Follow along with the video below to see how to install our site as a web app on your home screen.
Note: This feature may not be available in some browsers.
It's enough that I'm going to upgrade from the M2 Max.That only 4 p-core the 10 core model I think, seem like Apple really did hit an homerun for this gen.
When you put it like that then yea. Except Indigobench, all the Kvazaar tests, and Apache server which the M4 won but by a hair. Keep in mind that the M4 did come dead last in most tests with the FLAC audio coding as the only win for the M4. In some tests the M4 was extremely far behind compared to other chips. The Ryzen AI 9 HX 370 didn't exactly do better, but it's also a laptop chip that's trying to compete against desktop chips, just like the M4.Looking at bench on each page ?
Code compile, ffmepg / LLVM
Hx370: 1657 / 16043 joules
M4: 473 / 5707 joules
About 3 times more energy to do the compilation
7-zip
m4: 208 joules
HX 370: 1053
About 5x more energy
C-ray: m4: 2513 joules HX370: 7880
3x time more energy
V-ray: 808 vs 1739, 2xtime more energy on the HX
Indigo: 797 vs 1405, 1.8 times or something like that
Quantlib
No hx370 but 10,640 joules vs 54,475 for the 9700x
Apache: 573 vs 1687 joules, 3xtime more energy for the 370hx
DuckDB: 1359 vs 2821, 2xtime more energy for the 370hx
x265 video: 148 vs 269, 1.8x
Kvazaar: 479 vs 443, first test where they are close that I see
FLAC: 64 vs 137, 2x for 370hx
avifenc: 2086 vs 3063, 1.5 time for energy for the hx370
jpeg-xl: 590 vs 1082, a bit less than twice
Are you sure about the nearly matching ? The article in question conclusion say:
But where the M4 was really a standout winner was in the performance-per-Watt with the power efficiency typically well in the lead compared to the tested x86_64 desktop processors on Linux.
Which seem to be true looking at the number, not close at all.
I can agree with the ram, but the storage I can't. I'm assuming that the base Mac Mini M4 is meant to get peoples feet wet and help developers port their applications over since developers can't run MacOS in a VM. If this was going to be my daily computer then 1TB is a minimum. My old HPDV7 that does dual boot Linux and Windows 10 but on separate drives used to have a 256GB just for Windows 10. The Linux Mint installation is on a 256GB SSD because it was an extra drive I had laying around and I just wanted to be able to boot Linux to diagnose things. The Windows 10 installation is used for things I can't run on Linux like BMW ISTA/INPA and Toyota Tech plus some GM stuff for my Vette. I had to upgrade my SSD to 512GB just due to BMW Daten files because I didn't have enough space. If I were to buy a Mac Mini M4 then it would be for things like hosting Jellyfin because it does eat very little power, but not with 256GB. Adding 1TB does bring it to $999 which isn't that bad compared to Beelink's SER9 but the SER9 does come with twice as much ram. Upgrade it to 32GB along with 1TB SSD and the price is $1,400 which is right around desktop PC gaming prices.It's still the entry-level Mac mini. Yes, you're right that more RAM and storage balloons the price quickly, but you realistically don't need more than 16GB of RAM for most uses, and even storage beyond 512GB is less necessary than it once was.
And looking through Phoronix's tests, it's notable that a base M4 is drawing even with or beating the latest-generation AMD and Intel desktop chips in some tests. It does get its ass kicked in some benchmarks, but as has been mentioned... it's a $600 Mac mini competing against PCs where the CPU by itself sometimes costs as much as the entire Mac. Even the middling chips often cost half as much.
We can't test iMovie or Final Cut Pro because Apple doesn't port their software to other platforms. Keep in mind this is done on Linux which generally runs faster on x86 than Windows when it comes to these kinds of tests. People go nuts over Cinebench and Geekbench and no Mac user will ever run these.Let's also not forget that some of these tools are ones you wouldn't realistically use on a Mac. Am I really going to rely on x265 or Kvazaar for video encoding when I can use iMovie (for free) or, if I'm a creative professional, Final Cut Pro or Premiere Pro? I don't think so.
