ioDyne uses PCIe tunneling so it can reach 5000MB/s read and write over two TB4 cables. When TB5 launches it won’t even be a problem for people who thinks benchmarks represent real work. Honestly you would be hard pressed to find a workload that cares about sequential read and write speeds. It’s all about random read and writes.
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.
What would the M3 need to make you upgrade from an M1?
- Thread starter AltecX
- Start date
- Sort by reaction score
In decades of buying every type of Mac (except AIOs, which make no sense to me) I can say that I have never built a Mac purchase so poorly as to need to replace it in just two years. My premise is to buy hardware appropriate for 3-5 year life cycles.
Others may plan for shorter/longer life cycles for their own good reasons. Or an unanticipated new app's hardware needs might prematurely age out a Mac, of course. However in my experience short life cycles are most often the result of planning failures, especially buying inadequate RAM for a full 3-5 year life cycle. Constantly in these fora we see folks talking about today's RAM needs, when new purchase planning should be for RAM needs expected 2-3 years from now or later. RAM needs always increase over time.
Apple's Mac OS memory management does a great job and today's Mac SSDs are really fast for swop, but Apple's Unified Memory Architecture (UMA) is orders of magnitude faster. Top MBPs already can opt for up to 96 GB RAM. Mac buyers should expect app and OS engineers to be designing to take advantage of UMA and more of that newly available fast RAM.
Internal (fast) SSD capacity should also be over-specced to ~2x expected needs for various reasons. Spending an extra +$400 for fast internal SSD capacity is usually a sound investment.
Thunderbolt5.
I want my Samsung 980 Pro to run at the same speed at the internal.
Thunderbolt3/4 nerfs it.
I want my Samsung 980 Pro to run at the same speed at the internal.
Thunderbolt3/4 nerfs it.
The only way Apple would/should raise performance of the low end MBAs to equate to today's high end MBPs is if they concurrently raise the performance of MBPs to be substantially higher.
My personal opinion FWIW is that Apple should concentrate on reducing price of the low end MBAs, not on increasing performance. I.e. using whichever M-series chips have best cost/benefit to Apple for the low end MBAs. M1 or M2 or M3 all are already capable of providing adequate performance at the low end.
No. You are asking to defy physics. X-Box storage you describe is relatively very very slow, nothing like Apple's RAM and UMA.
My Windows laptop (Thunderbolt 4) has a pair of WD_Black 4TB NVMEs in there (rated at 7300). I was able to transfer 6TB of data, via an 8TB Sabrent NVME (rated at about 5000) in a Thunderbolt 3 enclosure, in about 50 minutes. Steady too, no drops.
Typically Macs get new OS versions for 6 years or so, but they can get security updates for much longer. E.g. my 2016 MBP topped out at Monterey but it still gets security updates to Monterey.
Often I have skipped entire OS versions anyway. Unfortunately I do want Handoff so I do now need Ventura. That plus the crippling effect of only 16 GB RAM in the 2016 MBP put it at end-of-life for my desktop usages. I will buy a Studio Max immediately if Apple would just put an M2 in it; no need to wait for M3 for my needs.
Last edited:
This will get you going: I can live with a 128 GB SSD; I have about 11 GB of data, total.
That idea does not "defy physics." It's just an engineering problem.
Last edited:
No. What you said was: "...where you can upgrade storage, with the same speed and performance as the internal, with a plug-in module and it all appears as one storage." What you suggest would defy physics because - unlike "the solution similar to X-Box X series... ...with a plug-in module" that you reference - Apple's UMA bakes on the RAM physically very very close to the CPU/GPU.
So it is not just an engineering problem, unless you are suggesting that creating a plug-in module at nanometer scale is the engineering problem. And even if you did so, electrons would still be taking the long way around through the printed circuit, which is much slower than what UMA does. Any plug-in module is simply orders of magnitude too physically large to reach the speeds Apple's UMA can reach. Physics.
Edit: Note that different, slower, levels of memory (L2, L3, whoopee or whatever they choose to call it) might be created, and could be made plug-in modular. Slower than the RAM in Apple's UMA but still faster than swop to SSD. I will be surprised if Apple does not create something like that to help compete with the huge RAM available in big Intel boxes.
Last edited:
Neither of us know what Apple engineers can or can't do, so don't be a google expert and pontificate on here.
Apple engineers cannot break the laws of physics - they can do amazing things but not that.
I agree with your first comment and believe that will be the case — as the M3 SOC is expected to deliver 2 generations of performance and efficiency improvement over M1. Given this, you can paraphrase my wish as “I hope the 2023 MacBook Air includes an SOC configuration that matches the 2021 M1 Max MacBook Pro CPU/GPU specs. This allows the 2023 MacBook Pro to exclusively offer M3 Pro/Max top-end performance.
This wish is not driven by a desire for a lower-cost device — after almost 1.5 years of loving and lugging my 2021 M1 Max 16” 32GB MacBook Pro through one too many airport security lines, I simple want a lighter, less-bulky device with comparable screen size and CPU/GPU performance. I’m happy to pay a premium AND trade away many of the current “Pro” features (Mini-LED, HDMI, CF slot, extra ports) for a lighter, thinner device with better SOC. This could potentially be a higher-margin product for Apple and suspect that I’m not alone in desire and willingness to pay for such a configuration.
That's good to know thanks for the reference, I've been looking for such a fast external storage device. Don't know why I never came across this brand. Always trust people from the movie industry! They make the real deal devices. Looks like a beast.
