Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

M4 vs. M5 Chip Buyer's Guide: How Much Better Really Is M5?

If you "need" to upgrade a M4 computer to the new M5, you bought the wrong computer, should have gotten a Pro or Max.
For 99+% of people computers nowadays last 3-5 years or even longer, buying year over year doesn't make sense anymore
 
  • Like
Reactions: wyrdness and Gummiwise
FloatingBones said:
Video encoders are "junk" -- unless you need them for your workflow.
GPUs are "junk" for many people.
NPUs are "junk" for people who don't use apps that access them.
Matrix-multiply accelerators are "junk" -- until you start needing matrix manipulation.
Efficiency cores are "junk" -- unless you want to conserve battery life when doing background tasks.

Apple solves many problems with each of their processor architectures. So does Qualcomm. So does ARM. So does everybody! Everything is a blend. If you obsess about single-processor speed, you would be better off inventing a time machine and going back to before 1999 when there were no multi-processors or GPUs. Or you could design and fab your own #!$$ processor. Good luck with that.

It is pointless to grouse about "wasted" real estate on any Apple general processor architecture. The M5 architecture does an amazing diversity of tasks amazingly well, and you have no idea when your needs may shift to another application. Appreciate the brilliance of its design and architecture. If you don't appreciate it, you can always go elsewhere.
Click to expand...
I will go farther than you and say that complaining about "wasted" real estate is ill-informed. The biggest design constraints today are power and heat. Moreover super scalar CPU design (read: single-threaded) hit a wall many years ago. It makes no sense to devote die space to designs which do not show any improvement when you scale them up when there is ample space to solve real problems. We were talking about the inevitability of heterogenous computing 15 years ago (I wrote a class assignment on it!), and now we're here, so just learn why it happened and deal with it.
 
Last edited:
  • Love
Reactions: Populus and FloatingBones
fyzle said:
In the previous gen there was already a 2x speed difference between the slow (1 chip) and fast (2 chip) SSD choices.
Click to expand...
Which is an achievement if it has happened again. SSDs have been around many years and Apple uses good ones. A 2X jump again would be amazing. It could mean over 10,000 MB/s read speeds (10 GB/s). Nice if you can get it. They'd have to be some of the fasted SSDs on the market, but regardless, those speeds are excellent. Gurman in his article called it "slightly increased". wtf.
 
Last edited:
  • Like
Reactions: montuori
MacBH928 said:
as for AI, if you are not running local AI models, is that any benefit to the user?
Click to expand...
While AI and LLMs get all the headlines these days, the underlying neural acceleration hardware boosts a host of other tasks that utilize machine learning routines. Things like the blurred backgrounds in FaceTime calls or copying text out of a photo in the Photos app all use ML and the neural hardware. Third party developers have access to this hardware so even if something isn't billed as AI (although everything today seems to be billed as AI for marketing purposes) it can still get a boost from the hardware improvements.
 
  • Like
Reactions: neuropsychguy, Leon Ze Professional and Populus
TheBeverage said:
So much effort wasted on “AI” junk. No mention of single threaded speeds, which is what the overwhelming majority of tasks rely on, especially the tasks a user buying a base chip is going to actually use it for.
Click to expand...
AI applications are the most compute intensive. For most user tasks current CPUs are plenty fast, it's the special intensive use cases that matter, and AI is the most extreme case. If you want to run a huge model fast, such as running 120b weight ChatGPT, that takes a very fast chip with a lot of graphics/unified RAM. And that's a specific use case that's of a lot of value to a lot of engineers, for example, and that's who buys high end CPUs. It doesn't hurt anyone else, worst case, and more and more the AI features are integrated into regular consumer apps, like photos, email, etc., where AI can be helpful, e.g., recognizing who's in photos so you can find them, correcting grammar, making Siri more intelligent and useful (one can hope), etc.
 
Hank001 said:
If I can make a suggestion; articles like this one would make sense if you were to compare it to the last 4 or 5 versions of the same hardware. Most people aren't going to bother upgrading to each iteration of said device.

Make a table, make it readable, make it crystal clear what the gains are over the years. That would make for some proper journalism, people will be able to make a sound decision whether to upgrade or not, and they will be happy to read in-depth articles on all differences. And the author will be vastly more proud of said work.
Click to expand...
This has been discussed for years and its ignored. Similar to how they only compare the latest iPhones to each other and ignore previous gens until a month or more later, if at all.