Oh, I'm not saying the storage upgrade prices are great... just that fewer people will be rushing to get larger SSDs than in the past. That and with a Mac mini you can realistically leave an external SSD plugged in if you don't need more primary drive space.I can agree with the ram, but the storage I can't. I'm assuming that the base Mac Mini M4 is meant to get peoples feet wet and help developers port their applications over since developers can't run MacOS in a VM. If this was going to be my daily computer then 1TB is a minimum. My old HPDV7 that does dual boot Linux and Windows 10 but on separate drives used to have a 256GB just for Windows 10. The Linux Mint installation is on a 256GB SSD because it was an extra drive I had laying around and I just wanted to be able to boot Linux to diagnose things. The Windows 10 installation is used for things I can't run on Linux like BMW ISTA/INPA and Toyota Tech plus some GM stuff for my Vette. I had to upgrade my SSD to 512GB just due to BMW Daten files because I didn't have enough space. If I were to buy a Mac Mini M4 then it would be for things like hosting Jellyfin because it does eat very little power, but not with 256GB. Adding 1TB does bring it to $999 which isn't that bad compared to Beelink's SER9 but the SER9 does come with twice as much ram. Upgrade it to 32GB along with 1TB SSD and the price is $1,400 which is right around desktop PC gaming prices.
You can't, but the point is that those are real-world apps that will meaningfully change performance and are often included in workflows. A video producer who needs to turn around a rough edit ASAP isn't going to leave Final Cut Pro or Premiere Pro and process the clip in x265.We can't test iMovie or Final Cut Pro because Apple doesn't port their software to other platforms. Keep in mind this is done on Linux which generally runs faster on x86 than Windows when it comes to these kinds of tests. People go nuts over Cinebench and Geekbench and no Mac user will ever run these.
I'm not the only one complaining about the storage.Oh, I'm not saying the storage upgrade prices are great... just that fewer people will be rushing to get larger SSDs than in the past. That and with a Mac mini you can realistically leave an external SSD plugged in if you don't need more primary drive space.
Why not Davinci Resolve? That's on Mac Windows and Linux.You can't, but the point is that those are real-world apps that will meaningfully change performance and are often included in workflows. A video producer who needs to turn around a rough edit ASAP isn't going to leave Final Cut Pro or Premiere Pro and process the clip in x265.
On that note, Just Josh posted its review of the MacBook Pro M4 line; Geekbench/Cinebench are there, but the review also includes DaVinci Resolve, a video editor that's available on Linux in addition to Mac and Windows... and surprise, the Macs stomp on the PCs. (Skip ahead a few chapters if you want to see those figures.)
View: https://www.youtube.com/watch?v=5rN6CEO31gM
?he Macs stomp on the PCs.
On that note, Just Josh posted its review of the MacBook Pro M4 line; Geekbench/Cinebench are there, but the review also includes DaVinci Resolve, a video editor that's available on Linux in addition to Mac and Windows... and surprise, the Macs stomp on the PCs. (Skip ahead a few chapters if you want to see those figures.)
View: https://www.youtube.com/watch?v=5rN6CEO31gM
So I looked at the video and it's mostly Geekbench and Cinebench. The Cinebench single core is rather impressive as it jumped up a lot, though I'd like to see how this applies to real applications. They used... wildlife extreme to tests graphics. Blender was impressive except the M4 ran out of GPU memory. Interesting that Apple made the M4 Pro and M4 Max equal in CPU configuration. Premiere Pro seems to like Intel chips as he points out. M4 in general seems to do really well at video editing. Other than the video editing tests, the rest are really pointless. I'd wish these guys used games to test the GPU as that would make sense. Especially since Cinebench claims much higher single core IPC and Wildlife Extreme shows the M4 being nearly as fast as an RTX 4090, so I'd expect games to perform nearly as well, in theory. All I can say is that AMD and Intel have better up their CPU performance because these results look damning. AMD has yet to release Strix Halo, though I don't have much faith in increase single core performance but just more cores. Intel... needs a miracle at this point. Though I do wonder if Windows is a part of this problem. Windows tends to lose against Linux, even applications like Davinci Resolve which are usually faster on Linux, and not by a little either.
