Macbook Pro 2019 Benchmarks and First Impressions! Post yours!

[SteveJUAE]

Y
That is the reason why I believe that Apple's cooling solution is pretty damn good. They chose a more complex design in order to make the laptop more versatile, even if this design is inherently less efficient than direct air suction through the bottom grills, and they still achieve a reasonable result. Funnily enough, one reason why it works is probably because it is so thin Its like a vacuum inside. It's quite smart IMO.

Bottomline: if Apple adopted a design like every other laptop and put inlet vents on the bottom, directly in front of the vents, then yes, the cooling performance of the MBP would most certainly improve. Would it make it a better laptop though? I doubt it. Many users choose Apple exactly because of those little things like absence of bottom grills. It makes the laptop just overall more useable.

This is probably an over simplification of the design choices Apple made

Whilst your assumptions on cooling are likely to be correct it negates the other issues that are equally likely simplified by not having vents on the bottom not least of all structural integrity of the chassis being so thin. It's also possible that placement of components on the logic board were simplified. It may also be a limitation and/or cost prohibitive aspect of the uni body design and materials

To imply all other aspects of design/construction are equal and Apple picked a more complex solution in isolation is unlikely.

There's no bragging rights here just the end result of many design decisions taken

 

[leman]

Intel had recommended disabling hyper-threading due to the MDS vulnerabilities

Ah, the Zombieload stuff? I am not particularly worried about that one, since its very difficult to execute properly, relies precise knowledge of the victim software and of course, requires the attacker and the victim run on the same physical core. MacOS does not offer any APIs that would allow the attacker to set this up reliably.
[doublepost=1558835160][/doublepost]

Whilst your assumptions on cooling are likely to be correct it negates the other issues that are equally likely simplified by not having vents on the bottom not least of all structural integrity of the chassis being so thin. It's also possible that placement of components on the logic board were simplified. It may also be a limitation and/or cost prohibitive aspect of the uni body design and materials

True, and of course everything I wrote is just conjecture. I do not have any insider knowledge about Apple's design process or decisions. However, they never used bottom inlet vents on any of their laptops, so it's obviously a deliberate choice. Given how Apple is focused on minimalist functionality and clean aesthetics, I like to think that my guess is mostly accurate

 

[OVERTASK]

Intel had recommended disabling hyper-threading due to the MDS vulnerabilities

Not just that, my workloads don't really benefit from HTT and it runs a little cooler (95º instead of 100 )

 

[souko]

 

[richinaus]

that was a positive video!

I am starting to get nudged to the Apple side of the fence again, right when I was ready to jump off.
I can imagine this hooked up with an eGPU could be quite awesome.

 

[wallysb01]

that was a positive video!

I am starting to get nudged to the Apple side of the fence again, right when I was ready to jump off.
I can imagine this hooked up with an eGPU could be quite awesome.

That is good, R20 score are about 100+ higher and the CPU is maintaining about 100MHz (3GHz vs 2.9GHz) higher than the 2.3. Looking like the jump from the 2.3 to the 2.4 won’t be nerfed by the thermals. Think I’ve nearly seen all I need to see. Just would like a complete run down by one of the more we’ll known review sites.

 

[Thysanoptera]

That is good, R20 score are about 100+ higher and the CPU is maintaining about 100MHz (3GHz vs 2.9GHz) higher than the 2.3. Looking like the jump from the 2.3 to the 2.4 won’t be nerfed by the thermals. Think I’ve nearly seen all I need to see. Just would like a complete run down by one of the more we’ll known review sites.

If you look at the power values, the machine with 2.3 runs at some 5W less and that would easily explain the difference in scores and clocks. That's a difference between 555x and 560x, or Vega 16 and 20. This has actually nothing to do with CPU model, but may be down to ambient temp or quality of particular assembly. If you remember, with 2018 we had a good 20% range on Cinebench 15 scores on the same CPU (from low 900 to 1100). I mean - buying 2.4 is not going to hurt, it is only $200 and you can walk into Apple store and pick it up - but just don't expect any measurable performance gains.

And I really hate, hate so much, when a reviewer pulls a statement like this (from this appleinsider review, text version):
"But, whatever Apple has done is working. The 2018 six-core machine is still an amazingly powerful machine, even if thermal conditions inside the case pull it back from what it could be.

That limitation doesn't seem to exist in the new eight-core machine."

