i dont have money for pay a cuadro rtx can i purchase a RTX 2080 for maya arnold ? to do this fast ?
Arnold GPU isn't a released product. It's too early to start making purchasing decisions.
Don't get my hopes up by posting questions about graphics cards. Though I have a 16 core Threadripper, which chews through Arnold like a champ!
I'm waiting for the first benchmarks to come out. I'm saving money to go either thread ripper 2990 WX or 2080 Ti.
TR 2990wx will do amazing job in Arnold and will remain all CPU features
After / If AMD will solve RAM bandwidth issue , then CPU rendering is back again
no need to wait GPU renderer at all , you're restricted in VRAM anyway , don't think GPU will come with all good features that Arnold already have .
Sorry guys, peace and love, but I absolutely do not agree with a lot of posters above....
- Arnold GPU render option will be a blast!
- Arnold have to release it's GPU render option ASAP it's simply madness that it's still not out. We already saw how well it can work out (I understand that they are probablly optimizing it for the Tensor cores, but still...)
- I'm planing to buy a threadripper as well, but not for Arnold CPU render, but for it's upcoming GPU render option, and also for zBrush, Houdini and specially that it has 60 PCI lanes AKA we can use a Threadripper system perfectly for a GPU render farm (2 × cost intel system has 24 lanes?)
- The mentioned above out of Vram issue due to GPU renders is the past. The last renderer was Octane that had issues with it, but not any more. All GPU renders can render out of core.
- Octane releasing it's unbiasd, PB, pathtraceing, realtime raytracing, tensor core suporting Octane V4 for FREE up to 2 GPUs and Arnold is still holding back it's GPU render option for 670 euro/ year : (
I love Arnold! It's a brilliant renderer but it's seems a bit like they were sit down after Autodesk bought them out. So go, go Arnold!
And yes! It's a valid question that how fast will be the 2070, 2080 and the 2080 Ti by using it under Arnold GPU render option. How well we will be able to render by using 8 or 11 Vram, how it will scale by using more than one GPU, how it will handle different models.