View: https://youtu.be/m9dZkRwWEj8?si=Ec26NUKNeq1GKYjr
Linux already supports a wider range of software and hardware compared to Windows. Linux also does a better job with legacy code, though a lot of it is also recompiled. We've already seen performance differences with Zen5 and Arrow Lake compared to Windows vs Linux. I feel that this is another indication that Windows isn't putting it's hardware in the best light. Even when it comes to battery drain when the laptop is in sleep, it still seems like a Windows problem. Here is a video from 5 years ago that shows Davinci Resolve on Linux being 15% faster compared to Windows. I can't find more recent benchmarks but I wouldn't be surprised if that's still true to this day.I do think that Windows is the issue, primarily because it has to support things in a way that is more generic across a large range of software and hardware. This, all well holding on the all the legacy code, which really is needed for backwards compatibility. It is a shame we cannot have the Windows desktop experience with the Linux performance combined.
I think a lot of testing is also flawed. Cinebench does paint the M4 as a massive performance increase compared to the M3, and of course compared to AMD and Intel. Forget multicore performance, as it shows a massive increase in single threaded performance. Either AMD and Intel are really that behind compared to Apple, or we have flawed testing tools.Not sure about the windows issue, for the things being tested here at least
take premiere pro export test at 11:54
HX 370: 13:28 minutes
M4: 22:43 minutes
On linux the HX370 performance advantage was pretty much the same.
https://www.phoronix.com/review/apple-m4-intel-amd-linux/6
So I looked at the video and it's mostly Geekbench and Cinebench. The Cinebench single core is rather impressive as it jumped up a lot, though I'd like to see how this applies to real applications. They used... wildlife extreme to tests graphics. Blender was impressive except the M4 ran out of GPU memory. Interesting that Apple made the M4 Pro and M4 Max equal in CPU configuration. Premiere Pro seems to like Intel chips as he points out. M4 in general seems to do really well at video editing. Other than the video editing tests, the rest are really pointless. I'd wish these guys used games to test the GPU as that would make sense. Especially since Cinebench claims much higher single core IPC and Wildlife Extreme shows the M4 being nearly as fast as an RTX 4090, so I'd expect games to perform nearly as well, in theory. All I can say is that AMD and Intel have better up their CPU performance because these results look damning. AMD has yet to release Strix Halo, though I don't have much faith in increase single core performance but just more cores. Intel... needs a miracle at this point. Though I do wonder if Windows is a part of this problem. Windows tends to lose against Linux, even applications like Davinci Resolve which are usually faster on Linux, and not by a little either.
View: https://youtu.be/m9dZkRwWEj8?si=Ec26NUKNeq1GKYjr
Still heavier on synthetic tests than I’d like, I’ll agree. But we won’t get the gaming benches in earnest until there are more high end Mac games, like Cyberpunk as you’ve mentioned.
I keep reminding myself that Apple only really made ray tracing an option as of November last year, and then only for some Macs. It might only be next year that we see RT-capable Mac games arrive in earnest.
AMD and Intel do need to be nervous right now. Laptop-oriented SoCs doing this well at much lower power draw doesn’t bode well for them. And remember, we haven’t seen an updated Mac Studio yet; that’ll have an M4 Max with more cooling than a laptop can provide, and of course the M4 Ultra.
Intel is just hurting full stop. Lunar Lake is the most impressive release it has right now, and it still struggles in multi-core.
