Apple M1 Pro Review - Is It Really Faster than Intel/AMD?

[Shirasaki]

This one covers a wider range of applications rather than “YouTuber exclusive”.

M1 Pro and x86 processor trades lead one after another. M1 Pro isn’t as amazing as some trying so hard for people to believe, but code compiling has been in its own league smoking all x86 competitors.

Due to exorbitantly expensive price of MacBook Pro, this video doesn’t benchmark devices on M1 Max with 32 GPU, so some of those GPU benchmarks M1 Max could potentially perform better.

The rest are all explained in this video.

Edit: So much hostility...

 

[mr_roboto]

This one covers a wider range of applications rather than “YouTuber exclusive”.

M1 Pro and x86 processor trades lead one after another. M1 Pro isn’t as amazing as some trying so hard for people to believe, but code compiling has been in its own league smoking all x86 competitors.

The video doesn't really support your take on it. So many of the not-amazing results are obviously due to factors which have nothing to do with M1 Pro's performance potential - lots of software run through Rosetta, other stuff with questionably mature ports, and so forth. And the section on power consumption makes it clear M1 Pro really is in its own league.

 

[spiderman0616]

People are doing some pretty impressive mental backflips trying to make sense of what Apple Silicon is about to do to the computer industry.

 

[TrueBlou]

I’ve given up on the X86/64 comparisons now, there seems little logical reason to it anymore.

If we’re Apple users, and considering this is MacRumors I’d guess many of us are (though I do still have a WinTel rig as well), then it makes no difference what the other side of the fence does, we’re not going to get it. Apple have moved on, personally I think for the better in the long run.

The only benefit now to Intel, AMD, Qualcomm and even ARM making advances is that it pushes their competitors. I’m all for Intel making better, faster (and hopefully more energy efficient) chips. But only because it’ll keep the pressure on Apple to improve Apple Silicon, which if the A series history is anything to go by, and it is, they will continue to do anyway. We’ll all reap the rewards, whichever platform we’re using.

 

[jeanlain]

Matlab and Microsoft Excel and a discussion on optimisation.

I don't know about Matlab, but Excel? It may be ARM native, but it's notoriously much slower on intel macOS than on Windows.

 

[Deadlywrong]

He is really just focus on the performance wise. He did mention the fan noise. That is what matters to me a lot.
There are other things like display and speakers that adds to the price tag. I do agree with him on those expensive upgrade costs for memory and SSD though. But at the end of day I was not aware of any PC laptop has that fast SSD and that impressive bandwidth memory.

 

[Shirasaki]

The video doesn't really support your take on it. So many of the not-amazing results are obviously due to factors which have nothing to do with M1 Pro's performance potential - lots of software run through Rosetta, other stuff with questionably mature ports, and so forth. And the section on power consumption makes it clear M1 Pro really is in its own league.

My take is normal people ain’t gonna run premier daily or play with Final Cut Pro professionally unless that’s their job. I have yet to see a single YouTuber outside of him testing anything that has nothing to do with video production or tasks around it, yet people can unzip archives using 7zip from time to time, same for exporting PDF files. Sure, those software isn’t fully optimised for M1 but after a full year if this is the software situation, we still have quite a way to go.

 

[Shirasaki]

I don't know about Matlab, but Excel? It may be ARM native, but it's notoriously much slower on intel macOS than on Windows.

I’d much rather call majority of “apple silicon native app” “apple silicon COMPATIBLE app”, without using any feature apple silicon designed for.

 

[chengengaun]

Alexander Ziskind did quite a lot of tests and benchmarks related to software development, which I think is worth checking out.

I hope the dramatic improvement in hardware capabilities will attract more developers dedicating effort to macOS-optimized apps and a wider range of users. “Pro” markets are usually lucrative and Apple can certainly continue to charge premium prices so long as the performance gaps with WinTel systems (using optimized apps) is significant.

 

[AlexESP]

My take is normal people ain’t gonna run premier daily or play with Final Cut Pro professionally unless that’s their job. I have yet to see a single YouTuber outside of him testing anything that has nothing to do with video production or tasks around it, yet people can unzip archives using 7zip from time to time, same for exporting PDF files. Sure, those software isn’t fully optimised for M1 but after a full year if this is the software situation, we still have quite a way to go.