Like WTF is he talking about???? It is around 55W thermal limit in both cases for crying out loud. And they show an image of cooling hardware differences between 2018 and 2019 Vega models, but the picture is actually between Vega and RX versions....

 

[ocnitsa]

that was a positive video!

I am starting to get nudged to the Apple side of the fence again, right when I was ready to jump off.
I can imagine this hooked up with an eGPU could be quite awesome.

Same. These incremental updates are a good thing imo. Work owes me a laptop and I was ready to go PC. I’m leaning back towards MBP, again.

 

[wallysb01]

If you look at the power values, the machine with 2.3 runs at some 5W less and that would easily explain the difference in scores and clocks. That's a difference between 555x and 560x, or Vega 16 and 20. This has actually nothing to do with CPU model, but may be down to ambient temp or quality of particular assembly. If you remember, with 2018 we had a good 20% range on Cinebench 15 scores on the same CPU (from low 900 to 1100). I mean - buying 2.4 is not going to hurt, it is only $200 and you can walk into Apple store and pick it up - but just don't expect any measurable performance gains.

And I really hate, hate so much, when a reviewer pulls a statement like this (from this appleinsider review, text version):
"But, whatever Apple has done is working. The 2018 six-core machine is still an amazingly powerful machine, even if thermal conditions inside the case pull it back from what it could be.

That limitation doesn't seem to exist in the new eight-core machine."

Like WTF is he talking about???? It is around 55W thermal limit in both cases for crying out loud. And they show an image of cooling hardware differences between 2018 and 2019 Vega models, but the picture is actually between Vega and RX versions....

Good points, but I’d be surprised if 5W is explained by room temp differences. I guess I’d have to break out some of my old physics to see if the flow rate and in/out temp differences really even stand a chance, but I’m skeptical. And I wouldn’t expect more than 3%, that’s what is advertised. Maybe you’re right though, and that’s within the margin of error for the silicon lottery.

 

[Thysanoptera]

Good points, but I’d be surprised if 5W is explained by room temp differences. I guess I’d have to break out some of my old physics to see if the flow rate and in/out temp differences really even stand a chance, but I’m skeptical. And I wouldn’t expect more than 3%, that’s what is advertised. Maybe you’re right though, and that’s within the margin of error for the silicon lottery.

We don't really know anything about the physical properties of the cooling system, so you can't really calculate this. But you will have more or less 1:1 relation between ambient and cpu temperature, and just for fun I checked how the power/temp ratio looks like on my GS65 with constant fan speed, and it is kind of 1C/1W, not kidding. So it looks like when the cooling is maxed out you can cause 5W of difference with 5C ambient temp difference, which is totally doable. Silicon would come to play if you wanted to max out the clock at given voltage, here it is just about cooling systems, Watts are Watts regardless the die quality.

EDIT - but honestly the biggest impact will come from the heatplate assembly, the reported package temperature is the temperature of the hottest core, and in laptops with their delicate heat transfer elements and their mountings it is very easy to cause one (or more) of them to be totally off the others, which in turn will drive the package clock down.

 

[smbu2000]

Curious to see what are the Cinebench r20 scores for the other variants, namely the 2016/2017 and the 2018 i7 2.2 and 2018 i9 2.9 models. Chime in your scores please! Cinebench r20 can be downloaded on Mac app store."

Since you asked: My max 2016 15" (2.9GHz i7-6920HQ/16GB/R460/2TB) scored 1867 on Cinebench R20.

When using the fans set to auto it turbos up to 3.4GHz (which is the turbo speed for 3 or 4 cores) and about halfway through will hit 99C and then drop down to 3.3GHz and then will fluctuate between 3.2 and 3.3. The score was about 1840-1850.
When I set the fans to always on max speed, then it will stay at 3.4GHz for most of the test only dropping to 3.3 near the end. Scores are only a bit higher though in the 1860 range.

 

[maflynn]

that was a positive video!

For me, I find AI to be a bit biased in the pro apple corner, along the lines of iMore. I usually take their videos with a grain of salt.

I am starting to get nudged to the Apple side of the fence again, right when I was ready to jump off.

My recommendation is before making any firm decisions, is to get more rounded reviews, such as Lisa from Mobile Tech Review and others I'm not saying the 2019 8 core machine isn't a good laptop but I think more reviews are needed to developed a fully rounded opinion.

