My Extremely Scientific MacBook Pro Benchmark - Years in the Making

[Skoobunny]

My Extremely Scientific MacBook Pro Benchmark — Years in the Making

First of all, please take this post with a grain of salt. It’s not meant to prove anything, settle any arguments, or start another Apple Silicon vs. Intel war.

It’s just for fun. 🙂

All of these machines have been — or still are — used for actual professional work.

But after hours… well, there’s Civilization VI.

So, ladies and gentlemen, here is my completely objective, scientifically rigorous, and at the same time entirely personal MacBook Pro benchmark.

Testing methodology:

A legally owned copy of Civilization VI.
A 4K external monitor.
 Exactly the same game settings on every machine.
 A game already past the year 1500, when the map is busy enough to make the hardware reconsider its life choices.

The benchmark measures one extremely important parameter:

NOISE.

From loudest to quietest — the fewer points, the better:

MacBook Pro M3 Pro — 100 points. Fans screaming pretty much the entire time. This is our reference value.

MacBook Pro M2 Max — 80 points. 
Fans screaming the entire time as well, but somewhat more politely. Still very much present, though.

MacBook Pro M5 Max — 30 points
. Fans running continuously, but dramatically quieter. You can hear them, but they no longer sound like the MacBook is preparing for takeoff.

And finally…

2018 MacBook Pro — six-core Intel i9 + Blackmagic eGPU — 0 points. Silence. The MacBook fans? Inaudible.
The Blackmagic eGPU fan? Inaudible.

Apparently, after all these years of technological progress, my ultimate Civilization VI machine is still an Intel MacBook Pro from 2018 connected to a giant black external GPU ;-)

Science has spoken.

Cheers, everyone.

Now I’m waiting for the M7 Max to finally beat a 2018 Intel MacBook at this highly sophisticated benchmark. ;-)

 

[OldNewMBP]

Can't wait for comments to roll in saying that you cheated with the 2018 MBP. 🤣

Anyways, what great effort that you put into it! I'd say that your work was worth it!

 

[BUDGETApple]

Don’t use the external GPU and test the Intel MacBook Pro again.

 

[DaveEcc]

My results:

Machine	Skip the dishes tracking driver phase	Lightroom bulk export
27" i7 late 2012 iMac	Untested	Audible fans, nowhere near PS4 levels
15" 2013 MBP	Over 100C, fans like a jet / loaded PS4	Audible fans, nowhere near PS4 levels
14" M1 Pro 2021 MBP	Cool and silent	Cool and silent

Clearly the most difficult task you can ask of a Mac is to track the location of your food delivery driver.

I'm still not sure if my M1 Pro has fans.

 

[Skoobunny]

Don’t use the external GPU and test the Intel MacBook Pro again.

But that would ruin the whole experiment! 😄

The Blackmagic eGPU is part of the 2018 setup — that’s exactly how I used that MacBook back then. This isn’t a CPU/GPU performance benchmark. It’s my extremely scientific “how much noise does my complete MacBook setup make while playing Civilization VI on a 4K monitor” benchmark.

Actually, I’d much rather do the exact opposite: connect the Blackmagic eGPU to the M3 Pro, M2 Max, and M5 Max and run the same test.

Unfortunately, Apple Silicon Macs don’t support eGPUs.

So removing the eGPU from the 2018 MacBook would make the comparison more “fair” only by taking away a capability the newer Macs no longer have.

And besides, remove the eGPU and we’re doing actual science.

That would be completely against the spirit of this test. ;-)

 

[OldNewMBP]

But that would ruin the whole experiment! 😄

The Blackmagic eGPU is part of the 2018 setup — that’s exactly how I used that MacBook back then. This isn’t a CPU/GPU performance benchmark. It’s my extremely scientific “how much noise does my complete MacBook setup make while playing Civilization VI on a 4K monitor” benchmark.

Actually, I’d much rather do the exact opposite: connect the Blackmagic eGPU to the M3 Pro, M2 Max, and M5 Max and run the same test.

Unfortunately, Apple Silicon Macs don’t support eGPUs.

So removing the eGPU from the 2018 MacBook would make the comparison more “fair” only by taking away a capability the newer Macs no longer have.

And besides, remove the eGPU and we’re doing actual science.

That would be completely against the spirit of this test. ;-)

I have a competitor:

2015 MBA 13" /w 1.6 GHz Intel Core I5 with a external laptop cooler: Absolutely no noise. (Although it gets to 100 Degrees Celsius under load, but is still quiet)

 

[Skoobunny]

I have a competitor:

2015 MBA 13" /w 1.6 GHz Intel Core I5: (ALMOST) Absolutely no noise. (Although it gets to 100 Degrees Celsius under load, but is still quiet)

Now we’re talking! 😄

So we officially have a competitor:

2015 MacBook Air — 0 points.

Reaching 100°C under load is completely irrelevant to this benchmark. We are measuring noise, not survival probability. ;-)

However, before I can officially accept this result, I need to verify the testing methodology:

Was the test conducted using a legally owned copy of Civilization VI, connected to a 4K monitor, with the game already past the year 1500?

