I tried a bunch of scenes on my base M1 MBA. The one with only 7 GPU cores and 8GB ram total.
GPU + CPU is pretty much always 2x faster across the board which is a huge upgrade for a free software only update. Still very far from the 2060 PC laptop I have but I’m not expecting miracles from a M1 which is a consumer grade basic system. I tried everything from a logo animation, a R2D2 model and a graveyard scene I did for a live Halloween DJ streaming event. All render 2x faster which is awesome. I look forward to any optimization juice that can be squeezed out of this.
is quite broken. Most rendering tasks lock it up. Some, including enabling denoising on the one project I actually did get to load, cause it to crash the instant viewport denoising is turned on. The Rivet project linked above causes it to lock up in the middle of loading it.
I was running the Dec 30 build getting absolutely amazing GPU speeds, 3-4x faster than CPU - suddenly blender started crashing and even after deleting and reinstalling I can’t get those speeds back. It is eating up all 32gb of RAM available even with Persistent Data turned off within 45 seconds of building the render.
Edit: Applied my Adaptive Subdivision modifiers and went back to Supported mode and renders are back to blazing fast on GPU.
A bug report would be appreciated. This can be done by selecting from the top of Blender Help -> Report a bug then filling out all the relevant information.
I’ve tested the January 9th build, and the render times have gotten worse than the last build I tested; back when the M1 GPU was first being tested in the alpha build.
my times went from 40-50 seconds to 50-60 seconds.
I just tested the December 30th build, because I heard some good things about it.
The results are closer to what I was getting with earlier builds.
One thing I noticed that was greatly improved over any other build I’ve tested so far; was the render window was able to close a ton faster than it did before if you interrupted the render. Before it would take up to a minute or longer just to close the window if you didn’t wait for the render to finish.
It looks like the current build is only using GPU for rendering even if CPU+GPU is enabled. That could explain the increased rendering times compared to builds where CPU+GPU rendering was effective.
I’m not seeing increased rendering times with the BMW benchmark file, but I can confirm that the CPU is not being engaged during rendering, even when the checkbox is checked in Preferences.
I’m having a repeatable crash always when you navigate to the shader editor.
The version of Blender is 3.1.0 alpha and running it on MPB 16* M1 Max with Monterey.
Keep up the good work, the speed with GPU &CPU renders are blazing fast compared to 2.9.
I can’t wait to get to hands on with the production version for the apple silicon.
Crashes and bug reports should be made to the Blender developers website. A bug report can be made by opening Blender, and selecting from the top Help -> Report a bug then filling out all the relevant information to help people reproduce the issue.
Michael Jones - by any chance did updated metal drivers make it into macOS 12.2? And if so, do they perhaps allow for Intel Mac render targets? (particularly AMD graphics cards). I’m not asking if that’s enabled in blender, I can run the dailies… and I am not asking you to rush anything - just if roadblocks have gone away?
AMD GPUs are planned for Metal support in Blender in the future according to the task linked below. It’s not implemented yet as further work needs to be done.