My personal opinion of the 2019 MBP has improved, initially I thought it was going to run too hot, I was wrong. It may not run as fast as other similarly configured 8-core laptops but then laptops are balance of compromises and the thin design of the MBP means something has to give. I was fairly critical of the 2019 MBP, but as I stated, the more I read/watch reviews the more I'm impressed. Personally, its not for me, I have no remorse over my departure from the mac ecosystem but overall it appears to be fine laptop.

The keyboard is the wildcard in the mix and apple's dedication to x86 architecture, in both cases time will tell. It took 4 to 6 months before the 3rd gen butterfly keyboard started showing signs of failure and so at the minimum we'll have to wait that long for the 4th gen. As for Apple moving to ARM, we may find out about that possible transition in the 2020 WWDC. My perspective is that i don't want to remain in or buy into a platform that is changing architectures and that architecture is counter to my needs. It may be fine for many people, and I'm not knocking that, but for my needs its not a move I could live with.

I can imagine this hooked up with an eGPU could be quite awesome.

I don't understand the allure of the eGPU, at least how its seems to be envisioned here at MR. The cost of a BlackMagic eGPU is 700 dollars, the 8 core MBP is easily 3k after taxes and adding in various configuration upgrades. It could top off in the 4k range. My point is you're looking at 4,000 to 5,000 dollar price tag for a MBP+eGPU arrangement. I think for many people there are other solutions that make more financial sense.

 

[William Payne]

Well I did some tests on my 2018 i7 MacBook Pro to get an idea of whether I would at all benefit from 8 cores. Now I chose the vega 20 gpu so factored that in.

Cinebench roughy gave me a 2450-2500 score. That however didn’t tell me much as 99% of the time I will be using Photoshop.

I found a photoshop benchmark that Puget Systems used for their testing and well it’s clock speed was fine. Staying consistently around 4.0GHZ even when hot. But most interesting for my use case was that with my 2.6GHZ 6 core my cpu utilisation was hardly ever above 50% and would spend a lot of time barely on 20%.

I think for me in my use case my Vega 20 gpu is used far more than my core count.

Also the top score on the benchmark is 1000 which is what their reference pc behemoth machine gets. Thats with a 2080, an 8 core cpu, 64GB ram. My 2018 MacBook Pro scores 800. I asked Puget Systems what they thought of that score and they said for a laptop that is really good.

 

[grmlin]

that was the wrong thread, somehow...

 

[malikkamran1]

When it will be water proof

 

[OSXconvert]

Well I did some tests on my 2018 i7 MacBook Pro to get an idea of whether I would at all benefit from 8 cores. Now I chose the vega 20 gpu so factored that in.

Cinebench roughy gave me a 2450-2500 score. That however didn’t tell me much as 99% of the time I will be using Photoshop.

I found a photoshop benchmark that Puget Systems used for their testing and well it’s clock speed was fine. Staying consistently around 4.0GHZ even when hot. But most interesting for my use case was that with my 2.6GHZ 6 core my cpu utilisation was hardly ever above 50% and would spend a lot of time barely on 20%.

I think for me in my use case my Vega 20 gpu is used far more than my core count.

View attachment 839195

William, did you publish your full puget sound photoshop cc test results on their page? I, too, evaluate a system on how it runs PS and LR. I badly need a new MBP, and have been wondering how much the 8-core 2.4 vs 2.3 compares, as well as the Vega options vs standard 560x Gpu. Please let us know your full extended score (16-bit) here. Thx.

 

[William Payne]

William, did you publish your puget sound photoshop cc test results on their page? I, too, evaluate a system on how it runs PS and LR. I badly need a new MBP, and have been wondering how much the 8-core 2.4 vs 2.3 compares, as well as the Vega options vs standard 560x Gpu. Please let us know your extended score here. Thx.

I got 800 overall on their extended test on my 2018 MacBook Pro with 2.6GHZ i7 32GB ram and Vega 20 GPU.

I would love to know what a 2019 8 core with the same ram/gpu specs would get but I am not in a position to do that test.

 

[richinaus]

For me, I find AI to be a bit biased in the pro apple corner, along the lines of iMore. I usually take their videos with a grain of salt.


My recommendation is before making any firm decisions, is to get more rounded reviews, such as Lisa from Mobile Tech Review and others I'm not saying the 2019 8 core machine isn't a good laptop but I think more reviews are needed to developed a fully rounded opinion.

