Logically unless supply chain issues interfere we will see new M2 MBPs in October. Hopefully my 2016 MBP keeps working until then and I can choose between even better last-generation M1 deals versus the substantially more refined M2.
And it's $1699 for Totaltech membersBest Buy also has the $1749 price on the base 14" with immediate availability.
Often some substantial deals show up on one-off boxes from various locations closing out their inventory after the new generation is on the street. In 2017 I got a 2016 MB with 2 TB SSD drive for about the previous cost of the same box with base level storage; savings of about $1k. There is no guarantee one will find such a deal of course.Hmm. What kind of discounts do you think we can expect on the M1 version when the M2 version is released?
$250-$300 off seems like just about the maximum that anybody ever discounts Apple laptops.
Often some substantial deals show up on one-off boxes from various locations closing out their inventory after the new generation is on the street. In 2017 I got a 2016 MB with 2 TB SSD drive for about the previous cost of the same box with 512 MB SSD; savings of about $1k. There is no guarantee one will find such a deal of course.
However my guess is that after the M2 MBPs are out Apple will still be wanting to sell M1 MBPs because at that point M1 chip production yields will be very profitable while M2 yields still in the low profitability range. So I expect M1 MBP deals after the M2 MBPs are out. Just my $0.02, and even with a deal on M1 the M2s may be a better choice, depending on one's finances/needs.
We are talking about yields, not about process. Chip yields always start out poor (like ~10%) and usually end up very good (like better than 80%). So in Q4 M2 yields will most likely not be the same as M1 yields that have been running for many more months. Mine is a very generic comment, however, as I have zero knowledge about Apple's specific manufacturing.I'm not buying the rumors that M2 Pros will be made on a different process vs. the M2 non-Pros. (Has that ever happened within an Apple chip model line?)
So I would expect the yields of M2 Pros to be the same as any other chip Apple is currently making.
I decided MBP 14" base over m2 MBA because when spec'd similarly the price difference isn' too far off and what you get with the MBP is worth it (at least to me: 120 hz, 4k, HDR, speakers sound incredible for a laptop, and the weight difference isn't a big deal to me coming from a 12" MacBook).SO WHAT is one to do? get the on sale powerful M1 MBP 14" that is a bit heavier... (so this machine is basically better than the M2 AIR, 16gb, 512) by the mini LED, anything else?
-I have the Air m2 now and its great, keyboard seems shallow? and the base seems flimsy... sure it is just me! LOL
So do you keep a M2 AIR because its the latest tech? or go for the deal, knowing something new will be coming in the fall?
ahh nice big change from a 12!I decided MBP 14" base over m2 MBA because when spec'd similarly the price difference isn' too far off and what you get with the MBP is worth it (at least to me: 120 hz, 4k, HDR, speakers sound incredible for a laptop, and the weight difference isn't a big deal to me coming from a 12" MacBook).
I don't need to worry about the M2 possibility because my uses won't warrant it. For my usage the logical choice was the MBA but I wanted everything mentioned in the "( )" above.
Midnight color almost sold me though...
how does this work exactly? I saw that as well and it seems too good to be true, lolAnd it's $1699 for Totaltech members
We are talking about yields, not about process. Chip yields always start out poor (like ~10%) and usually end up very good (like better than 80%). So in Q4 M2 yields will most likely not be the same as M1 yields that have been running for many more months. Mine is a very generic comment, however, as I have zero knowledge about Apple's specific manufacturing.
M2 and M1 are different chips by definition, they are different sizes physically in public photos, and they are almost assuredly made on physically different fabrication lines (because at least in the beginning M1 chips are still being made concurrently with M2). A new chip line gets tweaked over time to improve yields. It will not run at the same yield on week #40 as it did on week #1.Yields start out low when a new manufacturing (fabrication) process comes online. Each step of the process must be fine-tuned for yields to be good.
If you're making a new chip using an established process, there's no reason for yields to be low. Why would they be? And what would you even do about it if they were?
M2 and M1 are different chips by definition, and they are almost assuredly made on physically different fabrication lines (because at least in the beginning M1 chips are still being made concurrently with M2). A new chip line gets tweaked over time to improve yields. It will not run at the same yield on week #40 as it did on week #1.
No. My expectation is that Apple or Apple's vendors make enough M1s or M2s to efficiently dedicate a line to a single such chip at a time. IIRC, I have seen suggestions to that effect in the press. But - again - I do not really know.Uh, is your theory that a whole manufacturing line can only make one particular chip?
No. My expectation is that Apple or Apple's vendors make enough M1s or M2s to efficiently dedicate a line to a single such chip at a time. IIRC, I have seen suggestions to that effect in the press. But - again - I do not really know.
I kinda doubt that, because my understanding is that there is hella more to getting each usable chip than simply swapping masks, even if that one step is as simplistic as you state. However, I lack the expertise to argue the point.The way fabs work is that they can process X many silicon wafers per day. Chip designs are etched onto the wafers by blasting ultraviolet light through a mask and onto the wafer. To make a new chip design, you just make a new mask, and you can swap masks whenever you want. The rest of the fabrication process is the same. So are the yields.
I kinda doubt that, because my understanding is that there is hella more to getting each usable chip than simply swapping masks, even if that one step is as simplistic as you state. However, I lack the expertise to argue the point.
And individual chips need to be cut from finished wafers. I can see all kinds of ways that might be tweakable, including repeatedly evolving the mask design. Unfortunately we will not actually hear from inside the fabs thanks to NDAs.What else would be involved?
Detail: a design might be tweaked (and new masks made) to improve yields, because some component of the design wasn't able to be etched very well. But in this day and age of automated design tools I would be surprised if that happens much anymore.
And individual chips need to be cut from finished wafers. I can see all kinds of ways that might be tweakable, including repeatedly evolving the mask design. Unfortunately we will not actually hear from inside the fabs thanks to NDAs.
I don't think this is true. Following the masking process, the different parts of the chip need to be interconnected with copper wires, and so many connections are needed such that, according to https://en.wikipedia.org/wiki/Semiconductor_device_fabrication, as many as 11 metal layers are used.The way fabs work is that they can process X many silicon wafers per day. Chip designs are etched onto the wafers by blasting ultraviolet light through a mask and onto the wafer. To make a new chip design, you just make a new mask, and you can swap masks whenever you want. The rest of the fabrication process is the same. So are the yields.
I don't think this is true. Following the masking process, the different parts of the chip need to be interconnected with copper wires, and so many connections are needed such that, according to https://en.wikipedia.org/wiki/Semiconductor_device_fabrication, as many as 11 metal layers are used.
A different microarchitecture likely requires a different design for each of the metal layers. And creating these metal layers can take over 300 processing steps. Thus I expect if you switch to a different microarchitecture, there's a lot of very careful setup work that needs to be done, followed by extensive QA/QC tests to ensure it's dialed in before you can resume production. My guess is that this is a non-trivial process. I have no idea, however, how long this typically takes.