Jump to content

padre.ayuso

Members
  • Posts

    198
  • Joined

1 Follower

Personal Information

  • Country
    United_States

padre.ayuso's Achievements

Newbie

Newbie (1/14)

10

Reputation

  1. I just saw an article from ChaosGroup yesterday, downloaded it and read it from beginning to end. I was also under the false impression that GPU is faster. It is not usually faster. It can handle more, that is, it's parallel processing, its how much more it can do at any given time compared against a CPU, which is faster at handling individual buckets with each core. Having said that, if your scene does not require much computation per pixel or buckets, then each pixel gets handled that much faster and thus the GPU can handle more than the CPU by the sheer volume it can take on against the CPU. But if each bucket/pixel requires more speed, then the CPU gains advantage over the GPU. Its basically all relative. The more GPUs you have, then the faster it will go, because it can handle that many more pixels at any given time over the CPU. Hope that makes somewhat sense.
  2. I have a new machine that I got with two Titan X cards. I'm running 3DS Max 2013 and have just installed a trial version for 3DS Max 2016 to test these cards. These cards don't seem to get detected by 3DS Max/VRay/IRay. The solution to use the cards when doing GPU rendering in VRay RT is to unselect In Process RT engine, however that makes both my cards run when I'd like to use one only for interactivity purpose or to leave one rendering going while working on another one with the other card. Has anyone found or know a solution? I tried tweaking the Environment Variables per different threads I found on different forums and I have also tried installing 3 different drivers for the cards to no avail. To solve the problem temporarily, I have just reduced the Ray Bundle Size, but I'd like to be able to just select one GPU when doing this type of work to use the other for graphics work on another scene. Thank you!
  3. Is your setting in VRay RT Tab set to Open CL? I believe CUDA is only for NVidia. And CPU, well, is only CPU.
  4. No problems steve. So I have been trying DR with other slaves and it does work, the only problem with my machine I think I have some firewall settings that does not allow for the full communication to happen between the machines and so I can't DR neither with RT nor CPU, but you follow what I wrote earlier and DR should be a cinch. Let me know how that goes.
  5. Let me correct myself, I have tried myself today and I can't seem to make them appear. I'll be working on this some more, as I remember having done this successfully in the past. You can shoot also a question on the ChaosGroup forum, which is great in answering questions and giving you support. EDIT: and I have tried it again, and what you need to do is to turn on VRay RT Server on your slave machine and it gets detected. Then it should work.
  6. Steve, make sure you have enough CPU RAM. I found out through my own tests that all the data goes through the CPU first and then it gets sent to the GPU, however those calculations get done, etc. I had a rendering with Displacement grass on and my rendering reached about 24GB of CPU memory and hang there. Apparently in VRay LC also consumes lots of memory and if I had BF + BF it would have taken 19GB of CPU memory and it would have processed, as I sent my file to Chaos to review and that's what I got back.
  7. I'd not do it in the first place, it is quite a gross concept what this guy/s do.
  8. Makes me want to puke! But then again, if this guy would ever sell my photos for 2m dollars, I'd be honored and know that I can then sell that one or others of mine for half the price and have that serious 3D workstation that Santa couldn't afford this year.
  9. Yes it can, in both, production mode and Active Shade mode, under the VRay RT tab, it has a box for Distributed Rendering, add the servers that have the cards and render away. I tried it out before. What it does not do, however, is use the RAM of the multiple GPUs and combine them, which I guess is a silly thing for me to say; but at one point I was excited that if I added to 2 4GB GPUs I'd get 8GB of RAM, but I'm unhappy to say that this is not the case.
  10. Yes off topic, but can you replicate anything like that with MR? If so, great. I agree with you, it all depends on how good you are with it. But I think the rest of the people here are giving very valuable insight and whatever the industry standard is and what companies ask for. For us, it is VRay, it was MR as of a year ago, but we found we had to go VRay as VRay became the software the rest of the pipeline was using and so it became imperative to work wit VRay.
  11. Having used MR, I think what the above has been said happens to hit the nail, I'm just going to reiterate. I have used MR and it is limited. I have used Corona and it is simple, cheap and great to use. And I'm currently using VRay because it is within the pipeline I have to fit it through, and it is a learning experience but it produces great images and it has soooo many options to do so many things I had only wised from MR for 7 years. Having said that, you can go Corona if you don't want to pay the extra money for VRay (although worth it) and if you want to freelance.
  12. I suggest going to the Chaosgroup forum and posting your question there, or looking for solution. Try turning off the filter and then re-render that, I have had it where it depends on the filter I use too.
  13. First, probabilistic lights on or off? Then there is other things to check for, but I forgot most of them. I usually turn off probabilistic lights off.
  14. Thanks and that's a good datum to have, I had not thought about that even though it was probably mentioned before. Thanks Nik!
×
×
  • Create New...