it's an example bro. my point is you assume dsomething that wasn't my position.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
2017 MacBook Pro is Up to 20% Faster Than Last Year's Model in Benchmarks
- Thread starter MacRumors
- Start date
- Sort by reaction score
Good points...
Except if Apple optimized for the 2016 over time why would you assume they wouldn't do the same for the 2017 Kabylakes over time?
I wouldn't go that far to say I'm 'eagerly' defending it nor do i owe you a good reason really, I barely care abut the topic.
But now that you've engaged me on it: I guess the lack of Magsafe hasnt bothered me, nor the shorter changing cable length they supply. I haven't tripped over it yet and the reduced cabling footprint has been nice when traveling. You can always get this if you dont want to leash the kids: https://www.amazon.com/Griffin-Brea...6954199&sr=8-1&keywords=usb+c+magsafe+adapter
As an Amazon Associate, MacRumors earns a commission from qualifying purchases made through links in this post.
Taken at face value, sure makes sense. However, Skylake brought along a new CPU architecture for Apple to optimize in moving from the old Haswell processors of the 2015 model. Kaby Lake uses the same architecture as Skylake, but is clocked faster. That's not to say Apple can't optimize anything, but there's not going to be nearly the same increases that were seen with Skylake optimizations.
FACT: 2016 base model is 93.5% (single core) and 91.3% (multi core) as fast as the 2017 model, benchmark wise.
The GeekBench results takes the SSD into consideration as well, which widens the results gap even further.
IMHO, the SSD won't make a big difference in real life as the 2016 is blazing fast already.
I am personally happy that I had the base 2016 15'' for these months.
The GeekBench results takes the SSD into consideration as well, which widens the results gap even further.
IMHO, the SSD won't make a big difference in real life as the 2016 is blazing fast already.
I am personally happy that I had the base 2016 15'' for these months.
I got curious and calculated some numbers myself. I took the 23 most recent geekbench runs from the high-end stock Skylake 2.7 and Kaby Lake 2.9 models. I used 23 because that’s all the new 2017 model has currently. Then based on multi-core score, I used the middle 11 results (middle ~50%) and threw out the outliers, the lowest and highest 6 from each to be exact. The average of these 11 results gave me the following.
Skylake
Multi: 14605
Single: 4273
Kaby Lake
Multi: 15710 (7.6% increase)
Single: 4588 (7.4% increase)
Between Skylake and Kaby Lake, multi-core clock speed increased by 7.4% and single-core Turbo Boost clock speed increased by 8.3%. As can be seen, Kaby Lake is very slightly outperforming where it should be based on a clock speed increase alone for multi-core and slightly underperforming for single-core. It also shows that we’re only a little over a third of the amount of an increase the article so boldly proclaims.
On the contrary, a lot of people think those flat butterfly switches are great. I certainly do and I have RSI from years of typing too much and too fast. The butterfly switches are surprisingly comfortable and pain free for me to use. That said, everyone's RSI is different.
Now, lemme say that I love my keyboards. I don't really have a clear favorite. I like many of them and I own a lot of them. I have a mechanical keyboard in just about every switch type you can find including Kailh, Gaterons, original ALPS, reproduction ALPS, and Topres.
I love all of those keyboards and I'm also very satisified with Apple's keyboards. Just because they're not true mechanical switches doesn't mean that it's junk. I actually quite like the Apple Bluetooth keyboards or any of the flat aluminium series too. Crazy, I know, but once I got over my hangup about them not being mechanical, I learned that I quite liked them.
I still prefer my mechanicals, but the Apple keyboards are more than acceptable.
My mid-2012 2.6 Ghz Macbook Pro gets the following scores:
Single-Core: 3696. Multi-Core: 12459
The new 2017 2.8 Ghz system is less than 15% faster than my 5 year old machine.. I'm having a hard time at the moment justifying an upgrade - especially considering that the Pro model is missing ports that I use on an every day basis. Are my calculations off?