I’ll run BG3 on my m4 max when I get it. I suspect it’ll run twice as fast as compared to the M2 Max. That’s nearing the point that I’ll dump my 4080 gaming PC.Still heavier on synthetic tests than I’d like, I’ll agree. But we won’t get the gaming benches in earnest until there are more high end Mac games, like Cyberpunk as you’ve mentioned.
I keep reminding myself that Apple only really made ray tracing an option as of November last year, and then only for some Macs. It might only be next year that we see RT-capable Mac games arrive in earnest.
AMD and Intel do need to be nervous right now. Laptop-oriented SoCs doing this well at much lower power draw doesn’t bode well for them. And remember, we haven’t seen an updated Mac Studio yet; that’ll have an M4 Max with more cooling than a laptop can provide, and of course the M4 Ultra.
Intel is just hurting full stop. Lunar Lake is the most impressive release it has right now, and it still struggles in multi-core.
You are better than me at using it, because when clients send us their scene they have issue with, I often wish I had a workstation instead of a regular computer.Photoshop, Illustrator, and Solidworks on an older Ryzen 5 laptop with a 1060 and I have never once wished I had something faster.
Part is node and pricing power advantage, their competition to not spend tsmc 3NM die space for 16GB-128GB of "cache" and all the power that can save.Why do these perform so well?
In theory the M4 Max should be as fast or faster than the fastest gaming PC. Cinebench shows that the M4's have the highest IPC, which games really like. The M4 Max according to 3D Mark is nearly as fast as an RTX 4090.I’ll run BG3 on my m4 max when I get it. I suspect it’ll run twice as fast as compared to the M2 Max. That’s nearing the point that I’ll dump my 4080 gaming PC.
In theory the M4 Max should be as fast or faster than the fastest gaming PC. Cinebench shows that the M4's have the highest IPC, which games really like. The M4 Max according to 3D Mark is nearly as fast as an RTX 4090.
It's a lot better than it used to be — I honestly enjoyed using a Surface Laptop 7 when I got to try it for a little while this summer. But there was always that app compatibility asterisk, knowing that I couldn't really use it for gaming or any creative apps that weren't already ARM-native (the ones I normally use are).The real issue, in my opinion, is that Arm Based Windows is an dumpster fire, at least when compared to the X86 version. That alone is going to continue to be a pain up the backside of both AMD and Intel.
I suspect it won't be quite that great in practice, but the fact that an integrated GPU can be mentioned in the same breath as high-end dedicated GPUs is saying something.In theory the M4 Max should be as fast or faster than the fastest gaming PC. Cinebench shows that the M4's have the highest IPC, which games really like. The M4 Max according to 3D Mark is nearly as fast as an RTX 4090.
Make you wonder how close it will be on game with an mac silicon decent port:I suspect it won't be quite that great in practice, but the fact that an integrated GPU can be mentioned in the same breath as high-end dedicated GPUs is saying something.
There are a number of native ports on Apple silicon, but the only game in his tests that was native was Tomb Raider. He tested the M4 max on a 14" which people have found to actually thermal throttle the chip, so he wasn't showing the M4 Max in the best light. Most likely he tested the 14" because it was cheaper. The... MSI stealth AI studio A1 comes with an Ultra 9 185H mated with a mobile RTX 4090. Besides the stupid name, the laptop combines a powerful RTX 4090 mobile chip with Intel's not very fast 185H. I didn't think anyone would waste an RTX 4090 with a 185H but yep MSI absolutely built that mistake of a laptop. Would make more sense to include a i9-14900HX despite the terrible battery power that would come with it. I would recommend a laptop with something like a Ryzen 9 7945HX but majority of laptops with an RTX GPU seems to included Intel's 13th and 14th generation CPU's. Makes one wonder how this idiot found the one laptop with a 185H. As usual, the comments section is the most informative as people did point these things out. One guy had to return his M4 Pro as he tested it with Davinci resolve magic mask test. He claims 49 minutes to track the whole clip on M4 pro vs 8 minutes on the 4090.Make you wonder how close it will be on game with an mac silicon decent port:
In theory the M4 Max should be as fast or faster than the fastest gaming PC. Cinebench shows that the M4's have the highest IPC, which games really like. The M4 Max according to 3D Mark is nearly as fast as an RTX 4090.
