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.
M3 Pro Chip Barely Faster Than M2 Pro in Unverified Benchmark Result
- Thread starter MacRumors
- Start date
- Sort by reaction score
To expand on the historical context: IBM and Motorola were interested in that alliance because they thought they could compete with "Wintel". They wanted to — and briefly tried to — make desktops and laptops that ran PowerPC. But, it turned out that Windows + x86 was (at the time) too strong for a different architecture to compete. This was the early 1990s, so it was basically the tail end of "IBM still thinks it can do PCs". So, both of them eventually gave up on that dream, and IBM focused PowerPC (nowadays essentially the same as POWER) on server stuff, and some consoles; Motorola focused PowerPC on embedded stuff and eventually spun it off altogether to Freescale (now NXP), who ended up abandoning it in favor of ARM.
If AIM's position at the time had been stronger, and Intel/Microsoft's been weaker, things might've worked out differently. For example, if IBM ThinkPads with PowerPC had been a moderate success, they might've been more interested to continue making efficient cores.
I think neither is a good deal in 2023. You're keeping that thing for three to five years, and at that point, macOS Death Valley will make 8 GiB feel very limited.
Well, no.
Only the reviewers have them, and they're embargoed, probably until Monday.
Can’t really argue with that. 512 should be minimum at this price point.
Not at all. Qualcomm's Nuvia-based SoC is due out next summer, with 12 p-cores @80W it hit 15,000 on GB. So the M3 Pro with 6p/6e matches that performance, with a much lower TDP today, not 8 months from now. So the loss of those Nuvia engineers has not slowed Apple down. More than likely Apple has refined the CPU design to the point that the only gains come from the fab process. At least until they make some significant ISA changes.
I'd argue that the kind of people buying the base model are also likely to be the ones to keep that thing for 8+ years.
Range is relative to the product line, i.e., other products in the line, not the price of other computers on the market. What you might consider "expensive" someone else might consider reasonable. Especially someone who has used Macs before and gotten many, many years of use out of them.
Yes an m3 max performing as fast as an m2 ultra is “minimal incremental improvements”
Looking at it another way...
As a happy owner of an M1 Air with 8GB memory, it would now seem likely that macOS is going to have to be capable of running on 8GB machines for a bit longer than if the base had just been upped to 12 or 16GB... 🤷♀️ maybe?
(Obviously I like to see progress - whilst being a little self centred! 🤣)
I recently got my current phone three months ago. My previous phone was five years old. I got my current iPad a year ago. The previous iPad was five years old. I plan on getting a new computer soon given my current computer is twelve years old.
The thing is that Apple has tons of market research into what actual real users who aren’t tech geeks use. We don’t know exactly what that market research says, but if it said that nobody wants 8/256, Apple would immediately bump their minimum specs to 12/512 or 16/512. That they don’t tells you a lot of people are fine with 8/256. I know several people who get along fine with that much and still have room left over, including my wife who spent most of her career writing documents. If your drive is full of Word or Excel documents, you don’t need a whole lot. If your drive is full of 8K ProRes footage, 4TB might not be enough, but no sane company’s going to set their base models to that. Remember the base configurations are for the least needy users of that tier. This is why the M3 Max starts at 36GB of RAM and not 8GB. The M3 Max’s target audience can’t get by on 8GB, but the M3 users can.
You simply cannot base the entire market on your own needs. Me, I need a minimum of 1TB, but I’m not yelling at Apple to make 1TB the minimum because I know I’m not the typical user. Neither is anyone else in this forum.
No - Intel's problem is that they have been "too big to fail" and able to succeed through politics rather than innovation ever since IBM "made" them by picking their kludgey stopgap
Even the current x86-64 instruction set was designed by AMD, not Intel.
The first ARM in 1987 wiped the floor with the contemporary Intel 80286 performance-wise - but it didn't run Windows (except under emulation), so it went nowhere.
