Apple announce new 14" & 16" MBP with M1Pro and M1Max
-
- KVRist
- 313 posts since 20 Feb, 2005
Sidetrack here, sorry, but thought it worth mentioning for the fans going hard on an older macbook Pro: I opened it up and removed a LOT of dust from the cooling ducts and fans. It helped!cryophonik wrote: ↑Tue Oct 26, 2021 11:14 pmI'm starting to encounter the same thing with my fans and I'm getting some of the infamous "battery bulge", so the lid doesn't lay quite flat when I close it. Thanks for the feedback!
The battery bulge would require a battery swap, of course.
- KVRAF
- 1871 posts since 30 Mar, 2008 from MN, USA
Who, Apple? The chart shows the M1 Pro/Max being 1.7X faster than the M1 in the Pro 13", in multi-core. And that makes perfect sense based on the core count differences.
CLAP Software Database: https://clapdb.tech. KVR Discussion Topic.
-
- KVRist
- 313 posts since 20 Feb, 2005
The early benchmarks here indicate M1 Pro and Max performance at around 1.6--1.7x of M1 on a Logic benchmark (in the comments) https://music-prod.com/logic-pro-x-benchmarks/
A bit unclear as there are multiple benchmarks and people's reports on m1 vary quite a bit
A bit unclear as there are multiple benchmarks and people's reports on m1 vary quite a bit
- Banned
- 11467 posts since 4 Jan, 2017 from Warsaw, Poland
Yeah, sorry. I deleted the post after I've watched the video. But since then few new ones were posted, showing that multicore performance of regular M1 is around 7,5k, whereas 10-core M1 Pro is 12,5 (3 videos confirmed that) which is less than 2x I was expecting. So I guess the 2 efficiency cores aren't that weak as I thought (13'' has 4x eff + 4x perf, 14'' has 2x eff + 8x perf).
- KVRAF
- 1871 posts since 30 Mar, 2008 from MN, USA
el davo wrote: ↑Wed Oct 27, 2021 4:46 pm The early benchmarks here indicate M1 Pro and Max performance at around 1.6--1.7x of M1 on a Logic benchmark (in the comments) https://music-prod.com/logic-pro-x-benchmarks/
A bit unclear as there are multiple benchmarks and people's reports on m1 vary quite a bit
Raw multi-core performance metrics only go so far. So much of real-world performance relies on I/O, smart threading, etc., that the raw numbers are only a general guideline.
CLAP Software Database: https://clapdb.tech. KVR Discussion Topic.
- KVRist
- 426 posts since 30 Jan, 2015
Received my new 16.2 MacBook Pro M1 Pro chip with 16 RAM. First impressions seem quite odd. I am coming from an i5 iMac 2012. Therefore, I expected massive performance boosts.
In Ableton an empty session is taking 6% CPU and a small session of 5 tracks with only stock Ableton effects with simpler on midi tracks is putting the CPU at 20%. This compares with 4% on my i5.
Now obviously this is only what the CPU meter is telling me in Ableton.
Either my new shiny MacBook is very poor or the CPU meter is not telling me the full story?
In Ableton an empty session is taking 6% CPU and a small session of 5 tracks with only stock Ableton effects with simpler on midi tracks is putting the CPU at 20%. This compares with 4% on my i5.
Now obviously this is only what the CPU meter is telling me in Ableton.
Either my new shiny MacBook is very poor or the CPU meter is not telling me the full story?
-
- KVRAF
- 4751 posts since 22 Nov, 2012
It's not there yet, but they are getting there. They have plans for doubling the current processor, but the corporate investors want to keep selling you shit while they get there. Going back to the old keyboard and current connectors was YUGE from a hardware standpoint. Apple has been lacking in the hardware side of things since Jobs. I'm just thinking that by the time they release a finished processor, they will be switching back to intel, or moving on to quantum... so... not sure i will jump, but the software side is destroying windows for creatives rn.
