11G-P4-2281-KR @ stock (if matters, they are in NVLINK)
    System Memory: Ballistix Sport 32GB Kit (8GBx4) DDR3 1600 MT/s (PC3-12800)
    OS Drive: Samsung 840 EVO 500GB 2.5-Inch SATA III Internal SSD
    Asset Drive: 2 x Seagate Desktop 3 TB HDD SATA 6 Gb/s NCQ 64MB Cache 7200 RPM, RAID 1
    Operating System: Windows 10 Pro Build 17134
    Nvidia Drivers Version: 431.36
    Daz Studio Version: 4.11.0.383
    Optix Prime Acceleration: ON

    Benchmark Results
    DAZ_STATS

    2019-07-27 10:32:17.410 Finished Rendering
    2019-07-27 10:32:17.456 Total Rendering Time: 2 minutes 40.97 seconds
    IRAY_STATS
    2019-07-27 10:34:26.723 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Device statistics:
    2019-07-27 10:34:26.723 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : CUDA device 1 (GeForce RTX 2080 Ti): 917 iterations, 5.694s init, 151.359s render
    2019-07-27 10:34:26.723 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : CUDA device 0 (GeForce RTX 2080 Ti): 883 iterations, 5.735s init, 150.687s render

    Iteration Rate: (1800 / 151.359) = 11.892 iterations per second
    Loading Time: ((0 * 3600 + 2 * 60 + 40.97) - 151.359) = 160.97 - 151.359 = 9.611 seconds

    Thanks @RayDAnt for the final numbers. I did calc that but thought it was too high. Also, seeing that loading time and discussions, I may switch assets over to my 1TB M.2 850 EVO.

    rayDant-411-optixON-2x2080ti.png
    900 x 900 - 1M
    Post edited by AelfinMaegik on
  • RayDAntRayDAnt Posts: 584
    edited July 2019

    @AelfinMaegik see here for the right template to use in your case (adjusted the wording to hopefully make it more obvious.)

    Here's how it works out for your results:

    Iteration Rate: (1800 / 151.359) = 11.892 iterations per second
    Loading Time: ((0 * 3600 + 2 * 60 + 40.97) - 151.359) = 160.97 - 151.359 = 9.611 seconds
    Post edited by RayDAnt on
  • Same as above, just with 4.12 BETA

    Daz Studio Version: 4.12.0.33 BETA

    Benchmark Results
    DAZ_STATS
    2019-07-27 12:06:17.050 Finished Rendering
    2019-07-27 12:06:17.096 Total Rendering Time: 2 minutes 25.72 seconds
    IRAY_STATS
    2019-07-27 12:07:00.784 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Device statistics:
    2019-07-27 12:07:00.784 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : CUDA device 1 (GeForce RTX 2080 Ti): 904 iterations, 5.494s init, 136.003s render
    2019-07-27 12:07:00.784 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : CUDA device 0 (GeForce RTX 2080 Ti): 896 iterations, 5.289s init, 136.656s render

    Iteration Rate: (1800 / 136.656) = 13.172 iterations per second
    Loading Time: ((0 * 3600 + 2 * 60 + 25.72) - 136.656) = 9.064 seconds

  • A Paper TigerA Paper Tiger Posts: 3
    edited August 2019

    System Configuration
    System/Motherboard: Asus P8Z68-V LE
    CPU: Intel Core i7-2600 @ 3.40GHz (Stock)
    GPU: EVGA GTX 1070ti SC @ stock NVIDIA GTX 970 @ stock
    System Memory: Kingston 16GB DDR3
    OS Drive: SanDisk Ultra 3D SSD
    Asset Drive: Western Digital Mainstream (WD30EZRZ) 3TB
    Operating System: Windows 10 Home 1903 18362.239
    Nvidia Drivers Version: 430.86 Standard

     