My personal opinion of the 2019 MBP has improved, initially I thought it was going to run too hot, I was wrong. It may not run as fast as other similarly configured 8-core laptops but then laptops are balance of compromises and the thin design of the MBP means something has to give. I was fairly critical of the 2019 MBP, but as I stated, the more I read/watch reviews the more I'm impressed. Personally, its not for me, I have no remorse over my departure from the mac ecosystem but overall it appears to be fine laptop.

The keyboard is the wildcard in the mix and apple's dedication to x86 architecture, in both cases time will tell. It took 4 to 6 months before the 3rd gen butterfly keyboard started showing signs of failure and so at the minimum we'll have to wait that long for the 4th gen. As for Apple moving to ARM, we may find out about that possible transition in the 2020 WWDC. My perspective is that i don't want to remain in or buy into a platform that is changing architectures and that architecture is counter to my needs. It may be fine for many people, and I'm not knocking that, but for my needs its not a move I could live with.


I don't understand the allure of the eGPU, at least how its seems to be envisioned here at MR. The cost of a BlackMagic eGPU is 700 dollars, the 8 core MBP is easily 3k after taxes and adding in various configuration upgrades. It could top off in the 4k range. My point is you're looking at 4,000 to 5,000 dollar price tag for a MBP+eGPU arrangement. I think for many people there are other solutions that make more financial sense.

I am not jumping in by any means immediately, and am awaiting further reviews as you say.

As I have mentioned previously I am really OS agnostic, and use both Windows and Mac and am happy with either for professional work. A friend is lending me a brand new HP Zbook 360 today to see my impressions as that is the machine I would buy as an alternative. So I will see the performance, build etc and then make my mind up. Lisa Glade gave it a great review btw.

The ARM comment is actually interesting and you may be right - I cannot be risking my work to allow them to transition [I went through the PowerPC - Intel fun], however I don't believe they will be doing this on pro machines for at least 5 years. I could be wrong but just cannot see it. If it did happen, and I did get a MBP, I would just leave it in bootcamp and use windows. Very simple and easy for me.

The allure of a eGPU for me is this - I do VR work and a lot of 3D modelling / design testing and it is now all based on the GPU. I also need to have a laptop only solution. So the best of both is laptop and desktop GPU. I am not gaming at all, but is purely a work requirement. You may not think it makes sense in your field of work, but for me there is no laptop I would buy that I would be happy throttling the GPU to 99% on sustained loads, which is what happens.

In reality I rarely max out the CPU so won't throttle that, but the GPU will certainly get it. My iMacpro gets hot and loud under sustained VR / live rendering use. I would prefer that in a separate box tbh. and will still buy a egpu for whatever machine I end up getting, windows or Mac. Sort of irrelevant to the discussion at hand really though.

 

[OSXconvert]

I got 800 overall on their extended test on my 2018 MacBook Pro with 2.6GHZ i7 32GB ram and Vega 20 GPU.

I would love to know what a 2019 8 core with the same ram/gpu specs would get but I am not in a position to do that test.

Yes, me, too. It gives a good comparison with my hackintosh 8700K 5.0ghz desktop which scores 1109.

So photoshop users who bought a new i9 MacBook Pro 2019, please take the Puget Sound PS speed test and let us know how you do:

https://www.pugetsystems.com/labs/articles/Puget-Systems-Adobe-Photoshop-CC-Benchmark-1132/

 

[RumorConsumer]

For me, right now, the prospect of a new Macbook Pro with 8 cores is truly tempting. Money is OK, not great, but I could work it. The thing I would like to avoid is in 6 months or whatever when the new body style is released and if it ISNT what its cracked up to be its like ok you can get something we like even less which has the performance you want, or go with the best of the last generation which isn't really as good as what's out now but ya know it works. I have a 13" 2017 3.5ghz dual core w 16gb and 512 right now so I'm not suffering too bad but man that power sounds good right now.

 

[William Payne]

I don't know for me all this stuff is just tools for work with personal use on the side. If I was constantly waiting for "The Next big thing" I would never get any work done which would not be good for my wallet/career.
[doublepost=1559007047][/doublepost]

Yes, me, too. It gives a good comparison with my hackintosh 8700K 5.0ghz desktop which scores 1109.

So photoshop users who bought a new i9 MacBook Pro 2019, please take the Puget Sound PS speed test and let us know how you do:

https://www.pugetsystems.com/labs/articles/Puget-Systems-Adobe-Photoshop-CC-Benchmark-1132/

Our of curiosity what are the other specs of your hackintosh?

 

[richinaus]

I am not jumping in by any means immediately, and am awaiting further reviews as you say.

As I have mentioned previously I am really OS agnostic, and use both Windows and Mac and am happy with either for professional work. A friend is lending me a brand new HP Zbook 360 today to see my impressions as that is the machine I would buy as an alternative. So I will see the performance, build etc and then make my mind up. Lisa Glade gave it a great review btw.

The ARM comment is actually interesting and you may be right - I cannot be risking my work to allow them to transition [I went through the PowerPC - Intel fun], however I don't believe they will be doing this on pro machines for at least 5 years. I could be wrong but just cannot see it. If it did happen, and I did get a MBP, I would just leave it in bootcamp and use windows. Very simple and easy for me.

The allure of a eGPU for me is this - I do VR work and a lot of 3D modelling / design testing and it is now all based on the GPU. I also need to have a laptop only solution. So the best of both is laptop and desktop GPU. I am not gaming at all, but is purely a work requirement. You may not think it makes sense in your field of work, but for me there is no laptop I would buy that I would be happy throttling the GPU to 99% on sustained loads, which is what happens.

In reality I rarely max out the CPU so won't throttle that, but the GPU will certainly get it. My iMacpro gets hot and loud under sustained VR / live rendering use. I would prefer that in a separate box tbh. and will still buy a egpu for whatever machine I end up getting, windows or Mac. Sort of irrelevant to the discussion at hand really though.

OK - had my friends HP Zbook 360 for the last hour. It is one of the better PC workstation laptops apparently. Lisa Glade was very positive about it and we all know she is the best laptop reviewer.

I say apparently, as there is absolutely no way will I buy one. You guys [Q6, Maflynn etc] can preach to me all you like about the failures of Apple, but now I am convinced. It is a MacBook Pro for me and I don't care if the keys fail or it throttles a little under heavy load.

I was able to get the HP very hot and fans going instantly with the VR apps. Not a surprise really to be honest.

But the screen was very poor and disappointing. Trackpad average. Very heavy and clunky. I will never use it as a tablet as too cumbersome.

The only good thing about it is it is cheaper than the MBP. Now I know why - it just felt clumsy and un refined.

I would certainly entertain a Surface Book 3 if it was released now, but given end of tax year is here, and it doesn't exist, it will be a MBP for me.

I look forward to reading some more reviews and hitting buy.

 

[OSXconvert]

Our of curiosity what are the other specs of your hackintosh?

They're in detail on the puget sound page, but briefly: 64gb RAM, 1070Ti GPU, EVO 960 M2 drives, i8700k OC'd to 5ghz.

 

[RumorConsumer]

They're in detail on the puget sound page, but briefly: 64gb RAM, 1070Ti GPU, EVO 960 M2 drives, i8700k OC'd to 5ghz.

and it works well, runs mojave?

 

[Thysanoptera]

I say apparently, as there is absolutely no way will I buy one. You guys [Q6, Maflynn etc] can preach to me all you like about the failures of Apple, but now I am convinced. It is a MacBook Pro for me and I don't care if the keys fail or it throttles a little under heavy load.

I saw your previous posts, and since you're into VR and 3d, just to give you a general idea what kind of performance you can expect, since I'm just messing with my eGPU setup - Vega 20 is about equal to RX580 in eGPU. I have Vega FE, and when running as eGPU it is about 50% faster than RX580. On the Windows side, the internal, laptop low power (80W) RTX2060 is about the same as eGPU Vega FE , while RTX2080Ti (eGPU) is twice as fast. You have iMac Pro right? That's still going to be fastest GPU solution for Mac. Not sure about Radeon VII, if the drivers are finally working on Mac - but this baby ain't going to be cool&quiet.

If I was buying eGPU for Mac, it would be Sapphire Pulse Vega 56. It will be slower than the top GPU in iMac Pro, but it is fairly quiet - I have RX580 from them and it is almost dead silent even though it has larger power draw then Vega 56. That being said, if I was into 3d/VR I'm pretty sure I wouldn't ditch the largest manufacturer of graphics cards in the world, that still produces the fastest cards, just because I didn't like one laptop model out of hundreds. If that helps at all, I swore long time ago never to get into HP laptops again and I'm not planning to change my mind, regardless of what some popular youtuber may say.