-
- KVRist
- 313 posts since 20 Feb, 2005
Sure. The benchmarks I linked to are track counts people get running specific template Logic Pro X projects with defined settings though. Not as heavy on IO as a real world project, but closer to what I am looking for than Cinebench.teilo wrote: ↑Wed Oct 27, 2021 8:03 pmel davo wrote: ↑Wed Oct 27, 2021 4:46 pm The early benchmarks here indicate M1 Pro and Max performance at around 1.6--1.7x of M1 on a Logic benchmark (in the comments) https://music-prod.com/logic-pro-x-benchmarks/
A bit unclear as there are multiple benchmarks and people's reports on m1 vary quite a bit
Raw multi-core performance metrics only go so far. So much of real-world performance relies on I/O, smart threading, etc., that the raw numbers are only a general guideline.
-
- KVRer
- 24 posts since 6 Aug, 2021
I suspect it's the latter. Since the Apple Silicon CPUs are a different architecture from Intel, there are most likely optimizations that are not yet in place. Even though software can be recompiled to be native to the new architecture, there are still optimizations to make things even more performant which will take some time.
Bottom line, 20% usage on one CPU cannot directly compare to 20% usage on another CPU, even within the same general architectures.
I would try generating a project that really taxes your old system, then try the same project on your new MBP to see how it compares. I would also keep in mind that things will only get better with time as more software becomes optimized for the Apple Silicon architectures.
- KVRAF
- 2026 posts since 15 Mar, 2002 from Seattle, WA - USA
Which version of Live, the official v11.0.11 which runs under Rosetta or the M1 native 11.1b5 beta? Also, what sample buffer size? With M1 I actually get lower CPU usage with smaller buffer sizes.Rickskii wrote: ↑Wed Oct 27, 2021 8:41 pm Received my new 16.2 MacBook Pro M1 Pro chip with 16 RAM. First impressions seem quite odd. I am coming from an i5 iMac 2012. Therefore, I expected massive performance boosts.
In Ableton an empty session is taking 6% CPU and a small session of 5 tracks with only stock Ableton effects with simpler on midi tracks is putting the CPU at 20%. This compares with 4% on my i5.
Now obviously this is only what the CPU meter is telling me in Ableton.
Either my new shiny MacBook is very poor or the CPU meter is not telling me the full story?
-
- KVRian
- 912 posts since 18 Feb, 2004
Dasheesh wrote: ↑Wed Oct 27, 2021 9:06 pm It's not there yet, but they are getting there. They have plans for doubling the current processor, but the corporate investors want to keep selling you shit while they get there. Going back to the old keyboard and current connectors was YUGE from a hardware standpoint. Apple has been lacking in the hardware side of things since Jobs. I'm just thinking that by the time they release a finished processor, they will be switching back to intel, or moving on to quantum... so... not sure i will jump, but the software side is destroying windows for creatives rn.
- KVRist
- 426 posts since 30 Jan, 2015
I started with the official as you put it and that was way worse. Now using the beta 11.1b5. I am working at 48 and 512 samples.Tronam wrote: ↑Thu Oct 28, 2021 12:12 amWhich version of Live, the official v11.0.11 which runs under Rosetta or the M1 native 11.1b5 beta? Also, what sample buffer size? With M1 I actually get lower CPU usage with smaller buffer sizes.Rickskii wrote: ↑Wed Oct 27, 2021 8:41 pm Received my new 16.2 MacBook Pro M1 Pro chip with 16 RAM. First impressions seem quite odd. I am coming from an i5 iMac 2012. Therefore, I expected massive performance boosts.
In Ableton an empty session is taking 6% CPU and a small session of 5 tracks with only stock Ableton effects with simpler on midi tracks is putting the CPU at 20%. This compares with 4% on my i5.
Now obviously this is only what the CPU meter is telling me in Ableton.
Either my new shiny MacBook is very poor or the CPU meter is not telling me the full story?
- KVRist
- 426 posts since 30 Jan, 2015
Thanks for your detailed answer, it has relaxed me a little. I will try as you say and report back. Still in the process of setting up really.waxtrax wrote: ↑Wed Oct 27, 2021 11:44 pmI suspect it's the latter. Since the Apple Silicon CPUs are a different architecture from Intel, there are most likely optimizations that are not yet in place. Even though software can be recompiled to be native to the new architecture, there are still optimizations to make things even more performant which will take some time.
Bottom line, 20% usage on one CPU cannot directly compare to 20% usage on another CPU, even within the same general architectures.
I would try generating a project that really taxes your old system, then try the same project on your new MBP to see how it compares. I would also keep in mind that things will only get better with time as more software becomes optimized for the Apple Silicon architectures.