    Benchmark Results - GTX 1070ti Only
    Daz Studio Version: 4.11.0.383 64-bit
    Optix Prime Acceleratio: Yes
    CUDA device 0 (GeForce GTX 1070 Ti): 1800 iterations, 6.175s init, 770.149s render
    Total Rendering Time: 13 minutes 0.20 seconds
    Iteration Rate: (1800/770.149) = 2.337 iterations per second
    Loading Time: ((0 + 780 + 0.20) 780.20) = (780.20 - 770.149) = 10.51


    Benchmark Results - GTX 1070ti Only
    Daz Studio Version: 4.12.0.33 Beta 64-bit

    Optix Prime Acceleratio: Yes
    Total Rendering Time: 10 minutes 58.40 seconds
    CUDA device 0 (GeForce GTX 1070 Ti): 1800 iterations, 6.094s init, 648.264s render
    Iteration Rate: (1800/648.264) = 2.776 iterations per second
    Loading Time: ((0 + 600 + 58.40) 658.40) = (658.40 - 648.264) = 10.136

    Post edited by A Paper Tiger on
  • Doing these made me realize I needed to update Windows and Nvidia drivers. Not much of a difference, probably within statistical variance.

    Operating System: Windows 10 Pro Build 18362 (1903)
    Nvidia Drivers Version: 431.60 GRD
    Daz Studio Version: 4.12.0.33 BETA

    Benchmark Results
    DAZ_STATS
    2019-07-27 15:53:50.052 Finished Rendering
    2019-07-27 15:53:50.087 Total Rendering Time: 2 minutes 24.16 seconds
    IRAY_STATS
    2019-07-27 15:54:22.007 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Device statistics:
    2019-07-27 15:54:22.007 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : CUDA device 1 (GeForce RTX 2080 Ti): 909 iterations, 4.817s init, 135.982s render
    2019-07-27 15:54:22.007 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : CUDA device 0 (GeForce RTX 2080 Ti): 891 iterations, 4.546s init, 135.692s render

    Iteration Rate: (1800 / 135.982) = 13.2370
    Loading Time: ((0 * 3600 + 2 * 60 + 24.16) - 135.982) = 8.178

  • garyh.pubgaryh.pub Posts: 57
    edited August 2019

    The DS 4.10.0.123 table looks pretty bare.  I'll add the first entry...

    System Configuration
    System/Motherboard: MSI Meg Creation X399
    CPU: AMD Threadripper 1920X @ 4.00 GHz
    GPU: Nvidia Titan Xp @ 2050, Nvidia Titan Xp 2050, Nvidia Titan X Pascal @ 2050
    System Memory: G.Skill Ripjaws 32GB DDR4 @ 3200
    OS Drive: Samsung 970 NVMe 512 GB
    Asset Drive: Samsung 860 EVO 1TB
    Operating System: Windows 10 Home 64-bit 1809
    Nvidia Drivers Version: 416.34 Standard
    Daz Studio Version: 4.10.0.123 64-bit
    Optix Prime Acceleration: On

    Benchmark Results
    DAZ_STATS

    2019-07-27 20:44:42.604 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 01800 iterations after 146.093s.

    2019-07-27 20:44:43.305 Total Rendering Time: 2 minutes 29.56 seconds


    IRAY_STATS

    2019-07-27 20:45:04.494 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : CUDA device 1 (TITAN Xp): 617 iterations, 13.473s init, 132.657s render

    2019-07-27 20:45:04.499 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : CUDA device 0 (TITAN Xp): 620 iterations, 13.420s init, 132.414s render

    2019-07-27 20:45:04.499 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : CUDA device 2 (TITAN X (Pascal)): 563 iterations, 13.427s init, 132.636s render


    Rendering Performance: 1800 / 132.657 = 13.569 iterations per second
    Loading Time: [(0 + 120 + 29.56) - 132.657 = 16.903 seconds

    EDIT:  I corrected my calculations, and made a couple of other corrections/clarifications.

    Post edited by garyh.pub on
  • garyh.pub said:

    The DS 4.10.0.123 table looks pretty bare.  I'll add the first entry...