This makes me wonder: how hard do you think Apple is courting Rockstar to get a Mac-native version of GTA VI? That could literally reshape the computer gaming market when GTA Online is both very popular and has had very long legs so far.Make you wonder how close it will be on game with an mac silicon decent port:
if I could play my gacha games I might grab a beefy one since FF14 runs on mac even with not the most amazing performance. Most are on playstation but sony requires their own separate accounts because of course they do. Still not sure why we can't play iphone/ipad games on mac.I have been toying with the idea of selling my gaming PC, since I barely game anymore, buying the M4 Mac Mini Pro and returning my regular M4 Mac Mini. Faster processor, more P Cores, more GPU cores and faster ones, it would game well enough, perhaps.
GTA VI would pivot Apple's market share if they could convince Rockstar to release it on Mac. Especially since the PC version isn't getting released anytime soon. People would run out and buy a Mac just to play the game. The question is why is Rockstar holding back the PC version, because this may apply to Apple as well? If Rockstar is trying to get people to buy the game twice, then even Apple will take a back seat.This makes me wonder: how hard do you think Apple is courting Rockstar to get a Mac-native version of GTA VI? That could literally reshape the computer gaming market when GTA Online is both very popular and has had very long legs so far.
I did say that I hope Apple isn't paying for ports, and I don't want it to here, but I won't be surprised if executives and engineers are making frequent visits.
GTA VI would pivot Apple's market share if they could convince Rockstar to release it on Mac. Especially since the PC version isn't getting released anytime soon. People would run out and buy a Mac just to play the game. The question is why is Rockstar holding back the PC version, because this may apply to Apple as well? If Rockstar is trying to get people to buy the game twice, then even Apple will take a back seat.
I was going to say, they make some if not actually the most complex piece of software ever, the simplest target first and not pushing the release 1 year (with 500-600 millions of spending eating interest instead of revenue) is probably a simple enough possible scenario here.There's no mystery as to why Rockstar holds back PC launches, as it has for over a decade: it's easier to get the console versions done sooner as you're targeting a much narrower set of hardware that still has a large audience
You could use this: https://topnotch.appAfter a week with my M4 Max, I'm... impressed. It's stupid fast. It's ridiculously well built. All of the HMI bits - keyboard, trackpad, screen, speakers, mics, webcams - are the absolute best in class or right up there. This is the most well rounded laptop I am aware of.
Complaints? Not many, but I'd prefer softer edges when the lid is up. The backlit keyboard is pretty good but not best-in-the-world. Once a week or so, the notch annoys me for a few seconds.
But, it's amazing.
I’m contemplating replacing a lab with the basic M4 variant my only concern is the lack of an ability to secure them to the desk or wall. They would grow legs.After a week with my M4 Max, I'm... impressed. It's stupid fast. It's ridiculously well built. All of the HMI bits - keyboard, trackpad, screen, speakers, mics, webcams - are the absolute best in class or right up there. This is the most well rounded laptop I am aware of.
Complaints? Not many, but I'd prefer softer edges when the lid is up. The backlit keyboard is pretty good but not best-in-the-world. Once a week or so, the notch annoys me for a few seconds.
But, it's amazing.
Won't be surprised if there's a cage (specially designed or generic) you could chain or clamp. And they could still probably use the ports if it's large enough!I’m contemplating replacing a lab with the basic M4 variant my only concern is the lack of an ability to secure them to the desk or wall. They would grow legs.
Nothing like a Kensington lock? Weird.I’m contemplating replacing a lab with the basic M4 variant my only concern is the lack of an ability to secure them to the desk or wall. They would grow legs.