M4 Mac Announcements

It is semi-tangential, but ISTR hearing that N3B had not great yields (but still way better than the Samsung process). I am wondering, if the "relaxed standard" of N3E improves yields, would that give Apple more headroom to map out different chip patterns for different iterations of M4 (the way M3 Pro was a completely different layout from the Max, only moreso)? Are they inclined to order an Ultra that is something other than Max+Max at this point?
 
Apple still has the M2 Max and M2 Ultra studio and Mac Pro in their lineup. I really gotta ask myself why anybody would buy one of those machines after today’s Mac mini announcement. No idea why they are sticking around.
 
Apple still has the M2 Max and M2 Ultra studio and Mac Pro in their lineup. I really gotta ask myself why anybody would buy one of those machines after today’s Mac mini announcement. No idea why they are sticking around.
Apple always keeps the old model around until either the line is discontinued entirely or replaced by the new version.

It is semi-tangential, but ISTR hearing that N3B had not great yields (but still way better than the Samsung process). I am wondering, if the "relaxed standard" of N3E improves yields, would that give Apple more headroom to map out different chip patterns for different iterations of M4 (the way M3 Pro was a completely different layout from the Max, only moreso)? Are they inclined to order an Ultra that is something other than Max+Max at this point?

That is indeed the rumor - supposedly there will be a bespoke “Hidra” die making up the next set of desktop chips but whether it will replace the ultra or be a new or even exist is obviously not confirmed.
 
ISTR hearing that N3B had not great yields
As best as I can tell, this was just a lot of hotheads in internet fora doing a chicken little. There was never any evidence of this where you'd expect to see it (TSMC financials and disclosures).

Still, it's more expensive than N3E. So I'd still love to know why Intel picked it for Lunar Lake. Was N3E simply not available because Apple had sucked up too much of their capacity?
 
M1 Pro: 8+2
M2 Pro: 8+4
M3 Pro: 6+6
M4 Pro: 10+4

I've said it before, after having worked with the M1 for a while before the M1 Pro was announced, I felt that having only 2 E-cores wasn't the best design. One of the clear strengths of the Apple Silicon and macOS combination is that background tasks can run on the E-cores, while the P-cores are waiting to process the user's demands as quickly as possible.
This was corrected in the M2 Pro.
The M3 Pro is the perfect MacBook Pro processor, but I'm not too sure if I would want it in a Mac mini Pro or not. Apple skipped straight to the M4 anyway...
As a lot of people commented, the M4 Pro is a whopping upgrade.

Now I'm wondering about two things:
1. Will there be a MacBook Pro with an M4 Pro or not? Not that I'm interested in getting one, but I think the 6+6 in the M3 Pro is the better combination for a laptop.
2. Will Apple go back to the original scheme that the Max is just a Pro with additional GPU cores, since the M4 Pro already is a beast in the CPU core department?

For my CPU needs the M4 Pro would be more than enough. The only reason to go for the M4 Max would be the number of GPU cores.
 
It is semi-tangential, but ISTR hearing that N3B had not great yields (but still way better than the Samsung process). I am wondering, if the "relaxed standard" of N3E improves yields, would that give Apple more headroom to map out different chip patterns for different iterations of M4 (the way M3 Pro was a completely different layout from the Max, only moreso)? Are they inclined to order an Ultra that is something other than Max+Max at this point?
design work happens before Apple knows what the yields will be.

As a designer, I’d rather work with N3B than N3E. Of course, as a businessman, I’d rather work with N3E than N3B.
 
As best as I can tell, this was just a lot of hotheads in internet fora doing a chicken little. There was never any evidence of this where you'd expect to see it (TSMC financials and disclosures).
I agree that it was overblown but to be fair I think there was a little from the TSMC side that the node wasn’t as good as they’d hoped - eg renaming the node from N3 to N3B and declining to base any further iterations off it but rather using N3E as the base.
Still, it's more expensive than N3E. So I'd still love to know why Intel picked it for Lunar Lake. Was N3E simply not available because Apple had sucked up too much of their capacity?
Almost certainly + Qualcomm too, after all N3E Snapdragons for mobile are coming now too. Intel was probably later than the rest for putting their order in. I doubt there would’ve been sufficient volume at N3E for Intel.
I've said it before, after having worked with the M1 for a while before the M1 Pro was announced, I felt that having only 2 E-cores wasn't the best design. One of the clear strengths of the Apple Silicon and macOS combination is that background tasks can run on the E-cores, while the P-cores are waiting to process the user's demands as quickly as possible.
This was corrected in the M2 Pro.
The M3 Pro is the perfect MacBook Pro processor, but I'm not too sure if I would want it in a Mac mini Pro or not. Apple skipped straight to the M4 anyway...
As a lot of people commented, the M4 Pro is a whopping upgrade.