    System Configuration
    System/Motherboard: MSI Meg Creation X399
    CPU: AMD Threadripper 1920X @ 4.00 GHz
    GPU: Nvidia Titan Xp @ 2050, Nvidia Titan Xp @ 2038, Nvidia Titan X Pascal @ 2038
    System Memory: G.Skill Ripjaws 32GB DDR4 @ 3200
    OS Drive: Samsung 970 NVMe 512 GB
    Asset Drive: Samsung 860 EVO 1TB
    Operating System: Windows 10 Home 64-bit 1809
    Nvidia Drivers Version: 416.34 Standard
    Daz Studio Version: 4.10.0.123
    Optix Prime Acceleration: On

    Benchmark Results
    DAZ_STATS

    2019-07-27 20:44:42.604 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : Received update to 01800 iterations after 146.093s.

    2019-07-27 20:44:43.305 Total Rendering Time: 2 minutes 29.56 seconds


    IRAY_STATS

    2019-07-27 20:45:04.494 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : CUDA device 1 (TITAN Xp): 617 iterations, 13.473s init, 132.657s render

    2019-07-27 20:45:04.499 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : CUDA device 0 (TITAN Xp): 620 iterations, 13.420s init, 132.414s render

    2019-07-27 20:45:04.499 Iray INFO - module:category(IRAY:RENDER):   1.0   IRAY   rend info : CUDA device 2 (TITAN X (Pascal)): 563 iterations, 13.427s init, 132.636s render


    Rendering Performance: 1800 / 146.093 = 12.32 iterations per second
    Loading Time: [(0 + 120 + 29.56) - 146.093 = 3.467 seconds

    I believe this should be 13.56 i/s with a load time of 16.903

  • AalaAala Posts: 112
    edited July 2019

    (Just upgraded from 2x 1080 Ti's to 2x 2080 Ti's, did some benchmarks)

    System Configuration
    System/Motherboard: Gigabyte Z390 Aorus Master
    CPU: Intel i7-9900K @ 4.7 Ghz
    GPU: 2x Gigabyte GTX 1080 Ti Windforce OC
    System Memory: Corsair Vengeance LPX 64GB DDR4 @ 3000Mhz
    OS Drive: Samsung Pro 970 512GB NVME SSD
    Asset Drive: Samsung SSD 860 QVO 4TB
    Operating System: Windows 10 Pro version 1903
    Nvidia Drivers Version: 431.60

    ------------------------------------------------

    Benchmark Results - 2x Gigabyte GTX 1080 Ti Windforce OC
    Daz Studio Version: 4.10.0.123
    Optix Prime Acceleration: ON
    Total Rendering Time: 4 minutes 2.48 seconds
    CUDA device 0 (GeForce GTX 1080 Ti): Scene processed in 9.489s, 239.733s render
    CUDA device 1 (GeForce GTX 1080 Ti): Scene processed in 9.498s, 239.733s render

    Benchmark Results - 2x Gigabyte GTX 1080 Ti Windforce OC
    Daz Studio Version: 4.10.0.123
    Optix Prime Acceleration: OFF
    Total Rendering Time: 4 minutes 27.64 seconds
    CUDA device 0 (GeForce GTX 1080 Ti): Scene processed in 9.397s, 265.072s render
    CUDA device 1 (GeForce GTX 1080 Ti): Scene processed in 9.406s, 265.072s render

    ---

    Benchmark Results - 2x Gigabyte GTX 1080 Ti Windforce OC
    Daz Studio Version: 4.12.0.33 Public Beta
    Optix Prime Acceleration: ON
    Total Rendering Time: 4 minutes 12.83 seconds
    CUDA device 0 (GeForce GTX 1080 Ti): Scene processed in 2.286s, 250.252s render
    CUDA device 1 (GeForce GTX 1080 Ti): Scene processed in 2.290s, 250.252s render

