New release 2021.1 seems to be very very laggy on M1. Layout speed seems to have been cut to half with the new processor. Any ideas?
Are you comparing 2021.1 on M1 against 2021.0.1 on M1? There was no work done with regard to M1, so the performance ought to be the same.
Can you give an example file, and describe a test we could do, in order to see the change you’re seeing?
Hi Colin. Thanks for responding. See video attached. Also a problem with some references being higher resolution under raster mode. Google Drive: Sign-in
I have just changed from windows to Mac silicon since I have read all over that those m1 chip were more powerfull than most windows laptop.
Well, yesterday I installed sketchup 2021 and the result is quite disappointing. See the join picture. Not only the model is laggy but when I move the line tool a bug appears and draws lines all over.
Overall what surprises me more is that my old laptop from 2011 with a gtx 780m is running this same sketchup model much better.
Is there some kind of configuration that I am missing?
you could try if decreasing the demand of the OpenGL support by disabling the “Use Fast Feedback” option and decreasing the Anti-Aliasing at “SketchUp > Preferences > OpenGL” helps.
Thank you, I disabled the “Use Fast Feedback” (which I don’t know its use) and it solved the line problem. I let the Anti-Aliasing at x4 which is the maximum allowed in my version at least.
However, I still miss feeling a difference between my old ten years old laptop and this brand new Mac. In the speed of moving and the Anti Aliasing which I think was around x16… Do you think this is normal? Even in my Ipad pro 2018 with sketchup viewer the model moves more smoothly…
Maybe it is just an issue due to the fact that Sketchup hasn’t made yet a native M1 version…
Thanks again for your quick response.
SketchUp is not optimized for the M1 so you will get better performance from Intel based computer.
Ok. That’s what I was thinking. I had read so many things about how great sketchup was working on M1 but it seems that when you work heavy and detailed modeled is is not quite a beast… I will maintain my M1 mac in anycase hopping they finally make a M1 native version!
Thanks for the post.
“decreasing” means lowering the value, not keeping it at the maximum.
An emulation as here Rosetta v2 for emulating an intel CPU will always be slower than running natively on the physical CPU, with a SU 2022 version expected to be available natively for the Arm M# CPUs.