The iodyne Pro Data is the world's fastest external SSD system
It's a bold claim, but at a possible 5GB/s via combined Thunderbolt ports, the iodyne Pro Data would appear to be seriously fast.
www.redsharknews.com
Review: iodyne Pro Data - a next level SSD RAID by Scott Simmons - ProVideo Coalition
The iodyne Pro Data is quite a beast when it comes to fast data storage for editing and post-production. It isn’t just a fast RAID or a fast SSD: it is a very fast SSD RAID (using NVMe storage) but a bit more than that. Now, SSD RAIDs aren’t anything new. We’ve reviewed SSD RAIDs [...]<p><a...
www.provideocoalition.com
Yes absolutely regarding T5, the other feature I need is Thunderbolt 5, it will finally make the most of speed. I'll wait for Thunderbolt 5 model of this enclosure then, and T5 MBPs.
You spelled Steve Jobs wrong.
If anything, MacOS is quickly becoming more like the phone interface and with starker white flat interface elements just like Windows.
Windows is terrible, by the way.
This has nothing to do with the architecture.
Apple can have its phone development and continue on with it. They need to stop treating the real work OS like a stepchild or trying to turn it into a phone interface.
Apple doesn't guarentee security updates for anything other than the current version. They will backport most to N-1, but that's about it. Basically once you know that the new version won't support hardware (e.g. June last year) you have at most until the October the following year to upgrade without getting into serious risky territory.
We're a long way away from that happening for M1's. Intel support will drop first, and that's likely 2025-2027. After that? Well, the interesting thing about the Mx's is that since Apple controls the microcode, I have a guess that they'll have much longer lifespan than the intel machines did. What drove the rapid turnover there was specific CPU and GPU features, or supporting tech like the T2 chips. With M1's, at least so far, all we've seen are speed bumps, not that kind of breaking capability change.
Long way of saying, I expect M1's to have at least 4 and maybe as many as 7 years left. The exception might be the very first ones (pre-redesigns).
We're a long way away from that happening for M1's. Intel support will drop first, and that's likely 2025-2027. After that? Well, the interesting thing about the Mx's is that since Apple controls the microcode, I have a guess that they'll have much longer lifespan than the intel machines did. What drove the rapid turnover there was specific CPU and GPU features, or supporting tech like the T2 chips. With M1's, at least so far, all we've seen are speed bumps, not that kind of breaking capability change.
Long way of saying, I expect M1's to have at least 4 and maybe as many as 7 years left. The exception might be the very first ones (pre-redesigns).
If they simply hold the price flat, that's like an 8% annual discount these days.
Total disk space, I get (I'm somewhere north of 30 TB now), but how much actually needs to be in my machine vs on external drives? 2TB is just fine for me. Now if I was doing video editing, maybe it'd be different. But that's just for live working space - anything other than current files can easily live offline.
If the M3 delivers a significant performance upgrade and they hold the current price as is it’ll be a huge win. If they raise the price only $100 it’ll still be a win.
I know this thread has gone on 5 pages and has drifted slightly away from the question that the OP proposed, but to get back to OP's original question -- I'm not entirely sure what the M3 would need to make me upgrade. I've got a 10 core/16 core M1 Pro currently and to be honest, it's really doing virtually everything I need. I think, for many people, gone are the days when one powerful computer was needed. Now, I've got a fast iPad for web browsing, a fast Mac for doing things like large PowerPoint decks on two external displays, and a fast phone for mobile basic emails, small apps and such.
My need for a single fast computer is less important for me these days than it used to be in days past. I remember back in the day when I upgraded my PowerMac Dual G5 after one year to the next model dual G5 because I needed the power. Now, the only reason I upgrade is due to changes in form factor or if one of my kids or my wife needs an upgrade. I don't play games any more, so there's few applications I use that requires top of the line power anymore with one specific exception (Osirix MD), but that application is fast enough on the M1 Pro and it's a small percentage of my total use. I gave my 8 month old MacBook Air M1 to my daughter because she needed a new machine and replaced it with a 14" MBP. Otherwise, I'd still likely be using it.
Unless I need to hand down my current machine to someone else, which is always possible, I'll probably keep this one well beyond the M3.
My need for a single fast computer is less important for me these days than it used to be in days past. I remember back in the day when I upgraded my PowerMac Dual G5 after one year to the next model dual G5 because I needed the power. Now, the only reason I upgrade is due to changes in form factor or if one of my kids or my wife needs an upgrade. I don't play games any more, so there's few applications I use that requires top of the line power anymore with one specific exception (Osirix MD), but that application is fast enough on the M1 Pro and it's a small percentage of my total use. I gave my 8 month old MacBook Air M1 to my daughter because she needed a new machine and replaced it with a 14" MBP. Otherwise, I'd still likely be using it.
Unless I need to hand down my current machine to someone else, which is always possible, I'll probably keep this one well beyond the M3.
Have you tried to use Windows lately? From my perspective, as a 30+year Mac user, Windows has become an opaque nightmare. It is simply anathema to my workflows, much worse than macOS.
Yes, Apple has tightened the UI, and done some things I really dislike. But the doors to doing what you need to do are still there, just obscured a little by arrases. If they kill the ability to use Go to Folder to get to / or ~/Library, or disallow those directories in the sidebar, that will piss me off. But most of the annoying phone-like features can strangled or stomped on in Prefs.
They are making an OS that is easy to use, and maybe since 87.4% of computer users have ditched the big hot box in the study for the phone in hand, their increasing moves toward simplifying a niche-use product are ill-considered. But, as long as they keep the nerd doors and the prefs settings that make me happy, the other stuff is fine with me.