    Benchmark Results - 2x Gigabyte GTX 1080 Ti Windforce OC
    Daz Studio Version: 4.12.0.33 Public Beta
    Optix Prime Acceleration: OFF
    Total Rendering Time: 4 minutes 12.51 seconds
    CUDA device 0 (GeForce GTX 1080 Ti): Scene processed in 1.908s, 250.513s render
    CUDA device 1 (GeForce GTX 1080 Ti): Scene processed in 1.915s, 250.513s render

    ---------------------------------------------------------------------------------------------------------------------------

    System Configuration
    System/Motherboard: Gigabyte Z390 Aorus Master
    CPU: Intel i7-9900K @ 4.7 Ghz
    GPU: 2x Gigabyte RTX 2080 Ti Gaming OC 11G
    System Memory: Corsair Vengeance LPX 64GB DDR4 @ 3000Mhz
    OS Drive: Samsung Pro 970 512GB NVME SSD
    Asset Drive: Samsung SSD 860 QVO 4TB
    Operating System: Windows 10 Pro version 1903
    Nvidia Drivers Version: 431.60

    ---

    Benchmark Results - 2x Gigabyte RTX 2080 Ti Gaming OC 11G
    Daz Studio Version: 4.12.0.33 Public Beta
    Optix Prime Acceleration: ON
    Total Rendering Time: 2 minutes 22.0 seconds
    CUDA device 0 (GeForce RTX 2080 Ti): Scene processed in 7.877s (First time render), 139.184s render
    CUDA device 1 (GeForce RTX 2080 Ti): Scene processed in 7.880s (First time render), 139.184s render

    Benchmark Results - 2x Gigabyte RTX 2080 Ti Gaming OC 11G
    Daz Studio Version: 4.12.0.33 Public Beta
    Optix Prime Acceleration: OFF
    Total Rendering Time: 2 minutes 17.11 seconds
    CUDA device 0 (GeForce RTX 2080 Ti): Scene processed in 1.992s (Second time render), 135.201s render
    CUDA device 1 (GeForce RTX 2080 Ti): Scene processed in 1.992s (Second time render), 135.201s render

    Post edited by Aala on
  • outrider42outrider42 Posts: 2,194
    I like that comparison to 4.10. That really shows that while 4.12 beta is faster than 4.11, it is STILL not as fast as good old 4.10 with OptiX enabled. My own tests mirrored this as I still have 4.10 installed, too. Sadly we have no way of ever testing 4.10 with RTX cards.

    While we cannot directly compare GPU results to different versions of Iray, we can compare the different versions of Iray against itself when using the same hardware. 4.10 is still the fastest version of Iray for people who do not have RTX cards. This is why I kept 4.10 around. I suppose if I make the jump to RTX then I wont have much choice but to let 4.10 go. But until then, I will keep rocking 4.10 for a while longer.

    The denoiser can be an equalizer. There are specific times I like to use it. In those cases, 4.12 will do.

    BTW, there is one special advantage to Iray using full OptiX instead of Prime which may have been glossed over. OptiX does not need to be recompiled for every new GPU launch, unlike Prime. If you you all remember it took a painfully long time for Daz Iray to support Pascal after it launched. Pascal cards were paper weights when it came to Iray. I know Turing worked at launch, but I'll contend that we got lucky Turing worked with 4.11 because it is similar to Volta. Keep in mind that Volta support was added to 4.11 in the months prior to Turing's release. Of course major hardware changes might need updates for support, like the ray tracing cores.

    At any rate, what I'm saying is that hopefully the long agonizing wait for new GPUs to be supported by Iray should be a thing of the past now that Iray uses the true OptiX. So I would expect the next generation of GPUs to be fully zippered by Iray out of the box.
  • rrwardrrward Posts: 268
    ETA that's 4x 2080ti's in there now

    [HomerSimpsonDrool.jpg]

  • RayDAntRayDAnt Posts: 584
    edited September 2019

    A new day, a new DS Beta/Iray release!