Now I'm wondering about two things:
1. Will there be a MacBook Pro with an M4 Pro or not? Not that I'm interested in getting one, but I think the 6+6 in the M3 Pro is the better combination for a laptop.
2. Will Apple go back to the original scheme that the Max is just a Pro with additional GPU cores, since the M4 Pro already is a beast in the CPU core department?

For my CPU needs the M4 Pro would be more than enough. The only reason to go for the M4 Max would be the number of GPU cores.
I didn’t choose it in the end, but I agree. I thought the M3 Pro was a really interesting design direction.
 
Last edited:
The M4 Mac Mini is so powerful yet so cheap I'm seriously considering one just to tinker around, it's perfect for all kinds of projects.
Same. E.g., it's small enough to fit inside here:

1730278097453.png
 
The M4 Mac Mini is so powerful yet so cheap I'm seriously considering one just to tinker around, it's perfect for all kinds of projects.
Same here! 4 would allow a fully connected graph

Edit: and 4 would also make a great/odd t1000 pic
 
Just had a thought. My company only hands out laptops because they don’t want the computers to live in the office and expects us to bring the laptops home after work. In part for work from home and in part to deter break ins at the office.
I prefer desktops.

This Mac mini could fit in my jacket pocket. With nice displays both at work and home maybe there’s something here to discuss with the boss. It’d be lighter than the 16” M1 Pro I log around now at least.
 
Will Apple go back to the original scheme that the Max is just a Pro with additional GPU cores, since the M4 Pro already is a beast in the CPU core department?

I think that’s unlikely. The M4 Max should be faster than M3 Max or customers will get upset. We have a massive performance improvement in M4 Pro. I’d expect a similar improvement from Max. So either 15 or even 18 cores, depending on what they go with.
 
I think that’s unlikely. The M4 Max should be faster than M3 Max or customers will get upset. We have a massive performance improvement in M4 Pro. I’d expect a similar improvement from Max. So either 15 or even 18 cores, depending on what they go with.
If so, I’m worried that the 14” will have difficulty handling 15-18 P-cores thermally or at least quietly.
 
I think that’s unlikely. The M4 Max should be faster than M3 Max or customers will get upset. We have a massive performance improvement in M4 Pro. I’d expect a similar improvement from Max. So either 15 or even 18 cores, depending on what they go with.
I don’t see why, if the cluster size is 6 and the pro has 1 or 2 fused off per cluster, that 12 P cores wouldn’t be sufficient, the per core increases alone will bring a nice uplift and will also nicely balance the needs of battery life and performance.

I’ve always thought the M3 pro was a bad design, it was just too close to the base M3 especially in binned form, I think the M4 pro with 10 P goes goes too far in the other direction.

I’d love a lineup with
M4 4p-6e
M4 Pro 8p-8e
M4 Max 12p-8e
 
I think they're doing it right. Each of these SOCs has the core counts aligned to the chassis Apple intends them to be used in.
 
12 p cores. We were wrong.

Perhaps they will diverge the laptop/desktop lines? Wouldn’t be terrible.

Huh, so Gurman was right? Seems like M4 Pro is a chop of a M4 Max with two P-cores disabled after all. It makes sense economically I suppose. Still, I was hoping we would see Apple challenging desktop CPUs with this.
 
Huh, so Gurman was right? Seems like M4 Pro is a chop of a M4 Max with two P-cores disabled after all. It makes sense economically I suppose. Still, I was hoping we would see Apple challenging desktop CPUs with this.
Yes. I feel the Max is the least impressive update of the week, relatively speaking. The Pro feels like the sweet spot.
 
Yes. I feel the Max is the least impressive update of the week, relatively speaking. The Pro feels like the sweet spot.

I could see some folks opting for the Max for the extra GPU oomph. It is a pretty pricy jump for the extra 12 GPU cores on the low end M4 Max though, but it does get you another 50% or so.

The one apprehension I have dropping to the Pro is that I will lose some performance on the GPU. And not super jazzed having to CTO to get more than 24GB RAM.
 
Back
Top