I feel like in the past few years, processors are remaining as fast but better power efficient, but the end-user never sees longer use out of them. Because manufacturers skimp on battery catching up with power efficiency :-(
Up to 100 Wh / battery, up to 160 Wh with airline approval. All laptops I know of have a 100 Wh battery at most. Macbook pro's used to have that size, but the latest gen has a smaller capacity. You can carry spare batteries, as long as they are for "personal use".
> Later this year, ARM will appear in Windows 10 laptops powered by Qualcomm's Snapdragon 835. The laptops are being called "cellular PCs," which will include smartphone-like capabilities of cellular connectivity and long battery life.
It seems less significant but still noticeable when connected to power than on battery, even when the battery is full charged. IIRC my battery is around 50% design capacity with 1000–1500 cycles.
It would be nice to be able to definitively tell if macOS is underclocking or something.
Since I can't find too much complaints on the web about laptop slowing down and the reply to my yesterday question was that some users don't see much difference, I am now downloading Sierra and will install it on another partition. This should answer this question precisely.
I can't leave High Sierra as is because it made my MacBook Pro unusable.
Running Sierra, I noticed an immediate speed increase on tasks I have done hundreds of times. Everything was snappier again.
Comparing the i7 I have VS the latest i7, there is only about a 25% difference 3 years later. I can say my mbp likely dropped 25% in speed before the battery swap.
I wish I had taken more geekbench diagnostics before sending the laptop to Apple. I did notice several mbp's with the came cpu ad mine had a geekbench score in the 6000's while the spec was in the 8000's. The battery condition may be a factor.
It would be interesting to correlate geekbench (or something comparable) scores to charge cycles. This question is likely worth asking with Google made Android devices that have had battery issues like the Nexus 6p too.
I ordered a replacement battery off Amazon last week.
We just need a truthful answer from Apple when we do not have any evidence for that, or it's hard to get an evidence.
I had the older MacBook Pros (mid 2012 and 2013) and they are getting slower than years ago. So we need to make sure Apple is not doing the slow down like it does to old iPhones.
Apple has a special talent of breaking what was working just fine. While I'm sure some of it is is very worthwhile improvements under the hood that doesn't get the appreciation it deserves, do the pros outweigh a laptop that no longer runs for me?
The latest mbp's while looking great, have lead me to wonder if they are for technical creators/pros anymore..
The cpu is regularly at least a generation behind.
A missing escape key is humorous, while I can remap my escape key to the ~ key if I had a touch bar..
The fact that there's no Apple made dock for a mbp's after a decade of having flawlessly operating docks with HP. Apple display doesn't cut it for everyone as a dock.
Still I continue to use macos because it's the most usable linux/bsd out with respect to available apps. Maybe the Chromebooks will push on this front.
The Nvidia GPU is also on life support on this machine due to a bug which isn't covered under warranty. For that I use Gfxcardstatus and keep the machine's GPU in Integrated mode (the slower Intel CPU). Its a widely known problem with the 2010 series. All apps which use Electron or Chrome put the machine's GPU at boot in Dedicated mode. Which is annoying. Firefox doesn't do this, and since 57 yields good performance. Although I'm not sold on watching videos in Firefox; Chrome/Safari might yield better performance on Mac according to certain anecdotal evidence I found (including my own).
Now, there's a few reasons why newer macOS versions are slower than older. One reason being APFS   (Reddit thread  ). Especially when using FileVault. Keep in mind thats a state of the art MBP; its from 2017.
Regardless, we need benchmarks.
When I first checked earlier this year they weren't in stock, so I asked when they'd be in stock which was soon. Once they were in stock I lacked time IRL so had to postpone order. Hence I only replaced it a good month ago. Had to order some kit and accessoires to open up the machine (was on sale during Black Friday) including the magnetic project mat which you can use to group the screws (you can add notes as well on it). 10/10 recommend that! (I'm not affiliated btw.) I recommend to just check the guides. They're good, and if you need help you can ask questions over there. The people there are much more knowledgable on this subject than I will ever be.
Only bad thing is that the guide told me to not worry about disconnecting the iSight camera. I managed to break that cable. That was during thermal paste application guide though; not battery replacement. The other disconnections of cables all went very well or flawless. Nothing broken in these departments. In a way, no iSight is a plus anyway.
If the older machine doesn't have the same power as newer hardware, it will likely be slower: it's doing more work. Anyone developing a system needs to make a choice: allow the new feature to run on older hardware and accept that it may be slower, or don't allow it to run. They also need to make tradeoffs in what feature sets they make available to platforms (e.g., version-x-for-platform-a, version-x-for-platform-b). Each additional version/platform configuration is additional work to maintain. You may disagree with the choices they make, but they do have to make these choices depending on where they want to distribute their resources as a company.
If the underlying assumption is that Apple is doing this on purpose to make machines obsolete, I think that's mistaken. If OS updates make an older machine run slower, that happens because the update includes new features which requires more resources, and older machines may not have as many resources available.
† At least major versions, and most minor versions.
1) I can't speak for the newer models but many of the older models have an extremely blue monitors without colour profiles, however it's relatively easy to fish the icc colour profile from OS X, then it's just a matter of deciding on how to apply it to your xsession (simplest and lightest is manually with xcalib).
2) Dual graphics... these are a pain and I believe affect most of the newer models also, the solutions differ slightly depending on both the model and the combination of graphics cards. This can be confounded by old nvidia drivers that lack kernel mode setting and give you a nice black screen on ttys. My solution was to adapt my own tiny little utility for controlling the gmux switch (Apple specific, and also slightly model specific), this just basically just provides a switch between integrated and descrete graphics using slightly model specific GPIO ports... Once adapting and compiling that you need to automate calling the switch at various points, I found it was best to initiate the gmux (just by selecting you prefered GPU) when the xsession startsup, (this preserves the tty until X starts, and works around the annoying nvidia mode setting issue for startup at least), than also on wake when the gmux seems to forget itself.
The main reason to do the whole GMUX thing is not necessarily to have choice of GPU (my discrete one doesn't actually work properly with nvidia's linux driver anyway) but to allow it power off the unused one GPU, without this selection you will have an instant cooker and it's really not pleasant to use.
Newer MBPs or those without nvidia cards might not need so much messing around... a single card is always the preference because then you can forgo the whole GMUX nonsense (my model is 10 years old now). The linux kernel actually has some gmux stuff built in, but it didn't seem to work on my model, it may only apply to newer ones.
I switched from 10.6 when OS X was still pretty lean, but even then I noticed many performance improvements... It's difficult to do direct comparison though because I don't run a full DE, just i3wm, the main advantage really is customisability, and being able to run something relatively far lighter. Graphics were noticeably faster though, Although I suspect that's because there was some sort of soft underclocking going on in the OSX drivers (these were the models plagued by the low lead solder micro fractures on the discrete GPUs, which apple "fixed" (read: pushed out of warranty) by underclocking)
Both things are much more configurable under Linux than macOS of course, and you should just be able to set it up once and forget it. (Or keep a note, in case you want to replicate it on another machine...) But the flip side being that it is something you'll have to configure.
Does it bring down power usage in laptops enough to make it efficient?