     

    System Configuration
    System/Motherboard: Gigabyte Z370 Aorus Gaming 7
    CPU: Intel i7-8700K @ stock (MCE enabled)
    GPU: Nvidia Titan RTX @ stock (watercooled)
    System Memory: Corsair Vengeance LPX 32GB DDR4 @ 3000Mhz
    OS Drive: Samsung Pro 970 512GB NVME SSD
    Asset Drive: Sandisk Extreme Portable SSD 1TB
    Operating System: Windows 10 Pro version 1903
    Nvidia Drivers Version: 431.60 GRD WDDM (except where noted)
    Daz Studio Version: 4.12.0.042 Beta 64-bit
    Optix Prime Acceleration: On*

    Benchmark Results - Titan RTX Only (TCC mode enabled)
    Total Rendering Time: 3 minutes 56.96 seconds
    CUDA device 0 (TITAN RTX): 1800 iterations, 3.209s init, 230.966s render
    Iteration Rate: (1800 / 239.117) = 7.793 iterations per second
    Loading Time: ((0 + 180 + 56.96) - 230.966) = (236.96 - 230.966) = 5.994 seconds

    Benchmark Results - Titan RTX Only
    Total Rendering Time: 4 minutes 18.82 seconds
    CUDA device 0 (TITAN RTX): 1800 iterations, 3.117s init, 253.092s render
    Iteration Rate: (1800 / 253.092) = 7.112 iterations per second
    Loading Time: ((0 + 240 + 18.82) - 253.092) = (258.82 - 253.092) = 5.728 seconds

    Benchmark Results - Titan RTX + i7-8700K
    Total Rendering Time: 4 minutes 11.79 seconds
    CUDA device 0 (TITAN RTX): 1800 iterations, 3.263s init, 244.974s render
    CPU: 110 iterations, 2.862s init, 246.195s render
    Iteration Rate: (1800 / 246.195) = 7.311 iterations per second
    Loading Time: ((0 + 240 + 11.79) - 246.195) = (251.79 - 246.195) = 5.595 seconds

    Benchmark Results - i7-8700K Only
    Total Rendering Time: 1 hours 2 minutes 34.70 seconds
    CUDA device 0 (TITAN RTX): 1800 iterations, 2.461s init, 3750.235s render
    Iteration Rate: (1800 / 3750.235) = 0.480 iterations per second
    Loading Time: ((3600 + 120 + 34.70) - 3750.235) = (3754.70 - 3750.235) = 4.465 seconds

     

    PS: Should get to bringing the main tables up to speed over the next day or so.

    Post edited by RayDAnt on
  • bluejauntebluejaunte Posts: 1,358

    So where do we discuss this stuff now?

  • RayDAntRayDAnt Posts: 584

    So where do we discuss this stuff now?

    frown

  • ebergerlyebergerly Posts: 2,842
    edited July 2019

    So where do we discuss this stuff now?

    Just a thought...

    Clearly, our often contentious tech discussions are seen as very disruptive, and that's certainly understandable. 

    So I'm thinking it might be best for all of us interested in those types of tech discussions to just pick up and move to an unrelated discussion website somewhere. I imagine that would be a huge relief for all involved. Not sure where we'd go, but I imagine there are many choices. 

    Post edited by ebergerly on
  • RayDAntRayDAnt Posts: 584

    If someone decides to start up something new here/somewhere else, drop me a PM and I'll link to it at the top of this thread.

  • https://www.reddit.com/r/Daz3D/ ?

    maybe? not sure. maybe i don't actually want to post with my normal account on that.

  • AelfinMaegikAelfinMaegik Posts: 47
    edited July 2019

     

    Aala said:

    (Just upgraded from 2x 1080 Ti's to 2x 2080 Ti's, did some benchmarks)

    Benchmark Results - 2x Gigabyte RTX 2080 Ti Gaming OC 11G
    Daz Studio Version: 4.12.0.33 Public Beta
    Optix Prime Acceleration: ON
    Total Rendering Time: 2 minutes 22.0 seconds
    CUDA device 0 (GeForce RTX 2080 Ti): Scene processed in 7.877s (First time render), 139.184s render
    CUDA device 1 (GeForce RTX 2080 Ti): Scene processed in 7.880s (First time render), 139.184s render