Single-Core: 3696. Multi-Core: 12459
The new 2017 2.8 Ghz system is less than 15% faster than my 5 year old machine.. I'm having a hard time at the moment justifying an upgrade - especially considering that the Pro model is missing ports that I use on an every day basis. Are my calculations off?
How are you working out that you benefit less than 15% increase? I can’t see how you arrive at percentage...
Are you using the same version of Geekbench?
Also, if you machine is fine then don’t upgrade. The 500 nits brightness, faster SSD, Touch ID/Touch Bar and smaller foot print are all luxury features / nice to haves. If your machine fills all your needs then there’s no point.
Yes, I'm using the same version of Geekbench (4.1.0 tryout). I compared the results at http://browser.geekbench.com/v4/cpu/3038252 showing
Single Core: 4255. Multi-Core: 13727
The single core score is actually 15.12% higher, and multi-core is only 10.18% higher. I guess I was just expecting more of a change 5 years.
From what I understand (based on what a friend told me who is usually in the know on such stuff), the big deal about the move to Kaby Lake is that the TB3 implementation is on-chip, and not bolted on like the previous gen MacBook Pros. And, I guess that would dramatically impact battery life when plugging in anything to the ports. (i.e.: much better battery life under that circumstance now than previously.)
I just ran a Geekbench Open CL Compute test on my 2017 $2799 MBP and got 55644. A friend has a 2016 MBP at the same tier and got 41701.
https://browser.geekbench.com/v4/compute/810282
https://browser.geekbench.com/v4/compute/810282
That's interesting, a quick search of some of the recent runs on the high-end stock 2016 model shows results between 46721 and 51842. Your friend's seems to be underperforming for some reason.
https://browser.geekbench.com/v4/compute/search?utf8=✓&q=platform:"Mac"+processor:"Intel+Core+i7-6820HQ"+frequency:2700+bits:64
Actually, I forgot that the 460 was a BTO option on the high-end base last year, which are the values I was using. Your friend's seems to be in line with other 455 results, my mistake.
https://browser.geekbench.com/v4/compute/810297
[doublepost=1496976338][/doublepost]
So, roughly 25% increase there when comparing the 2016 $2799 machine to the 2017 $2799 machine?
Yeah that seems to be what the numbers say. I wonder why you got such a high result compared to others with the 560? In the link below, most others were getting in the low 40,000's with the 560. Your result actually makes more sense though since models with the 460 were between 46721 and 51842.
https://browser.geekbench.com/v4/co..."Intel+Core+i7-7820HQ"+frequency:2900+bits:64
I ran it again and got 56432 this time
[doublepost=1496977208][/doublepost]I could tell even before benchmarking that there was a significant video card difference by the way the 2017 machine handles my 4K display.
Attachments
Those with low-40,000 results on the 560 could have been doing other things on their computer while benchmarking. 55k-56k sounds much more likely.
IMHO, anyone who bought the high-end 2016 model, but opted not to get the 460 was a bit foolish. This was a ~$3k computer and it was only a $100 upgrade and the performance gain was rather significant, as these benchmarks show.
I'm in the same boat as you, i have a maxed out 2012 rMBP, and I'm holding off till next gen.
The only significant thing you get right now is:
SSD speed (3000gb/s+, over 500 gb/s)
Video RAM (if you use external)
Better GPU
You don't get (if you're maxed out):
More RAM
Significantly more CPU power
I went back to Windows on my desktop and Windows 10 has been very good. I've come to the conclusion I need two laptops. 1 very thin and a thicker, workstation class laptop for when I do need to run a load of VM's on the move. I already bought the surface Pro 4 (massive discount on them at the moment as the new one is just about to come out) and that is brilliant. I was hoping Apple would increase the spec of the MBP, but nope so I'll just forget having a Mac entirely and look at the Lenovo P51 - it's a bit thicker, but only weighs 2.5 KG, which isn't too bad.
So add a fan or a heatsink, they make it work in the MacBooks and iMacs.
Or Make a 32" screen iMac.
I'd just like to use a large screen with an I7 4.2..that would work for me!