Agreed. We have to play the waiting game now. If these are correct for the 32 core, I will also be a little underwhelmed. I just can't see how they can be correct, but I've been wrong before so who knows!
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
So M1 Max with 32 Core GPU matches RTX 3080 Mobile.
- Thread starter iBug2
- Start date
- Sort by reaction score
Given the scaling from 8-16 cores basically doubles the score, but 16-32 doesn't, I wonder if that could mean there is some kind of chiplet / interconnect in play on the M1 Max and we're essentially looking at 2x16 core GPUs instead of 1x32.
People have been saying that the M1 Max die shot was fake, so I guess it's a possibility...
People have been saying that the M1 Max die shot was fake, so I guess it's a possibility...
I've been wondering the exact same. Truthfully, I wouldn't be surprised if this is the case, but we've seen spectacular results from other 32 core (like GFXBench), so I'm curious as to why those would perform well and geekbench wouldn't, with all other things being equal aside from the potential layout.
Even on the die shot they put out it looks like two GPUs with shared logic. So it could still be the case that something goes wrong with the compute scaling. Dunno, confusing.
I had thought the M1 and the M1 Max are both dies. From Mark Gurman the M1 Pro is basically the "chop' (half dies). I dont think there is any chipset with these, that should come with the 2 or 4 x Jade Die. We're all guessing for now!
I noticed that on the Apple MacBook site if you go down to "GPU performance" you can actually toggle between different tasks and they have Redshift there, which is basically a 100% pure test of GPU horsepower.
They claim the 16 core is 2.5x faster than the old machine, and the 32 core is 4x faster. That basically confirms these GeekBench scores for me.
They claim the 16 core is 2.5x faster than the old machine, and the 32 core is 4x faster. That basically confirms these GeekBench scores for me.
What do you mean? Whys that?I noticed that on the Apple MacBook site if you go down to "GPU performance" you can actually toggle between different tasks and they have Redshift there, which is basically a 100% pure test of GPU horsepower.
They claim the 16 core is 2.5x faster than the old machine, and the 32 core is 4x faster. That basically confirms these GeekBench scores for me.
'Steve Moser discovered references to High Power Mode in the macOS Monterey beta. This new setting is the opposite of "Low Power Mode," which aims to decrease system performance to prolong battery life. The new mode will only be available on the 16-inch MacBook Pro with the M1 Max chip, not the 14-inch model or models with the M1 Pro."
Aren’t they comparing performance compared to the previous amd gpu? Not sure much can be extrapolated from that.I noticed that on the Apple MacBook site if you go down to "GPU performance" you can actually toggle between different tasks and they have Redshift there, which is basically a 100% pure test of GPU horsepower.
They claim the 16 core is 2.5x faster than the old machine, and the 32 core is 4x faster. That basically confirms these GeekBench scores for me.
So wait if Redshift is scaling at 4x and these geekbench benchmarks aren't even close to 4x performance, how does this confirm the benchmarks? I'm missing something here. @jmho
Sure, but I’m not convinced you can draw conclusions from one test and another.
I think they are drawing attention to the difference between 2.5x increase and 4x increase.
If the old GPU is 100 "performance units" and the M1 Pro is 2.5x faster, then it would be 250 performance units. If scaling was linear from 16 to 32, then you would 2x250 aka 500, or 5x faster, but we only get 4x faster.
Thus the dream of taking the 8 core M1 GPU and multiplying it by 4 to get ~80k dies, and we have to multiply our 80k target by 4/5 which gives us ~65k Geekbench, which is what we're seeing.
It uses 4 RAM chips on the M1 Max. On the M1 pro it uses only 2and is limited to only 200MB/s. (For small values of limited).
I'm confused by how 5x is target number here. 5x to me reads as 5 times faster. 4x would be four times, which is what we should be expecting. In this case, you're using 4x as if its 3 times faster.
I get everything else you're saying, but wouldn't it be 3x here instead of 4x in your numbers?
In that case though wouldn’t we see a 2.5x increase in geekbench? Instead the 16 core is more or less even with the amd gpu. I don’t think we can draw conclusions.
The GPU they’re comparing to isn’t the M1 baseline. It’s the previous GPU in the MBP.
Yeah, I know that, but that would still give us larger numbers since the previous GPU was better than than the M1, right?
EDIT: Ok yeah, per Apple's site:
"16-inch MacBook Pro with Radeon Pro 5600M and 8GB HBM2"
And the 5600M score:
| AMD Radeon Pro 5600M | 42510 | |
I don't see why this high power mode is not available for the 14 inch.. Use a better cooler then or heatsink, that can't be the issue. Bummer, I ordered a 14 M1 Max with 32 cores, but prolly gonna cancel. 16 inch is way to big for comfort. Ah well initially I wanted a Mac mini M1X, let's go for that.
I get that you’re saying the difference between 16 and 32 cores is smaller than double, but that’s due to a massive jump for 16 core one. That isn’t reflected geekbench which makes me wary of comparing different benchmarks.