    Benchmark Results - 2x Gigabyte RTX 2080 Ti Gaming OC 11G
    Daz Studio Version: 4.12.0.33 Public Beta
    Optix Prime Acceleration: OFF
    Total Rendering Time: 2 minutes 17.11 seconds
    CUDA device 0 (GeForce RTX 2080 Ti): Scene processed in 1.992s (Second time render), 135.201s render
    CUDA device 1 (GeForce RTX 2080 Ti): Scene processed in 1.992s (Second time render), 135.201s render

    WOW! This really got me wondering since the prime off was faster. I downloaded 4.12.0.42, turned NVLINK/SLI to OFF and tried this, too. I'm really surprised. I am running air cooled, non OC, and  I'm shocked at these numbers. Why is prime off faster? (I am kinda peeved at myself- I changed two variables here... I updated beta version AND i turned off nvlink/sli)

    Daz Studio Version: 4.12.0.42 BETA
    Optix Prime Acceleration: ON

    Benchmark Results
    DAZ_STATS
    2019-07-30 21:54:57.007 Finished Rendering
    2019-07-30 21:54:57.042 Total Rendering Time: 2 minutes 18.77 seconds
    IRAY_STATS
    2019-07-30 21:55:31.063 Iray [INFO] - IRAY:RENDER ::   1.0   IRAY   rend info : Device statistics:
    2019-07-30 21:55:31.063 Iray [INFO] - IRAY:RENDER ::   1.0   IRAY   rend info : CUDA device 0 (GeForce RTX 2080 Ti): 882 iterations, 4.569s init, 129.627s render
    2019-07-30 21:55:31.063 Iray [INFO] - IRAY:RENDER ::   1.0   IRAY   rend info : CUDA device 1 (GeForce RTX 2080 Ti): 918 iterations, 4.755s init, 128.987s render
    Iteration Rate: (1800 / 129.67) 13.881 iterations per second
    Loading Time: ((0 * 3600 + 2 * 60 + 18.77) - 129.67) 9.1 seconds

     

    Daz Studio Version: 4.12.0.42 BETA
    Optix Prime Acceleration: OFF

    Benchmark Results
    DAZ_STATS
    2019-07-30 22:08:00.921 Finished Rendering
    2019-07-30 22:08:00.956 Total Rendering Time: 2 minutes 12.65 seconds
    IRAY_STATS
    2019-07-30 22:08:17.102 Iray [INFO] - IRAY:RENDER ::   1.0   IRAY   rend info : Device statistics:
    2019-07-30 22:08:17.102 Iray [INFO] - IRAY:RENDER ::   1.0   IRAY   rend info : CUDA device 0 (GeForce RTX 2080 Ti): 883 iterations, 4.572s init, 123.675s render
    2019-07-30 22:08:17.102 Iray [INFO] - IRAY:RENDER ::   1.0   IRAY   rend info : CUDA device 1 (GeForce RTX 2080 Ti): 917 iterations, 4.722s init, 122.562s render
    Iteration Rate: (1800 / 123.675) 14.554 iterations per second
    Loading Time: ((0 * 3600 + 2 * 60 + 12.65) - 123.675) 8.975seconds

    Post edited by AelfinMaegik on
  • outrider42outrider42 Posts: 2,194

    Because you have RTX. Also, if by chance Nvlink was working, that can actually slow performance in Nvlink mode. Nvlink mode should only be used when you need to pool the VRAM, assuming that is possible. This is because of how Nvlink works, since the cards must talk to each other more and share more data this slows them down a little bit VS letting them render apart without Nvlink.

    The Daz Reddit is a ghost town. Last post was 4 DAYS ago, and scrolling just a few posts takes you 27 days back. It is a shame that Daz has to be like the NFL...No Fun League.

