The iray cometh

DoctorMarimbas · 2010-06-27 14:32:33 UTC

Just so you people know… I got the 3.1 CUDA drivers for Mac OS X
3.1.1 actually. It seems to be 64-bit but iray still throws this message:

IRAY 0.2 warn : no CUDA capable devices found
IRAY 0.2 info : using CPU backend with 8 threads

deviceQuery reports this:

CUDA Device Query (Runtime API) version (CUDART static linking)

There are 2 devices supporting CUDA

Device 0: “Quadro FX 4800”
CUDA Driver Version: 3.10
CUDA Runtime Version: 3.10
CUDA Capability Major revision number: 1
CUDA Capability Minor revision number: 3
Total amount of global memory: 1610285056 bytes
Number of multiprocessors: 24
Number of cores: 192
Total amount of constant memory: 65536 bytes
Total amount of shared memory per block: 16384 bytes
Total number of registers available per block: 16384
Warp size: 32
Maximum number of threads per block: 512
Maximum sizes of each dimension of a block: 512 x 512 x 64
Maximum sizes of each dimension of a grid: 65535 x 65535 x 1
Maximum memory pitch: 2147483647 bytes
Texture alignment: 256 bytes
Clock rate: 1.20 GHz
Concurrent copy and execution: Yes
Run time limit on kernels: Yes
Integrated: No
Support host page-locked memory mapping: Yes
Compute mode: Default (multiple host threads can use this device simultaneously)
Concurrent kernel execution: No
Device has ECC support enabled: No

Device 1: “GeForce 8800 GT”
CUDA Driver Version: 3.10
CUDA Runtime Version: 3.10
CUDA Capability Major revision number: 1
CUDA Capability Minor revision number: 1
Total amount of global memory: 536674304 bytes
Number of multiprocessors: 14
Number of cores: 112
Total amount of constant memory: 65536 bytes
Total amount of shared memory per block: 16384 bytes
Total number of registers available per block: 8192
Warp size: 32
Maximum number of threads per block: 512
Maximum sizes of each dimension of a block: 512 x 512 x 64
Maximum sizes of each dimension of a grid: 65535 x 65535 x 1
Maximum memory pitch: 2147483647 bytes
Texture alignment: 256 bytes
Clock rate: 1.50 GHz
Concurrent copy and execution: Yes
Run time limit on kernels: Yes
Integrated: No
Support host page-locked memory mapping: No
Compute mode: Default (multiple host threads can use this device simultaneously)
Concurrent kernel execution: No
Device has ECC support enabled: No

deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 3.10, CUDA Runtime Version = 3.10, NumDevs = 2, Device = Quadro FX 4800, Device = GeForce 8800 GT

PASSED

What am I missing?

xtrm3d · 2010-06-28 10:01:52 UTC

Dear polygon engineer, What is your favorite workflow for GPU rendering? exported obj in Octane or Maya/Iray?

did not try iray in maya for now …
only in bunkspeed shot ( wich is build around iray )
i guess that i would stick with it for now …
still love octane …
but somehow … i prefer to have the abillity to use cpu or hybrid mode if i have to

pixym · 2010-06-28 10:04:27 UTC

Thanks for the reply Christopher.

xtrm3d · 2010-06-28 10:49:21 UTC

no problem … as a matter of fact i did just add two new car rendering in my cgportfolio …

pixym · 2010-07-11 20:21:15 UTC

Any new iray tests/results guys?

raggy2k4 · 2010-07-18 17:33:44 UTC

Can work firepro v8700 with iray?
Or only with nvidia cards?

moulder6 · 2010-07-21 01:20:48 UTC

Iray takes advantage of the Nvidia CUDA architecture. So, for non Nvidia cards, there’s no GPU processing.
From what i saw so far, there’s no GPU processing for most of the Nvidia cards, either, when u try to integrate iray, into max/maya, as discussed here.

cavemen · 2010-07-29 21:41:41 UTC

If i copy all the dlls from the standalone do i still need it to be installed on my machine ?

rBrady · 2010-07-30 08:14:15 UTC

All of the nvidia gpus from the last couple years will run cuda. The 8 series and higher all do. Nothing has shipped with a 7 series card in a while. To really take advantage you need to have a higher end card though. The cuda performance of a 9200m (16 processors) and a GTX480(480 processors) is absolutely huge. Not sure there is any point in running cuda on less than a 9600gt (64 processors).

I have a gtx470 on one machine and two gtx 260s on the other and am impressed with the speed every time I run iray.