“Normal people” are not the target audience of the M1 Pro. Yes, most people who buy a MBP will probably use Final Cut Pro, or Logic, Xcode… “Normal people” have the M1, which crushes every similar priced competitor in performance and efficiency.

 

[jeanlain]

My take is normal people ain’t gonna run premier daily or play with Final Cut Pro professionally unless that’s their job. I have yet to see a single YouTuber outside of him testing anything that has nothing to do with video production or tasks around it, yet people can unzip archives using 7zip from time to time, same for exporting PDF files.

OTOH, Excel, compression and office work do not stress the hardwware much, for most people. We're talking about a few seconds saved here and there.

 

[the8thark]

If we’re Apple users, and considering this is MacRumors I’d guess many of us are (though I do still have a WinTel rig as well), then it makes no difference what the other side of the fence does, we’re not going to get it. Apple have moved on, personally I think for the better in the long run.

Agreed but with one caveat. We do want the others (AMD, Intel etc etc) to keep innovating. This will give Apple the push it needs to keep innovating as well. Sure for a while Apple's only competition is Apple itself, but external competition is always welcome and benefits the end consumer.

 

[TrueBlou]

Agreed but with one caveat. We do want the others (AMD, Intel etc etc) to keep innovating. This will give Apple the push it needs to keep innovating as well. Sure for a while Apple's only competition is Apple itself, but external competition is always welcome and benefits the end consumer.

I couldn’t agree more. Which is why I said as much in my post. Competition in any industry is essential to product advancement and innovation. Without it there would inevitably be stagnation as there would be less drive to push forward.

 

[yitwail]

Alexander Ziskind did quite a lot of tests and benchmarks related to software development, which I think is worth checking out.

After watching one of those videos, I canceled order for an upgraded base 14in MBP and instead ordered an upgraded M1 MBA. If I were wealthier, I wouldn’t have done that, and I do think ports, display, and MagSafe in addition to higher performance justify the extra cost, but I have better uses for the money I saved.

 

[Analog Kid]

The power usage section is revealing - Turbo Boost really spiked power usage on Intel CPUs. Also nice to see other apps featured esp. Matlab and Microsoft Excel and a discussion on optimisation.

Didn't watch the video to see how this is addressed, but last I checked Matlab is still running under Rosetta. Not a very good test candidate to compare M1 against Intel.

 

[shenfrey]

I couldn’t agree more. Which is why I said as much in my post. Competition in any industry is essential to product advancement and innovation. Without it there would inevitably be stagnation as there would be less drive to push forward.

Not strictly true. Apple could keep pushing so far ahead that competition never catches up. Why quit when you're so far ahead, just keep going and make it impossible for anyone else to catch you.

You should never rest on your laurels.

 

[kpluck]

They say that the video is not a review of the MacBook Pro but benchmarking the M1 Pro SOC. If that was the case, they should only be using tests that run natively on the M1 Pro processor. They go so far as showing a Blender benchmark with GPU rendering even though that is not supported on the M1 Pro. They also complain about Apple prices and upgrade costs which seems to me to be something other than benchmarking the M1 Pro.

The fact is a lot of what they did is closer to a review on Mac vs Windows using the MBP M1 Pro, rather than looking at the performance of the M1 Pro SOC.

I really like Hardware Unboxed and have been subscribed to their YouTube channel for quite some time time. However, I think this is probably the worst review I have seen them do.

-kp

 

[TrueBlou]

Not strictly true. Apple could keep pushing so far ahead that competition never catches up. Why quit when you're so far ahead, just keep going and make it impossible for anyone else to catch you.

You should never rest on your laurels.

Which is why I also said in my post that if the A series chip development history is anything to go by, and as they share the same architecture, it should be. That Apple will continue to improve Apple Silicon regardless.

 

[mi7chy]

They haven't run into the memory leaks.

https://www.macworld.com/article/549755/m1-macbook-app-memory-leaks-macos.html

https://developer.apple.com/forums/thread/693701

 

[playtech1]

I think this is a reasonably fair assessment of M1 Pro performance. Basically it's very good (and incredible in certain task) and not so good in unoptimised applications. No problem at all with that conclusion.