  • AelfinMaegikAelfinMaegik Posts: 47
    edited July 2019

    Because you have RTX. Also, if by chance Nvlink was working, that can actually slow performance in Nvlink mode. Nvlink mode should only be used when you need to pool the VRAM, assuming that is possible. This is because of how Nvlink works, since the cards must talk to each other more and share more data this slows them down a little bit VS letting them render apart without Nvlink.

    The Daz Reddit is a ghost town. Last post was 4 DAYS ago, and scrolling just a few posts takes you 27 days back. It is a shame that Daz has to be like the NFL...No Fun League.

    i understand your points about nvlink. totally expected a minor improvement. but why is prime acceleration OFF faster than ON?

    Post edited by AelfinMaegik on
  • outrider42outrider42 Posts: 2,194
    Because Prime is not made for RTX. RTX is the whole reason Iray needed to be overhauled. Prime has some key differences between it and normal OptiX. These differences allowed normal GPUs to render faster in the past, but the new ray tracing cores caused problems that required this change. I'm rather surprised that Prime is even available to RTX cards at all. I suppose it still there in case you have a GTX card and a RTX card in your system. Or maybe its for CPU. Full OptiX has no CPU support at all.

    It would be great if Daz could add a toggle for ray tracing cores, even if there is no practical reason for it other than gauging how much faster RTX on might be. However, people are quite curious about such things.
  • RayDAntRayDAnt Posts: 584
    edited July 2019

     

    Aala said:

    (Just upgraded from 2x 1080 Ti's to 2x 2080 Ti's, did some benchmarks)

    Benchmark Results - 2x Gigabyte RTX 2080 Ti Gaming OC 11G
    Daz Studio Version: 4.12.0.33 Public Beta
    Optix Prime Acceleration: ON
    Total Rendering Time: 2 minutes 22.0 seconds
    CUDA device 0 (GeForce RTX 2080 Ti): Scene processed in 7.877s (First time render), 139.184s render
    CUDA device 1 (GeForce RTX 2080 Ti): Scene processed in 7.880s (First time render), 139.184s render

    Benchmark Results - 2x Gigabyte RTX 2080 Ti Gaming OC 11G
    Daz Studio Version: 4.12.0.33 Public Beta
    Optix Prime Acceleration: OFF
    Total Rendering Time: 2 minutes 17.11 seconds
    CUDA device 0 (GeForce RTX 2080 Ti): Scene processed in 1.992s (Second time render), 135.201s render
    CUDA device 1 (GeForce RTX 2080 Ti): Scene processed in 1.992s (Second time render), 135.201s render

    WOW! This really got me wondering since the prime off was faster. I downloaded 4.12.0.42, turned NVLINK/SLI to OFF and tried this, too. I'm really surprised. I am running air cooled, non OC, and  I'm shocked at these numbers. Why is prime off faster? (I am kinda peeved at myself- I changed two variables here... I updated beta version AND i turned off nvlink/sli)

    Daz Studio Version: 4.12.0.42 BETA
    Optix Prime Acceleration: ON

    Benchmark Results
    DAZ_STATS
    2019-07-30 21:54:57.007 Finished Rendering
    2019-07-30 21:54:57.042 Total Rendering Time: 2 minutes 18.77 seconds
    IRAY_STATS
    2019-07-30 21:55:31.063 Iray [INFO] - IRAY:RENDER ::   1.0   IRAY   rend info : Device statistics:
    2019-07-30 21:55:31.063 Iray [INFO] - IRAY:RENDER ::   1.0   IRAY   rend info : CUDA device 0 (GeForce RTX 2080 Ti): 882 iterations, 4.569s init, 129.627s render
    2019-07-30 21:55:31.063 Iray [INFO] - IRAY:RENDER ::   1.0   IRAY   rend info : CUDA device 1 (GeForce RTX 2080 Ti): 918 iterations, 4.755s init, 128.987s render
    Iteration Rate: (1800 / 129.67) 13.881 iterations per second
    Loading Time: ((0 * 3600 + 2 * 60 + 18.77) - 129.67) 9.1 seconds

     