Intel's Unique Selling Point is legacy compatibility, and this is now being undermined by a perfect storm of industry changes: There's Intel's almost total failure in the emerging mobile market, which is now dominated by ARM. The rise of Linux and web-based tech with its roots in Unix/Linux - and its culture of hardware-agnostic software design and open source (there's some advantage to running Linux on x86 but most of the key Unix/Linux software packages have supported ARM and other architectures since forever) - is giving Intel serious competition in the server space where power consumption is also a big issue. Apple - even if they remain the underdog in the personal computer market - have demonstrated to the world that you can make high-performance laptops and low-end desktop workstations without x86, and Microsoft seem to be having a serious attempt at Windows on ARM this time. Then, in general as computers have got more powerful, more and more software is written in portable, high-level code and uses operating system frameworks for things like graphics, multithreading, vector processing and neural networks, rather than direct hardware access. An increasing amount of stuff is written in scripting languages like Python or Javascript Even modern Windows ships applications as Common Language Runtime bytecode rather than x86 or ARM binaries, Android apps typically ship as bytecode and even the Apple App store has the capability to distribute "compile on delivery" bytecode. Apple have also demonstrated - with Rosetta 2 - how effectively x86 binaries can be translated to ARM once you've kicked out all of the legacy stuff. I think Windows on ARM's main problem is with bits of legacy Win 16, Win 32 etc. binaries still floating around.
Put simply, Intel's business model relies on people needing to run legacy code, and people ain't writing legacy code any more - so the clock is ticking.
The one area where Apple Silicon has failed to convince is in the high-end personal workstation market - the new Mac Pro satisfies a very limited niche of people who need high PCIe bandwidth but not discrete GPUs. If you really needed the 2019 Mac Pro an x86 tower is probably still the tool for the job. Trouble for Intel there is that the only people buying Intel are the dogmatic "workstations need Xeon because they just do" crowd and AMD are slaughtering them on price/performance. Plus, the whole idea of a high-powered personal workstation isn't for the ages, as the industry shifts to cloud computing (and NVIDIA have some nice ARM-based datacentre-grade iron such as Grace/Hopper to show you).
Whatever you do with x86 tech, it is always going to be carrying around the extra weight of supporting the complex x86 instruction set(s) that more modern RISC-like ISAs don't need. Intel's future is probably to make ARM, RISC-V or some other ISA - and rely on Rosetta-esque translation for x86 support. Trouble is, then, they're going to have to compete with multiple competitors on technical merit, which they haven't needed to do for 40 years.
While a part of this is down to Apple shaking up the industry from time to time, it certainly isn't about what Apple released recently - it goes back as far as the Newton (when they invested money in ARM) plus of course the iPhone's role in promoting the modern mobile scene.
Last edited:
I mean… was anyone with an M2 Pro really looking to jump at an M3 Pro? I understand why Apple compared these new chips to Intel and M1.
may not be a big jump from M2, but folks with older chips will be bigger differences. 🤷🏾♂️
may not be a big jump from M2, but folks with older chips will be bigger differences. 🤷🏾♂️
Love that macOs name!!! Haha!
Disagree though. No one but geeks and nerds give a crap about specs. Most young people use their phones as their only computing device. Apple's base systems are made for these people when they want to move "up". They're not going to need much more than what their phone already provides.
I've had an M1 mini with 8/256 for almost 3 years now, and haven't had any problems, and I'm not an average user; programming hobbyist using Xcode, drawing floor plans, designing furniture, etc.
Well, it tells you two things: that there are people who get by with 8/256, and that there are also people who will begrudgingly give Apple another $400 to upgrade each of those.
Sure. But the question should always also be: "will this still work for me four years from now". Because you're going to want to keep that machine for a while.
Neither am I, although you'd be surprised how cheap a 1 TB SSD is today.
Don't sugar coat it like that, kid, tell 'em straight.
I'm gonna wait a couple of years, save up my money and buy an M5. The BMW, not the Mac. This M1 MacBook Air I'm using right this minute is way more powerful than I need. My 2013 Volvo S60, not so much.
The Air won’t get the Pro chip and won’t need to as the base M3 beats the M1 Pro chip and is near equal to the M2 Pro.
This will likely be Apple’s practice going forward: the iMac and MacBook Airs will get the base chips (even as they continue to get more powerful) while the Pro and Max chips will be for the MacBook Pros and Mac Studio.
Apple isn't "limiting" anything. They've made a machine that is faster and more efficient, at the same base price point, than the previous. In the end, the overall performance counts, not the design designs made to get there.
Exactly. Most users don't care about the specs, they just want a machine that does what they need at a price and performance level they feel is a good value.
I disagree. Unless yo are saving large amounts of videos or massive number of photos, most users never get near filling a 256 GB drive.
Anecdotally, I have 20+ years of client work on my drive, and the only way I ever need more than 256 is when I decide to take over 50 ripped videos with me on a trip (most of which never get watched anyway.). An external drive would work just fine in such a scenario.
M5 lol. Try a panamera. Porsche >>> everything