what performance test does MBP2016 TB fail to catch up MBP 2015?

[jjiangweilan]

Many people say this is not a professional machines, but what performance test does it fail? Decreasing from 32GB to 16GB is a decreasing on number, but what about the reality on performance compared to these previous professional Macs?
[doublepost=1481916103][/doublepost]this video editing test says everything is better or the same.

what is the performance these people talk about...

 

[mikeo007]

I don't believe there's ever been a Macbook capable of supporting 32gb of RAM, has there?
The only performance people cite as being worse is synthetic benchmarks. I think we've seen dozens of examples of real-world usage always putting the 2016 at par or better.

 

[Naimfan]

If you look at Geekbench results, you'll see the top Mid-2015 is noticeably faster than the 2016. People debate how much meaning that has, but there you are.

 

[jjiangweilan]

If you look at Geekbench results, you'll see the top Mid-2015 is noticeably faster than the 2016. People debate how much meaning that has, but there you are.

that doesn't make sense. We know Apple is always working on integrating software and hardware. But Benchmark only test your hardware, right? That's why you cannot happily use flash on MacOS.

 

[linguist]

that doesn't make sense. We know Apple is always working on integrating software and hardware. But Benchmark only test your hardware, right? That's why you cannot happily use flash on MacOS.

yes and no.


CPU is made to throttle at lower capacity on 2016 , to achieve less heat, less fan noise, less GPU throttling ( CPU and GPU share same heatsink and power draw is regulated ), it is controlled compromise.

on application that use GPU more 2016 wins hands down, CPU only application 2015 wins.

 

[Creep89]

Coming from a 15" rMBP Late 2013 with a 4960HQ I can definitely say that I made the wrong choice opting for the 6820HQ. Should have gone for the 6920HQ. The 6820HQ is slower than the 4960HQ. So calculating a lot of stuff it's 0.47s vs 0.51s on average. This may sound not so much of a difference but with a lot of data every ms counts.

 

[leman]

As I haven't received our new units yet, I can't benchmark them yet (I certainly will!). But from what I have seen so far, the older Haswell CPUs have higher potential burst performance, while Skylakes are much better at sustained performance because of their better thermal characteristics. Basically, if your workflow involves short (10-20 seconds) CPU intensive work, you will most likely see better figures with the 2015 machine. However, Skylake will do better both at microbursts (e.g. browser/office) — because it can change clocks faster, and at sustained performance — because it can maintain high speed for way longer without overheating.

 

[daveak]

Complaints. Lower is better.

 

[Jaekae]

If you look at Geekbench results, you'll see the top Mid-2015 is noticeably faster than the 2016. People debate how much meaning that has, but there you are.

Maby look at the turbo speed of the processors and you see why, skylake have bit lower maximum turbo speed but it can hold on to that higher speed for a longer time than the old processors. Geekbench dont run for so long, which is why they can score higher on that but the skylake beat the old one in real life tasks all the time because maintained speed and better thermal management.

 

[Charlesje]

The 2,9 2016 macbook pro - tested it, and seeing this myself on a daily basis - performs around 15 to 20 % faster cpu wise than the previous generations maxed out versions. In fact previous generations measured mostly on par when considering cpu performance because macbooks suffered greatly under throttling. Skylake brings a noticable change to this since long, contrary to what some (too much) people on macrumors are stating. Please benchmark the machine or look up cinebench results (or other real world cpu tests). And please dont make statements solely on the basis of some one elses statements. It's confusing and creates quite an infinite complaining loop! We should complain about other things eg the choice for a meager memory capacity