Page 11 of 43 FirstFirst ... 91011121321 ... LastLast
Results 151 to 165 of 639

Thread: 11.5's BenchmarkMarbles.lws - share your machine's render time here

  1. #151
    yep, you could do that. Make 3 scenes, tint all the colours to r,g and b, alter each glass to have the appropriate refractive index... and do 3 renders... strip each one down to a single channel and add them together. So yes basically 3 times as slow. It's why dispersion is slow as it literally does this for you, but the mc radiosity doesn't. Lw needs a new photon-mapping based caustics engine and or improved radiosity to handle it properly (directional rays is painfully slow!) Hybrid photon / final gather engines are usually the way to go as you can importance-sample the best solution. LW is a little old in this regard and could use a little love. The old caustics engine is just to awful to even try!

    Pretty photo, and yes, to get that kind of quality, you need a dedicated engine to handle it, with ohoton mapping.

  2. #152
    digital artist azman_821009's Avatar
    Join Date
    Aug 2005
    Location
    Malaysia
    Posts
    79
    Custom - built PC

    Render time: 1h 16m 5s
    CPU: Intel i7-3960X 3.30Ghz
    RAM: 64 GB DDR3
    Overclocking: None
    OS: Windows 7 Pro 64-bit

    Click image for larger version. 

Name:	ScreenShot009.jpg 
Views:	156 
Size:	1,015.9 KB 
ID:	111857

  3. #153
    I realized I forgot my laptop. I brought forward the earlier tests for comparison, this laptop is the last one.

    Custom-built PC
    Render time: 2h 42m 16s
    CPU: 3.0 GHz AMD Phenom 1075T hex core (Observed threads =6)
    RAM: 32 GB DDR1333 8-8-8-20 (this is actually underclocked DDR1600 CL9)
    Overclocking: None (3.0GHz per CPU-Z)
    OS: Windows 7 Pro 64-bit

    Custom-built PC
    Render time: 1h 37m 26s
    CPU: 3.5 GHz i7 3770k Quad core with HT (Observed threads = 8)
    RAM: 8 GB DDR1333 9-9-9-24
    Overclocking: to 4.2GHz (per CPU-Z)
    OS: Windows 8 Pro 64-bit

    Custom-built PC
    Render time: 2h 27m 43s
    CPU: 3.4 GHz i5 3570k Quad core (no HT) (Observed threads = 4)
    RAM: 8 GB DDR1333 9-9-9-24
    Overclocking: none (though CPU-Z reports the CPU at 3.7GHz during render)
    OS: Windows 8 Pro 64-bit

    HP DV6t laptop
    Render Time: 2hr 7m 0s
    CPU: i7-3720QM 2.6GHz with HT (Observed threads = 8)
    RAM: 8GB DDR1333 11 11 11 28
    Overclocking: that's a laugh (the BIOS is so lacking in features it may as well be blank -- so, no) (CPU-Z shows 3.4GHz during render)
    OS: Windows 7 Home Premium 64 bit
    Last edited by CourtJester; 02-24-2013 at 04:52 PM. Reason: brought in old data

  4. #154
    I evaluate at about 21 hours since i got 31% done in 7 hours and then i gave up. Athlon 64 X2 3600+....i know.
    That might make you feel good, but 2 hours/frame is also too long if you intend to make a movie.
    How long with Octane?

  5. #155
    I am also searching free time to test the scene in Octane ... anyway soon you are going to be able to test scenes with Octane yourself

    -Juanjo
    Juanjo Gonzalez
    Computer Graphics Software Developer

  6. #156
    Electron wrangler jwiede's Avatar
    Join Date
    Aug 2007
    Location
    San Jose, CA
    Posts
    6,772
    Quote Originally Posted by Dave Jerrard View Post
    I don't know how you figure that LightWave's glass isn't realistic, since I'm using the Dielectric material, which uses the same math functions that pretty much every other renderer uses. I've tested this against physical objects and it's a match. I'm now wondering if VRay is taking shortcuts.
    LW's dielectric always seems to be missing details at edges as if internal reflections weren't quite correct. In real objects, as well as Maxwell and Vray renders of physical glass, they just always have extra details at the boundary edges. I'm not doing a good job of explaining what I mean, I'll show rendered examples of what I mean here in a bit.

