Johnathan Joy

Pixelised
Definitely going to follow this thread, as I'm thinking about an upgrade and I'm running a 1950X from 6 years ago, it's not bad but as we always "need" faster machines I'm interested to see how the M3 Pro will stack up.

I did a CPU benchmark and an M2 Pro was slightly slower that a 1950X, albeit impressive in itself given it's a laptop. I've also heard that you never hear the fans kick in. So it will be interesting to see where the M3 Pro sits, I'd anticipate a 15% increase judging by the M3 presentation the other day.

With all this in mind, what would it be like on day to day usage, and over a year! The 7950X is about 2.5x faster and probably makes more sense to upgrade as a workstation scenario. But I like the small form factor it has to be said.

Will be interesting to see how and what comes out soon I hope I haven't derailed this thread...

Where did you do the CPU benchmark?

osong
that's why we are asking - to know whether this particular development is happening soon™️ or otherwise

In my opinion, based on how SideFX have consistantly delivered over the years, it's a resonding YES. However all I was saying was to fully deliver every feature of an app the size and complexity of Houdini will take a little time. We just need to be patient. From what I am seeing Houdini on Apple Silicon is already very capable. So with every update going forward I think that it is a given that Houdini on Mac Silicon is only going to get faster with all the features operating and running to comparable levels of Houdini on the Intel/Windows platform, for which Houdini has-been developed to run on for decades. The fact Sidefx have delivered Houdini on Apple Silicon shows they're commited to the platform. I guess the only question we are asking here is when are we going to get XPU taking full advantage of the Apple Silicon chiptset! My guess H20.5 or H21.

Heres a video I found on Houdini on Mac Studio vs an Intel Box for anyone interested.

Is a Mac Studio fast enough for Houdini? [www.youtube.com]

osong

cgshorts
I presume that with each subsequent release of Houdini for Mac Silicon, more code will get optimised to take advantage Apple's new chip set. It must be a fairly large task to bring a program of the size and complexity of Houdini over to a new platform.

unironically it is already on the new platform, and runs natively. we are asking about XPU acceleration with M-chip GPU.

Agreed, Houdini is on a new platform already, but not everything is optimised yet to take full advantage of said new platform or indeed working at all, for example XPU. What I am saying is it will take a little time. There is a difference between having something working and something working and taking full advantage of the system it's running on.