I believe its down to the fact that they are using the latest device so don't care about other ones.

Also this article having no real data about what it does, because up to 45% faster includes 0% as a number, makes it all pretty useless anyway. I'm sure the new device is faster, but I'm not sure we can state for a fact what it is until they do real world benchmarks.
 
Leatherm@n said:
I am coming from an M1 Pro. I do photo editing in Lightroom Classic. Is there a significant difference between M4 and M5 for my use case?
Click to expand...
2x to 3x I'm assuming. I'm coming from M1 Max and I'll jump to M5 Max when it comes. I'm expecting huge gains in photo and video editing after 4 generations, at least in many common tasks.
 
  • Like
Reactions: Leon Ze Professional
Can't wait to see Blender results. The up to 30% in graphics, and 45% in Raytracing is HUGE. Apple is doing amazing things with their GPU's and even with the M4 on my Mac Studio, I have seen myself use my Nvidia Windows machines less and less.

While today consumer GPU's for Nvidia is merely chicken-feed for Nvidia, I still think they should be worried if Apple ever decides to scale things up for AI workloads.
 
it looks like most of that GPU boost comes from the increased memory bandwidth, which is 27% more than the M4. Will the M5 Pro and Max get a similar memory bandwidth increase to match, or will it be less impressive of an improvement on them?
 
ilostmynick said:
How is the single threaded CPU performance? This is the most important for performance. I suspect there's a reason why this number isn’t publicly disclosed.
Click to expand...
M5 is about 15% faster in single core than M4. It's almost certainly the fastest single core performance of any currently available CPU.
 
I have a M4 Max MBP coming as my M1 Max is struggling with FCP. I wonder if in some ways the M5 would be better for me. I make YouTube 4k videos but nothing crazy. My M1 is just struggling with 32 GB RAM. I would appreciate any advice. I wish the Max was available now as well. I also want the Thunderbolt 5 ports.
 
  • Like
Reactions: unomas77
Innovati0n said:
Talk about cherrypicking numbers for a debate
Click to expand...

You're not the target market for these machines. M5 is a big deal for people that work in AI

1990's-2000's: apple's target market was graphic designers, publishers, writers
2000's-2010's: apple's target market was video editors and photographers
2020's-: guess who
 
AeroUK said:
I have a M4 Max MBP coming as my M1 Max is struggling with FCP. I wonder if in some ways the M5 would be better for me. I make YouTube 4k videos but nothing crazy. My M1 is just struggling with 32 GB RAM. I would appreciate any advice. I wish the Max was available now as well. I also want the Thunderbolt 5 ports.
Click to expand...
If you can wait for the M5 Max, wait. This thing has pretty serious GPU boosts over the M4 which for your use case would be very beneficial. If my M1 Max MBP didn't get a bucket full of water I would have got the M5 Max Studio
 
AeroUK said:
I have a M4 Max MBP coming as my M1 Max is struggling with FCP. I wonder if in some ways the M5 would be better for me. I make YouTube 4k videos but nothing crazy. My M1 is just struggling with 32 GB RAM. I would appreciate any advice. I wish the Max was available now as well. I also want the Thunderbolt 5 ports.
Click to expand...
I honestly think you could get by with an M5 and still notice improvements, but I feel you’ll benefit from at least a Pro chip. So either getting the M4 max or waiting a few months for an M5 Pro, both seem like a good option. If you can wait a few months for the M5 Pro, I’d wait.
 
real question . should I go for a baseline m5 or the base m4 pro... thats what I really want to know..I already know m4 comes with thunderbolt
 
So the MacBook Pro only has a base chip now? Will the Air also have this base M5 chip? Is it the same chip inside the iPad Pro?

Why do they keep screwing with the naming? Pro should have the Pro chip like it's been for M1-M4, what's wrong with these people.
 
Last edited:
gsprs said:
So the MacBook Pro has a base chip now? Will the Air also have this base M5 chip? Is it the same chip inside the iPad Pro?

Why do they keep screwing with the naming? Pro should have the Pro chip like it's been for M1-M4, what's wrong with these people.
Click to expand...
Because it’s the MBP, for the masses. Also, it is admittedly more powerful than the upcoming M5 MBA, as this MBP has a cooling system that allows the M5 to perform much better than on an Air.

You also get the same display than the M4 Pro MBP, the same body, etc…
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back