We do have scientific standards to maintain here. ;-)

 

[tiguanito]

My Extremely Scientific MacBook Pro Benchmark — Years in the Making

Testing methodology:

A legally owned copy of Civilization VI.
A 4K external monitor.
 Exactly the same game settings on every machine.
 A game already past the year 1500, when the map is busy enough to make the hardware reconsider its life choices.

Did you set the energy mode to high power during the tests ? if it's left on Automatic, the fans on the M5 max (I don't know about the other chips) tend to run at around 60% and the GPU is throttled. When set to High power, the fans can ramp up to 100%, there is less throttling. I observed this while running local LLMs, with higher tokens/sec and fan noise in high power mode.

 

[Skoobunny]

Did you set the energy mode to high power during the tests ? if it's left on Automatic, the fans on the M5 max (I don't know about the other chips) tend to run at around 60% and the GPU is throttled. When set to High power, the fans can ramp up to 100%, there is less throttling. I observed this while running local LLMs, with higher tokens/sec and fan noise in high power mode.

That’s a good point, but thinking about it, I’m not sure High Power mode would actually make sense for this particular test.

The whole idea of my extremely scientific benchmark is not to squeeze the maximum possible FPS out of each MacBook. It’s to see how quietly the machine can run Civilization VI at 4K while still providing perfectly smooth gameplay.

On the M5 Max I normally use Automatic mode anyway, because I mostly use this machine for local LLMs — running it permanently in High Power mode would probably turn it into a rather expensive cooking appliance. ;-)

And even in Automatic mode, Civilization VI is perfectly smooth. High Power might give me a few more FPS and reduce GPU throttling, but if the price is higher fan speed and more noise, that would actually make the result worse for the purpose of this test, not better.

I wasn’t monitoring clocks, temperatures or fan speeds during the benchmark either, since Civilization VI was running full screen. But in this case I’m mainly interested in what I can actually hear while playing.

So I think Automatic is probably the right setting for this particular piece of extremely serious scientific research. ;-)

 

[OldNewMBP]

Now we’re talking! 😄

So we officially have a competitor:

2015 MacBook Air — 0 points.

Reaching 100°C under load is completely irrelevant to this benchmark. We are measuring noise, not survival probability. ;-)

However, before I can officially accept this result, I need to verify the testing methodology:

Was the test conducted using a legally owned copy of Civilization VI, connected to a 4K monitor, with the game already past the year 1500?

We do have scientific standards to maintain here. ;-)

Bump it up to 10-15 points, I didn't do that kind of test, but I did use Ukeysoft as the one putting all the pressure on the MBA. Exactly how resource hungry is Civilization 4(VI)?

 

[Skoobunny]

Bump it up to 10-15 points, I didn't do that kind of test, but I did use Ukeysoft as the one putting all the pressure on the MBA. Exactly how resource hungry is Civilization 4(VI)?

I'm afraid I can't give you 10–15 points just yet. ;-) Ukeysoft may put plenty of pressure on the MBA, but this is a Civilization VI benchmark and the Scientific Committee is very strict about methodology.

Civ VI itself isn't an especially demanding game by modern standards, but I'm running it full screen at 4K, so it does put a reasonable load on the GPU. More importantly, it gives me exactly the same real-world workload on every MacBook I've tested over the years.

So there is only one way to get those points: a legally owned copy of Civilization VI, a 4K display, and the two factory-installed, uncalibrated human ears as measuring equipment. ;-)

Then we can talk!

 

[OldNewMBP]

I'm afraid I can't give you 10–15 points just yet. ;-) Ukeysoft may put plenty of pressure on the MBA, but this is a Civilization VI benchmark and the Scientific Committee is very strict about methodology.

Civ VI itself isn't an especially demanding game by modern standards, but I'm running it full screen at 4K, so it does put a reasonable load on the GPU. More importantly, it gives me exactly the same real-world workload on every MacBook I've tested over the years.

So there is only one way to get those points: a legally owned copy of Civilization VI, a 4K display, and the two factory-installed, uncalibrated human ears as measuring equipment. ;-)

Then we can talk!

I can't guarantee that I would be able to do that just yet, but I will see what I can do.

 

[Skoobunny]

I can't guarantee that I would be able to do that just yet, but I will see what I can do.

Excellent! The Scientific Committee will be waiting for your results. ;-)