Samuka3000 · 2010-07-30 15:22:09 UTC

I was thinking, iRay and OpenCL, is it a viable option for the future?

cgbeige · 2010-07-30 16:43:54 UTC

I asked NVIDIA about that at SIGGRAPH and they basically said “no” in marketing terms. I wrote a piece on GPU renderers that was published today, for anyone interested:

http://arstechnica.com/gadgets/news/2010/07/gpu-renderers-proliferate-show-newfound-maturity.ars

Samuka3000 · 2010-08-01 14:42:24 UTC

Money talks!

malloriwooseung · 2010-08-01 21:33:40 UTC

Hello rBeady
I have some question…
i copy 3 iray dlls to maya2011 but iray dosn’t work with gpu.
-[The render message OK.But render black]-
I-ray cpu renders ok.

and MR -standalone was same …(not work with gpu)
i have 9600 GT.
Do you have any special setting?

DuttyFoot · 2010-08-01 22:01:32 UTC

if you didn’t see the nvidia presentation at sig its over in the area. iray is on the list of things that is shown. it starts at 10:06 and the demo is done with max

http://area.autodesk.com/siggraph/areatv-guest

wizlon · 2010-08-02 01:51:47 UTC

Just a quick note to say caustics work with Vray RT GPU.

SreckoM · 2010-08-02 07:54:18 UTC

Check also Shot at http://www.bunkspeed.com/shot/index.html, it is based on Iray, I am very impressed with speed and workflow. Worth checking.

rBrady · 2010-08-02 12:38:45 UTC

My guess would be that you need to update your drivers, 192(or was it 193?) was the last nvidia driver that I used with iray. Its beta type software, so your are bound to have issues. Hope you can get it figured out.

As to OpenCL, I am confidant that just about everything will move to OpenCL eventually. Thats probably a long ways out. Every developer I have talked to said that CUDA is a pain to develop for and OpenCL is horrible, almost unusable. Hopefully the folks at kronos, or whoever is managing that project, get their act together soon so we can get more OpenCL software out there. For now CUDA is a much more mature API, that is evident in how much software is being released for CUDA and how little is being released for OpenCL. I think that will change, and the sooner the better. But don’t get mad at the developers for using CUDA, get mad at the folks putting together OpenCL.

LinchpinZA · 2010-10-25 05:12:31 UTC

dagon1978:

this is a long story, iray is using the internal builtin shaders (builtin_architectural is the mia_material simplification), those shaders assume the light is irradiance, but most traditional light shaders don’t return irradiance, but rather irradiance divided by pi (the average radiance for diffuse surfaces)
mia_material and the daylight system are built to work in the old way, so there is a discrepancy between the two systems
obviously in an integrated version of iray all of this should be under the hood, but…

this is also why i’m using this with IBL when using mia_materials:
setAttr -type “string” miDefaultOptions.stringOptions[31].name “environment lighting scale”;
setAttr -type “string” miDefaultOptions.stringOptions[31].type “float”;
setAttr -type “string” miDefaultOptions.stringOptions[31].value “0.318”;

if you use iray you need to change the value to 1.0

Hi Dagon,

Thanks for this useful post. I’m using an IBL testing iray. I was just playing around and I noticed changing the ‘environment lighting scale’ string has the the same effect as adjusting the color gain for the IBL node (using maya here). Are there any differences between adjusting these two variables?

Also should you multiply direct lights and indirect lighting( IBL) by 3.14 when using iray ? I understand the reasoning behind multiplying your lighting by pi when using iray, but theres one thing I don’t understand this magical multiplication of ‘3180’. I’m saying this after seeing your lighting scale of 0.318.

Off the topic of iray, I see this multiplication of 3180 used alot with creating HDR scene lighting. For example djx’s sun&sky script automatically multiplies the rgb scale of the physical sky by 3180 taking it from the default rgb scale of 0.0001 to 0.318. There was also a thread a while ago where people were testing the physical correctness of lights in maya, and they found that multiplying the intensity of lights by this number ‘3180’ gave the most physically correct result.

I have since been using this as a base for creating HDR scenes, multiplying my lighting by this number and leaving my cm2 factor in my exposure photographic shader set to ‘1’. I have never quite understood the reasoning behind this. Would greatly appreciate it if you could explain this for me.

-Chris

chuckie7413 · 2010-11-23 03:57:24 UTC

Hi guys,

I am getting back into iray testing and trying to render some old interior scenes we did just with final gather.

One of my issues is with textures.

Before all textures had maya gamma correct nodes flugged into them to linearise them. However these nodes do not appear compatible with iray.

Can I ask how others have rendered their textures for linear workflow? Have you just coverted them to linear in photoshop/nuke first or are you able to get a gamma correct node/mip_gamma_gain node to work with iray?

Many thanks,

Richard

TearGasJazz · 2011-02-14 09:06:22 UTC

If using Maya 2011 you can simply put Input to say sRGB, meaning that all your textures you plug in are handled as gamma corrected, which is the same as using the gammacorrect node.
Then under Output, you put it to Linear. Using a exposure_simple node on the camera is recommended with gamma set to 2.2.

The 3ds Max 2011 Subscription Package obviously comes with iRay support, and by support I mean some kind of GUI to control it all, at least thats what I heard.