I think what's missing from this assessment in terms of buying choices are the platforms that M1 are offered in versus the platforms that the competing x86 chips are offered in.

A MacBook Pro is a fairly unique beast in terms of its particular combination of OS, build quality, screen, speakers, support, trackpad, keyboard, battery life, performance on battery, heat/noise, etc.

IMHO there simply doesn't exist on the market a PC that matches up to what Apple is selling right now. A few come close, but none beats it overall (if you exclude Mac OS vs Windows as a factor).

If a Ryzen chip performs, say, 30% better in a number of applications that matter to me then that's all well and good, but if I can only buy that CPU in a flimsy plastic Asus shell with terrible battery life, noisy fans and speakers that sound like a wasp in a tin can... well I'll stick with the MacBook Pro thanks.

Now if Dell, HP, Razer or some other company were to actually get to Apple levels of quality all around, then I might be tempted. But that seems some distance away right now.

 

[Pro Apple Silicon]

A MacBook Pro is a fairly unique beast in terms of its particular combination of OS, build quality, screen, speakers, support, trackpad, keyboard, battery life, performance on battery, heat/noise, etc.

IMHO there simply doesn't exist on the market a PC that matches up to what Apple is selling right now.

This ^

These silly comparison videos are ridiculous clickbait nonsense. The above is what reality is. No one is going to decide to buy some x86 legacy crap because of some misleading benchmark, instead of the above, for the above reasons.

 

[CMMChris]

I don't get why in his gaming comparison he didn't test Baldurs Gate 3 which is native and does perform extremely well. Would've been a great comparison candidate.

 

[fgengineer]

This video validates my own findings which is why I returned my MacBook Pro. Performance was mixed for me and there are some work loads I couldn’t run.

If I were more mobile, I would probably not have returned it because it’s efficiency cannot be beat, but I mostly just work from home right now.

 

[bill-p]

This one covers a wider range of applications rather than “YouTuber exclusive”.

M1 Pro and x86 processor trades lead one after another. M1 Pro isn’t as amazing as some trying so hard for people to believe, but code compiling has been in its own league smoking all x86 competitors.

I don't see how this one covers "a wider range". It's simply just another set of benchmarks.

In benchmarks that M1 isn't doing so well, note that Matlab R2020 doesn't even support ARM natively. If M1 Pro could actually beat anything in that comparison, it's already a miracle brought about by Rosetta 2. And in fact, if M1 Pro did beat everything, that would just mean AMD and Intel might as well throw their processors away since an ARM processor simulating an x86 chip runs faster and more efficient than literally ALL of their designs. Supposedly, Matlab R2021 will support ARM.

Same for 7-zip. I noticed he was using 7-zip version 19.0 in that video. Note that 7-zip for ARM is in Alpha right now (version 21) and heck, it's not even available on MacOS. To run this 7-zip benchmark, you'd have to go through Parallels. Even more overhead.

Again, if the M1 or M1 Pro/Max could beat a lot of CPUs under these "benchmarks", it would mean the machine can smoke Intel and AMD in... virtual machines.

So I actually see this as a positive thing for M1. It means M1 can actually emulate x86 using either Rosetta 2 or running Windows in Parallels faster than most x86 processors on the market. If anything, it just highlights why I no longer want an Intel or AMD laptop.

 

[bill-p]

My take is normal people ain’t gonna run premier daily or play with Final Cut Pro professionally unless that’s their job. I have yet to see a single YouTuber outside of him testing anything that has nothing to do with video production or tasks around it, yet people can unzip archives using 7zip from time to time, same for exporting PDF files. Sure, those software isn’t fully optimised for M1 but after a full year if this is the software situation, we still have quite a way to go.

Normal people on Mac don't run Matlab... nor do they run 7zip. Heck, 7zip doesn't even exist on Mac. We have Keka. Just the fact that you mentioned 7zip is already a signal to me that you don't use a Mac computer on a regular basis (or at all).

Also Mac doesn't need to export PDF files. The option to print and save as PDF has always been there and is much faster than some built-in PDF engines in some apps.

Some of the software shown are either being optimized for M1 or already has an alternative. The person doing the benchmark was just treating the MacBook as if it was a Windows machine going through Parallels. That's not realistic at all. Most Mac users just use MacOS.

Please get to know the target platform before you criticize it.