    Quote Originally Posted by Dave Jerrard View Post
    It's working exactly as intended. Any irregularities at a glancing angle will have more exaggerated effects on the refraction than they will when viewed straight on; that's the nature of refraction. Again, I've used actual physical reference for glass when setting this up.
    Can you clarify precisely what precisely was the physical aspect the turbulence+fresnel was supposed to replicate? Are you using that to approximate minor surface displacements? Or trying to replicate something within the glass itself? I need a better understanding of the "physical intent" behind the nodes, because that's how both Maxwell and Vray expect such things to be defined. I can just hook up something akin to what you did, but as I'm not quite sure what physical aspect you were going for, it's difficult for me to express it in physical terms.

    Quote Originally Posted by Dave Jerrard View Post
    The swirled marble was also a test of the the color attenuation of Dielectric in a multiple IOR environment. An earlier version of the marble had the a single IOR for everything, and it looked pretty boring...
    It definitely adds visual interest, no question. One choice I do question a bit is to leave LW's caustics turned off, despite it being an image of a collection of glass spheres. Was that because you already felt you'd reached an adequate benchmarking level? Still, easy enough to compensate.
    John W.
    LW2015.3UB/2019.1.4 on MacPro(12C/24T/10.13.6),64GB RAM, NV 980ti

  7. #157
    With MC radiosity, use transparency and directional rays, you get caustic like effects. Not as good as a dedicated system, but pretty enough, and visible in the render. Certainly better than the awful caustics engine in LW.

    I'm interested to see the added 'edge effects', and how other renderers compare. to be fair, LW is far from a pbrt system. It's very flexible, but it takes a lot of coaxing to get good results, and it mostly it's 'close' and quite often pretty, just not quite there...

  8. #158
    Electron wrangler jwiede's Avatar
    Join Date
    Aug 2007
    Location
    San Jose, CA
    Posts
    6,772
    Quote Originally Posted by Dave Jerrard View Post
    I don't know how you figure that LightWave's glass isn't realistic, since I'm using the Dielectric material, which uses the same math functions that pretty much every other renderer uses. I've tested this against physical objects and it's a match. I'm now wondering if VRay is taking shortcuts.
    Attachment 111733
    Take a close look at the edges of the glass twists "inside" the glass marble, and tell me if you see anything obvious missing?

    Here's a hint: It's quite visible inside the marble with the glass bubbles, and while wouldn't be visible viewed directly on, should be quite visible towards the edges of the glass twists.

    P.S. Had a moment when I finally recognized the cause of the issue I've been trying to describe in the "glass twists" marble in Dave J's pic (attached), I blame too much work and fatigue for not recognizing it sooner.
    Last edited by jwiede; 02-26-2013 at 07:50 PM.
    John W.
    LW2015.3UB/2019.1.4 on MacPro(12C/24T/10.13.6),64GB RAM, NV 980ti

  9. #159
    pass:sword OFF's Avatar
    Join Date
    Feb 2003
    Location
    Russia
    Posts
    1,050
    48min 24 sec on lw 11.5 trial version.
    2xE5-2650/32gb ram/120gb ssd

  10. #160
    Super Member JonW's Avatar
    Join Date
    Jul 2007
    Location
    Sydney Australia
    Posts
    2,235
    Haven't got a marble to compare twist!

    But I have been fiddling with the clear glass. Added 3 sets of clear marbles & changed the refractive index of the clear class (the coloured spirals left as is) for Red Green & Blue lights for each set of 3 marbles. 2 of 3 of each of the clear marbles are excluded from each of the other coloured lights. Also changed the air bubble refractive index.

    Red 1.5092
    Green 1.5153
    Blue 1.5169

    Would have liked to add these as well
    Yellow 1.5112
    Violet 1.5214


    This is a nice scene it would be nice to have a proper refractive setup to render.
    Attached Thumbnails Attached Thumbnails Click image for larger version. 

Name:	Marbles_clearx3xRGBlights.jpg 
Views:	124 
Size:	144.4 KB 
ID:	112029   Click image for larger version. 

Name:	Marbles_clearx3_RGBlights.png 
Views:	123 
Size:	1.53 MB 
ID:	112030   Click image for larger version. 

Name:	Marbles_clearx3_RGBlights_cropx16.jpg 
Views:	121 
Size:	96.6 KB 
ID:	112031   Click image for larger version. 

Name:	Marbles_clearx3xRGBlights_cropx16.png 
Views:	115 
Size:	1.64 MB 
ID:	112032  
    Procrastination, mankind's greatest labour saving device!

