Not a single review mentions performance gains for those coming from 2016, 2017 models to the 16". Anyone go from 2016 to 16"?

[fs454]

Has anyone found any concrete performance comparisons (ideally real-world and not just Geekbench numbers) between the older 15" models and the 16"? Every review only seems to mention the outgoing 2018 model when it's far more likely that tons of us on 2016/2017 and even pre-butterfly 2015 models want to know what we're getting for our cash if this is the point where we're looking to upgrade.

I'm specifically looking for anyone who's upgraded from a 2016 model. I've got a loaded 15" 2.9ghz 4core i7, 16GB RAM, Radeon Pro 460 4GB, 1TB model and am trying to make the call whether this jump is significant enough to noticeably improve overall day-to-day performance as well as Premiere Pro, Photoshop, etc. tasks. I know FCPX is improved pretty significantly but 95% of my work is in the adobe suite and nobody in my professional space is using it.


Thanks all.

 

[rdowty]

I noticed the same thing. I agree that the typical person, even a power user, probably isn't going to be upgrading from a maxed out 2019 15" MBP with a Vega 20. I'm assuming that since there seems to be a lot of improvements there the difference coming from my 2017 with it's 4-core CPU is going to be quite noticeable. Its too bad there aren't any 10th gen CPU's from Intel right now because I'm going to spend the extra money for 32GB of RAM this time so I'll want to keep it more than a year and I know when those CPU's are available that will likely be big upgrade as well.

 

[kave]

Same here, I am quite happy with my 2016 but considers an upgrade. Geekbench single core numbers is not a huge difference to the 2019.
Barefeats.com has done some real world tests but between the two i9 models.

 

[yonafunu]

I have exactly the same setup as yours (2016, 460, 1to, i7 2.9ghz and 16go). I have made the jump and I’ve bought the 16“ (not in my hand yet). The 5500m 8gb ( even the 5300m in fact) is maybe 5 or 6times more powerful than our current gpu. And I need 32gb of ram for my database processing. Of course you can always wait for a better MacBook , but there will always be a new one more powerful waiting in the corner.
For the CPU I think my new i9 will not be so much more powerful, the gap is lesser important than the gpu, but still there.

 

[danwells]

The CPU will be at least very close to double (it should be 1.7x at a minimum in a reasonably threaded application).

The GPU is several times as fast - it looks like ~2x on average (3DMark on Notebookcheck), but more than that in certain benchmarks.

The RAM (if upgraded) could be huge in RAM-intensive applications - or could mean nothing in low RAM applications.

Twice as fast overall?

 

[Fravin]

The GPU is several times as fast - it looks like ~2x on average (3DMark on Notebookcheck), but more than that in certain benchmarks.

Even the 5300 runs very hot. The 460 or 560 chips are much cooler.

In fact what will set the performance is how heat will be handled. In some reviews people are saying that the new 16” has a better Thermal Solution that can keep performance up while controlling the heat generated for more time.
This could look attractive but I don’t think so.

Look, heat is generated by CPU. The CPU controls how much power is delivered to the chip by processing load. So when a higher load is detected, more power is delivered to the CPUs, which start to run all cores at a higher clock to handle the job.
While it gets to a higher clock, more heat is generated. When the CPU gets closer to 100c the power is reduced and so the clock of the cores. This is preserved until the temperature is controlled. Then it gets back to higher clocks and more power... And the things go on and on...

The key is how much heat the whole system generates (inside these tiny space). Understand that the same heat pipe goes trough CPU and GPU. So the amount of heat generated by one affect the way the other will handle higher temperatures.

It’s a complicated thing.

If you let the MacBook handle the temperature and power consumption alone, it will always prefer iGPU to prevent heat.
It’s a clever solution by the way as much users don’t mind if they need or not a GPU.

 

[baypharm]

I have been using a 2016 MBP i7 with 2TB SSD in my work along with an iPad Pro. With the 2016 MBP I can process, for example, six 96 well plates in about 30 seconds - running Probabilistic Mixture, Fragment and Variant DNA analysis. I can process about 2000 sequencing files (.ab1, ESD, and .scf) in approximately 20 minutes. This is under a Windows environment on my MBP, by the way. I just got my 16" in yesterday (fully loaded). I ran some preliminary tests with it today using software for whole human exome sequencing and it performed well. I was able to process about 1000 samples looking for variants by chromosomal position only in about 12 minutes. Not sure at this point in time if I will keep it. There are other issues unrelated to this thread that make me suspect. But to answer your question specifically, yes, it is speedier - by how much? I can't say for sure at this stage. But I would recommend you purchasing a new one and trying it out for yourself. Everyone has different uses and expectations.

 

[Eee1111]

I had the 2016 2.9ghz i7 16gb mbp.
The 2.3ghz i9 32gb 5500m is a step up and it’s obvious.

Go for it! It’s a beast so far.

my time with the i9 shows me that it gets a lil toasty fast compared to that old i7. But the laptop as a whole doesn’t get as hot as the 2016 mbp did. The 2016 mbp got ****ing hot near the top where the oled strip. The i9 can jump around it’s temps but the metal doesn’t get hot like the 2016 did.

if that helps.

 

[matram]

I have gone from a 2016, 4-core 2.9 GHz, 16 GB, 2 TB, to the 16", 8-core 2.4 GHz, 32 GB.

I have run some comparisons on my workload which is Visual Studio Windows Development in VMWare Fusion. These are the results for 2019 / 2016

Full solution rebuild: 55s / 75s
Incremental build: 5s / 7s
ReSharper solution inspect: 71s / 95s

This matches fairly accurately the Geekbench single-core results that I posted in the Geekbench thread.
2016: 855 / 3609
2019: 1184 / 7361
Or +38% single-core and +103% multi-core

A couple of points to note.

• Fusion has been assigned the same 4 vCPU in both cases, so there is no added benefits from the 8-core at this point. I will play around with that a bit more when I have time
• The solution being built / analyzed is comprised of 233 source files and 22.049 LOC