I’ve recently made the switch from Rhino 5 and V-ray to using Rhino 7 and it’s built in rendering engine. We’re getting good results with the rendering engine, but on one of our machines it takes FOREVER to render. The same rendering will take 8 minutes on one computer, and upwards of 55 minutes on another from the same file. The renderings are about 2500x1400 pixels at 300 dpi. I’m trying to find the bottleneck and see what the best way to reduce rendering time is on the one machine. The specs of the computers are as follows:
Computer 1 (8 minute render time)
AMD Ryzen 5 5600X
32 GB ram
Nvidia GTX 1660 Super
Computer 2 (55 minute render time)
Intel i7 8700K
16 GB ram
AMD Radeon RX660
Rhino render exclusively uses CPU correct? The Ryzen is newer I realize, but it shouldn’t be 6 or 7 times faster than the i7, should it? I tried messing around with Power Render and I was able to drop the render times, but the quality was worse than regular Rhino render. Any suggestions for how to optimize render time, or any upgrades that may help in my situation? Would more RAM help? I appreciate your help, I like Rhino Render and don’t want to go back to V-ray, assuming we can get these render times in check.
Your computer 1 has a CUDA-capable GPU, the Nvidia GTX 1660 Super. CUDA hardware will be used by default when found.
Your computer 2 has a AMD RX 6600, if it is selected as render device it will be used assuming the OpenCL kernels required for the rendering process are successfully compiled.
Best is to see under: Tools > Options > Cycles.
You can choose to render on the CPU. When that is selected you’ll be able to set the targeted utilization level from one core to all cores in the same Cycles settings dialog when CPU is chosen as render device.
I don’t know what Power Render is, we don’t provide anything named that…
Addendum: OpenCL rendering is much less performant in Raytraced and Rhino Render in Rhino 7 compared to CUDA rendering.
My mistake, the other product I tried was called ProRender and was available through the package manager.
I have OpenCL selected under cycles for the computer with the AMD card, but when I check the resource manager it looks like my card usage never exceeds 3 or 4 percent while my CPU is being maxed out. Is this normal for Rhino Render and OpenCL?
If I swapped out the card for an Nvidia product (say the same card as the other machine) will I see a decrease in render time?
This leads me to believe that either the OpenCL kernel compiling hasn’t finished yet, or maybe even failed. In either case the CPU will be utilized, since the AMD GPU is not ready for usage.
Ah yes, that will work great with the AMD GPU (the renderer is created by AMD).
Yes. The GTX 1660 Super, or if you can get your hands on newer Nvidia GPUs, will decrease your rendering times. I know that currently the GPU markets are not ideal, but RTX 20xx series or RTX 30xx would be great.