    Daz Studio Version: 4.12.0.42 BETA
    Optix Prime Acceleration: OFF

    Benchmark Results
    DAZ_STATS
    2019-07-30 22:08:00.921 Finished Rendering
    2019-07-30 22:08:00.956 Total Rendering Time: 2 minutes 12.65 seconds
    IRAY_STATS
    2019-07-30 22:08:17.102 Iray [INFO] - IRAY:RENDER ::   1.0   IRAY   rend info : Device statistics:
    2019-07-30 22:08:17.102 Iray [INFO] - IRAY:RENDER ::   1.0   IRAY   rend info : CUDA device 0 (GeForce RTX 2080 Ti): 883 iterations, 4.572s init, 123.675s render
    2019-07-30 22:08:17.102 Iray [INFO] - IRAY:RENDER ::   1.0   IRAY   rend info : CUDA device 1 (GeForce RTX 2080 Ti): 917 iterations, 4.722s init, 122.562s render
    Iteration Rate: (1800 / 123.675) 14.554 iterations per second
    Loading Time: ((0 * 3600 + 2 * 60 + 12.65) - 123.675) 8.975seconds

    Yeah, this is extremely interesting.

    Logically neither you nor Aala should be seeing any dscernable difference in render times between OptiX Prime On vs OptiX Prime off tests here because, in the case of rendering with only RTX GPUs on Daz Studio 4.12+, OptiX Prime is disabled internally by Iray and all raytracing is farmed directly out to the RTCores on the cards themselves. Currently the only programmatic difference between OptiX Prime checked/unchecked with 20XX cards in 4.12+ is that one state pops up errors in the log file about optix prime being improperly called for when it isn't supported versus a clean log file with no extraneous error messages (renders still finish successfully either way.) At least that's been the case with my Titan RTX under 4.12.0.033. Could you post log files (at least the portions starting after "Rendering image") for both an Optix Prime On and Off test run? I'm gonna try doing the same with my Titan RTX tomorrow on 4.12.0.042 as well, and would love to be able to compare what my logs look like compared to yours.

    Btw could you also try OptiX Prime on vs off with just a single 2080Ti activated and see if the same patterns are true? It could be that what both you and Aala are seeing here is a multi-GPU thing. Specifically I'm wondering if perhaps the current DS/Iray version has a bug where having more than one RTX card activated at once is causing Iray to revert back to its pre-RTX GPU era behavior where OptiX Prime is used rather than RTCores for raytracing. Meaning that there may be a DS/Iray bugfix in the near future that could net both you and Aala a noticeable speedup in dual 2080Ti rendering scenarios regardless of whether OptiX Prime is checked on or off..

    Post edited by RayDAnt on
  • RayDAntRayDAnt Posts: 584
    edited July 2019
    I'm rather surprised that Prime is even available to RTX cards at all. I suppose it still there in case you have a GTX card and a RTX card in your system.

     That's been what my testing so far has indicated. OptiX Prime still registers as on/off in the log file as it always has in the case of my GTX 1050 laptop, whereas my RTX machine says its unsupported. I'm assuming that if I had a GTX card in my RTX system as well and had it selected instead, OptiX Prime would once again appear as a valid configuration option in the log file like it has in the past.

     

    Or maybe its for CPU. Full OptiX has no CPU support at all.

    So far it seems to be just Embree used for all raytracing functionality in CPU based rendering on 4.12+ (based, again, on my testing so far.) I remember reading somewhere that speedups in Embree's codebase have pretty much made it independently as good (if not better) than OptiX Prime ever was in accelerating raytracing. Thus the utilization change.

     

    It would be great if Daz could add a toggle for ray tracing cores, even if there is no practical reason for it other than gauging how much faster RTX on might be. However, people are quite curious about such things.

    Would be nice, but I'm pretty sure that's soemthing Iray's developers would have to first develop and then Daz's developers implement in order for us to have use of it. Not inclined to hold my breath on all that happening.

    Post edited by RayDAnt on
Sign In or Register to comment.