    W5580 x 2 24GB, Mac Mini, Spyder3Elite, Dulux 30gg 83/006 72/008 grey room,
    XeroxC2255, UPS EvolutionS 3kw+2xEXB

  11. #161
    Super Member JonW's Avatar
    Join Date
    Jul 2007
    Location
    Sydney Australia
    Posts
    2,235
    I had noticed Memory Segment limit has been increased from 2000.

    It is now 16384 so if you are doing massive images & have lots of ram stick it on 16384!
    Procrastination, mankind's greatest labour saving device!

    W5580 x 2 24GB, Mac Mini, Spyder3Elite, Dulux 30gg 83/006 72/008 grey room,
    XeroxC2255, UPS EvolutionS 3kw+2xEXB

  12. #162
    Registered User
    Join Date
    Feb 2013
    Location
    North East US
    Posts
    113
    47 mins 13 seconds

    Click image for larger version. 

Name:	Screen Shot 2013-03-03 at 1.20.20 AM.png 
Views:	141 
Size:	590.2 KB 
ID:	112148

    2008 Mac Pro Upgraded
    2 x 3.33 GHz Intel Xeon X5680
    16 GB 1066 Mhz Ram
    ATI Radeon 5890 1 GB

    Not too bad for how old this computer is and how little I spent to upgrade it I guess
    Last edited by KBS756; 03-03-2013 at 12:57 AM.

  13. #163
    Quote Originally Posted by KBS756 View Post
    47 mins 13 seconds

    Click image for larger version. 

Name:	Screen Shot 2013-03-03 at 1.20.20 AM.png 
Views:	141 
Size:	590.2 KB 
ID:	112148

    2008 Mac Pro Upgraded
    2 x 3.33 GHz Intel Xeon X5680
    16 GB 1066 Mhz Ram
    ATI Radeon 5890 1 GB

    Not too bad for how old this computer is and how little I spent to upgrade it I guess
    very nice indeed! so this 6 cores are working without any further modification in macpro's from 2008? is this a late 2008 nehalem macpro 4,1 or an E54XX CPU based macpro 3,1? i have both, an early 2008 macpro 3,1 and a 2009 nehalem macpro, both with dual quads here, so this is really a tempting upgrade! i have already upgraded two 2006 macpro's with success, so this shouldn't be any more difficult. any specific instructions for this model which you have used? any hint where to geet good priced CPU's besides from ebay? thanks for any info on this one...

    cheers

    markus
    Last edited by 3dworks; 03-03-2013 at 08:21 AM.
    3dworks visual computing
    demo reel on vimeo
    instagram

    OSX 10.12.x, macpro 5.1, nvidia gtx 980 ti, LW2015.x / 2018.x, octane 3.x

  14. #164
    Registered User
    Join Date
    Feb 2013
    Location
    North East US
    Posts
    113
    It's a 4,1 Mac Pro that originally had dual 2.93 GHz Processors, I upgraded the firmware to make it think it was a 5,1 so it could handle both 6 core processors. Which I got over here

    http://forum.netkas.org/index.php/topic,852.0.html

    I bought my processors from http://www.itcreations.com/view_prod...oduct_id=22787, but their price has gone up quite a lot since I did.

    If you have a dual processor Mac Pro 4,1 the only real annoyances with installation is that there are no clamps holding the processors down, so you have to put the heat sink down carefully with uniform pressure. Also the original CPUs in a dual processor Mac Pro are lidless so you need to space them to not crush them in the socket with washers .... If you have the one after mine you don't have to worry about these.

    Pretty straight forward install other then the pressure thing ... drove me nuts at first but it is very doable just have to be very careful with being uniform with the pressure and using an X pattern to tighten it so not to have the cpu come out of it's socket.

  15. #165
    eye kan kode gud jrandom's Avatar
    Join Date
    Dec 2009
    Location
    Seattle, WA
    Posts
    1,143
    Quote Originally Posted by KBS756 View Post
    I bought my processors from http://www.itcreations.com/view_prod...oduct_id=22787, but their price has gone up quite a lot since I did.
    Drat. I was all hopeful that I might be able to upgrade from the 2.66GHz dual 6-cores to the 3.3GHz ones, but for $1500 it's not worth it.

Page 11 of 43 FirstFirst ... 91011121321 ... LastLast

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •