Macbook Pro 2019 Benchmarks and First Impressions! Post yours!

[Queen6]

This has been the case since at least Ivy Bridge... all MBPs I've tested in last 6+ years run close to 100C under load. And talking about thermals, the current ones have better thermal performance than the 2015 chassis. It wouldn't exactly choke with Coffee Lake, since it was perfectly capable of dissipating 45Watts, but it would be in more trouble when CPU+GPU performance is considered.

Cooling solution is inadequate to see optimal performance simple as that. Mind it's not terrible for such a diminutive chassis, nor do the competitors fair any better in the same form factor, likely worse. As for 45W TDP an illusion, Intel is reasonably clear in the details.

Apple's failing is being overly "grippy" 9th Gen i9, 32Gb RAM, Vega 20, 1 TB SSD for the high tier 15" with 3 year warranty should be the default, for me this would equal Sale. Personally I'm of the opinion that Apple should serve me, not vice versa...

Q-6

 

[leman]

What's your take on the current thermal situation? Does Apple have a choice in front of itself or are we actually in good shape with the current chassis?

I'd say that it depends on what you want. Strictly speaking, every MBP matches the spec, since it meets (and exceeds) the CPU manufacturer's guaranteed performance levels. At the same time, Coffee Lake can perform well beyond its guaranteed levels and larger laptops that give the CPU more thermal headroom (i.e. run it more like a desktop part) do get better performance. Queen6's above post sums this up nicely (even though I disagree with him on the conclusion).

Personally, I am fine with the status quo. The MBP is currently just around 30% or so (regarding CPU) slower than some of the fastest (and bulkiest) laptops on the market. I believe this is an OK tradeoff. I would not want to trade the mobility and versatility of my MBP for more performance. At the same time, I do understand that for many users performance is the most important criterion (although I am still confused why these users are looking at Apple products, they were never the first choice when raw performance is considered).

All that said, I am convinced that Apple could potentially improve performance without too much compromise. Their cooling solution is quite good (it is extremely efficient relative to its size), but many people that repasting the CPU yields noticeable performance benefits. This suggests that there are inefficiencies on the assembly line. If Apple wanted, they could come up with a more technically involved solution (e.g. going as far as soldering the heat transmitters to the CPU directly) without sacrificing the laptop form factor or weight. After all, Mac high price gives Apple more room to play with advanced tech (which they do anyway). I remain cautiously optimistic that Apple will increase their efforts in this area, since for a while now, it seems that they are getting more serious about the prosumer user (the fact that they brought Vega to the MBP, some Mojave features and now the fact that they are among the first laptop manufacturer to adopt Coffee Lake Refresh CPU, even using CPUs that are not formally announced yet).

 

[spiderpumpkin]

This guy does a really good job comparing the 2018 15" i7 and 2019 15" i9.

 

[RumorConsumer]

As for 45W TDP an illusion,

What’s that mean exactly

 

[Thysanoptera]

This guy does a really good job comparing the 2018 15" i7 and 2019 15" i9.

If I had his 2018 I would take it back to Apple and exchange. 900 in CB and barely holds 45W. Theoretically within specs but exact same model can do 1100 and hold 55W - weird he didn't notice this.

 

[leman]

What’s that mean exactly

Coffee Lake CPUs are configured in such way that they will comfortably run above their stated design power of 45Watts. This is why larger laptops see higher performance, they often run the CPU at 55W or even higher. While you could also do it with earlier CPUs, they would quickly hit their clock limit. But Coffee Lake's clock limit is so high that, for example, you would need to run an i9 at well above 100Watts for it to hit the limit. To put it somewhat bluntly, Intel took a desktop CPU and sold it as a laptop CPU.

 

[maflynn]

Their cooling solution is quite good

All in all, I agree with most of your post, but I do disagree on this comment.

Edit:
I had provided an example of the cooling on thinkpad, but I left the power slider set to better battery thus providing results that are not applicable to be compared to the MBP. My apologies for mistating the cooling/performance, it was not my intent to deceive.


Image removed
View attachment 838847

 

[Queen6]

If I'm paying a premium I fully expect performance to be in place, I don't want to hear excuses...

8th Gen i7 8750H

This OEM puts performance first and delivers...

Q-6

 

[Thysanoptera]

All in all, I agree with most of your post, but I do disagree on this comment.

TBH the cooling of the current MBP is pretty amazing for its size, just looking at it one would never expect it being able to get rid of ~55W of heat on continuous basis. And its noise characteristic is just, don't know how to put it, but somewhat pleasing, even at full blast it doesn't annoy me. Of course, when it works correctly, the guy in the Youtube video above couldn't go above 45W. Just makes you wonder, what could be achieved if they scaled this by 20% by adding a millimeter or two to the thickness.

And on your screenshot the CPU runs at 25W, Sunday afternoon walk in the park. Did you have some ThrottleStop profile running?

 

[William Payne]

99% of my needs are photoshop. I recently found out that Puget Systems have a photoshop benchmark. I ran it on my 2018 MacBook Pro with i7 2.6 6 core with 32GB of ram and Vega 20. I got a score in the 780’s overall. Puget Systems said that that was really good for a laptop compared to the high spec desktop machine they used as a baseline.

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

 

[leman]

My laptop is not that much bulkier (something like 3 millimeters thicker) but it performs faster and its cooling is excellent.

Yeah, we talked about it already. Your laptop runs its airflow though the holes in the bottom chassis, this way the fans can suck in maximal possible amount of air. That is the design used by most laptops in this segment, since it is very simple and efficient way to maximise heat exchange. An obvious drawback is that you have to make sure that your intake vents are not occluded (and also, the bottom of the laptop can get very hot with this design, which is not optimal for lacquered wooden desks).

MacBook Pro on the other hand performs its air intake through inlet vents located on the sides of the laptop and its bottom is solid. This means that you can use the laptop in different environments, such as on your knees or on a blanket (like I am doing right now) with minimal risk of blocking your vents. But the drawback of this design is a) the air inlet vents are obviously smaller so you get less airflow and b) the fans are quite far from the inlet vents, which reduces the airflow efficiency. Essentially, MBP relies on creating a negative pressure inside the laptop which causes the air to flow though the sides and the already very cramped internals.

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.

 

[maflynn]

And on your screenshot the CPU runs at 25W, Sunday afternoon walk in the park. Did you have some ThrottleStop profile running?

Not knowingly, and yeah I saw that as well. The intel power gadget shows 55 watts and the package pwr at 25. Maybe I do have a setting ¯\_(ツ)_/¯

I'm not about to mess with throttestop while that process is chugging away
[doublepost=1558737713][/doublepost]I found what it was, I mistakenly left better battery turned on, and that limited the power. To that end, my point is invalidated and I am wrong. I'll edit my post, as it has misinformation.

 

[Thysanoptera]

I found what it was

It happens, at least you didn't make a youtube video out of it containing comments of thousand people agreeing with your findings and being cited by major tech outlets

 

[RumorConsumer]

Not knowingly, and yeah I saw that as well. The intel power gadget shows 55 watts and the package pwr at 25. Maybe I do have a setting ¯\_(ツ)_/¯

I'm not about to mess with throttestop while that process is chugging away
[doublepost=1558737713][/doublepost]I found what it was, I mistakenly left better battery turned on, and that limited the power. To that end, my point is invalidated and I am wrong. I'll edit my post, as it has misinformation.

Wow this thread is at least 65W

 

[RumorConsumer]

A sample 15" 2018 2.9Ghz x 32gb running a Handbrake video conversion for about 8 minutes at this point:

[doublepost=1558750279][/doublepost]

A sample 15" 2018 2.9Ghz x 32gb running a Handbrake video conversion for about 8 minutes at this point:

And then after 10.14.5 w Supplemental Update, but possibly a different codec so different properties and activating differing aspects of the hardware:

 

[MacRS4]

Oh wow good catch. Hmmmmm. I hate to ask about the 10.14.5 Supplemental update and whether it was applied...

It was applied, yes. More general review video will be up Monday/Tuesday. Quick summary - generally pretty good, seems terrible in BootCamp.

 

[RumorConsumer]

This guy does a really good job comparing the 2018 15" i7 and 2019 15" i9.

This is awesome

 

[RumorConsumer]

It was applied, yes. More general review video will be up Monday/Tuesday. Quick summary - generally pretty good, seems terrible in BootCamp.

Normal fit and finish aside anything you look at and say eh I’d be fine w the 2018?

 

[OVERTASK]

I'd be curious to see the difference between 2018 i9 vs 2019 with HTT disabled (thanks Intel)

 

[leman]

I'd be curious to see the difference between 2018 i9 vs 2019 with HTT disabled (thanks Intel)

What’s HTT?

 

[Howard2k]

Hyperthreading.

 

[ashcairo]

I'd be curious to see the difference between 2018 i9 vs 2019 with HTT disabled (thanks Intel)

You lose 300 points on Cinebench, but everything else is about the same. Some tasks like Crypto mining run faster. I recently did a video about it giving examples for the 2018 MacBook Pro, and just finished some perf tests on the 2019 which I'll have up soon.

 

[OVERTASK]

You lose 300 points on Cinebench, but everything else is about the same. Some tasks like Crypto mining run faster. I recently did a video about it giving examples for the 2018 MacBook Pro, and just finished some perf tests on the 2019 which I'll have up soon.

2 physical cores should technically give a noticeable bump in perf, dare I say I prefer (not happy though) having HTT disabled. Though I'm wondering how non-linear perf increase would be from throttling.

 

[leman]

2 physical cores should technically give a noticeable bump in perf, dare I say I prefer (not happy though) having HTT disabled. Though I'm wondering how non-linear perf increase would be from throttling.

What reason would you have for disabling hyper-threading?

 

[maflynn]

What reason would you have for disabling hyper-threading?

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