Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
iOS benefits very much from multi-core. Apple designs its own chips after all and they’re all multi core.

Safari also happens to be one of the fastest browsers out there too.

Some things benefit a lot, others little or nothing, but please let's not turn this into a dispute
 
You’re the one claiming that iOS browsing doesn’t benefit much from multi-core. However, that’s simply not accurate.
I am and I maintain what I said. I have done quite a lot of tests with multiple devices (A8, A8X, A9X, A12 etc.). People that want to know who is more accurate will do their own tests.
 
I am and I maintain what I said. I have done quite a lot of tests with multiple devices (A8, A8X, A9X, A12 etc.). People that want to know who is more accurate will do their own tests.

Putting my nose where it may not belong... all web browsers, regardless of who makes them, are at their core single threaded. Most browsers have moved to a one-thread-one-tab model which gives the appearance that they're faster since things will load in background tabs and be ready when you want them, but the actual downloading and rendering of content into a web browser is single threaded which is why Safari rendering benchmarks are basically comparable to every other web browser out there.

Safari operates on a main UI thread that renders the actual content to the screen. This, in turn, spins off child threads for each tab (but all but the visible one operate on a background priority). When the child thread has finished it's loading and rendering of content, it's then passed back to the UI thread for screen rendering. Again, it's the illusion of speed in a web browser. The UI never blocks, content requested in the background is there when you need want it, scrolling the window feels snappy (because it's handled by the UI thread). But everything that happens inside the window is single threaded. This is still the big blocker for JavaScript engines. JavaScript runs on the same thread as everything else in its tab and has to share its resources and interruptions from the UI.

So, when it comes to browser benchmarks, it's single core performance that matters most.
 
Putting my nose where it may not belong... all web browsers, regardless of who makes them, are at their core single threaded. Most browsers have moved to a one-thread-one-tab model which gives the appearance that they're faster since things will load in background tabs and be ready when you want them, but the actual downloading and rendering of content into a web browser is single threaded which is why Safari rendering benchmarks are basically comparable to every other web browser out there.

Safari operates on a main UI thread that renders the actual content to the screen. This, in turn, spins off child threads for each tab (but all but the visible one operate on a background priority). When the child thread has finished it's loading and rendering of content, it's then passed back to the UI thread for screen rendering. Again, it's the illusion of speed in a web browser. The UI never blocks, content requested in the background is there when you need want it, scrolling the window feels snappy (because it's handled by the UI thread). But everything that happens inside the window is single threaded. This is still the big blocker for JavaScript engines. JavaScript runs on the same thread as everything else in its tab and has to share its resources and interruptions from the UI.

So, when it comes to browser benchmarks, it's single core performance that matters most.
That would make sense. That also explains why A8X feels significantly faster than A8 even for browsing, although I may not agree with the term “illusion” in this context.
[doublepost=1555612278][/doublepost]
I am and I maintain what I said. I have done quite a lot of tests with multiple devices (A8, A8X, A9X, A12 etc.). People that want to know who is more accurate will do their own tests.
We will have to agree to disagree if you don’t think the iPad Air 2 feels faster than the iPad mini 4 in Safari.

I’m not talking about just running benchmarks.
 
Last edited:
That would make sense. That also explains why A8X feels significantly faster than A8 even for browsing, although I may not agree with the term “illusion” in this context.
[doublepost=1555612278][/doublepost]
We will have to agree to disagree if you don’t think the iPad Air 2 feels faster than the iPad mini 4 in Safari.

I’m not talking about just running benchmarks.

What spudworks said is pretty accurate. We do not need to disagree on "faster" but we do need to disagree on significantly in web broswing. I was not talking about benchmarks either (I have mini 4, air 2, and first gen pros). A8X feels slightly faster than A8 in browsing, but has a much higher multicore since it has 3 cores. A9X (I consider the 9.7 pro, to have equal ram, but I have both the 9.7 and the 12.9) is only dual core but with much higher clock. Multicore is only slightly more than ipad air 2 for ipad pro 9.7. So based on multicore benchmarks, ipad air 2 should be much faster than ipad mini 4 and close to ipad pro 9.7. Truth? Ipad air 2 is only slightly faster (like less the a second usually) than ipad mini 4 (more in line with the slightly higher single core performance) and ipad pro 9.7 is much faster (like several seconds) than ipad air 2 (again in line with the much higher single core). This both in Safari and in Chrome (which is not suprising given that they all use the same engine). Which again empirically proves what has been true for a long time and still is and what spudworks explained
 
What spudworks said is pretty accurate. We do not need to disagree on "faster" but we do need to disagree on significantly in web broswing. I was not talking about benchmarks either (I have mini 4, air 2, and first gen pros). A8X feels slightly faster than A8 in browsing, but has a much higher multicore since it has 3 cores. A9X (I consider the 9.7 pro, to have equal ram, but I have both the 9.7 and the 12.9) is only dual core but with much higher clock. Multicore is only slightly more than ipad air 2 for ipad pro 9.7. So based on multicore benchmarks, ipad air 2 should be much faster than ipad mini 4 and close to ipad pro 9.7. Truth? Ipad air 2 is only slightly faster (like less the a second usually) than ipad mini 4 (more in line with the slightly higher single core performance) and ipad pro 9.7 is much faster (like several seconds) than ipad air 2 (again in line with the much higher single core). This both in Safari and in Chrome (which is not suprising given that they all use the same engine). Which again empirically proves what has been true for a long time and still is and what spudworks explained
Well, perhaps if you're talking about single page rendering. But many of us like to multi-tab (and multi-task). It's smoother on A8X than A8. I can't comment on A9X vs A9.
 
Well, perhaps if you're talking about single page rendering. But many of us like to multi-tab (and multi-task). It's smoother on A8X than A8. I can't comment on A9X vs A9.
I am talking about both. A8X is slightly faster/smother than A8. But the dual core A9X is much faster than 3-core A8X. Trust me. Multi-tab on IOS does not work like on Windows, so it's not the same kind of multitasking (one could even argue it's not even true multitasking). But that's another story... Split screen is indeed true multitasking, but that would mean using 2 different browsers (might change in ios 13).
 
I am talking about both. A8X is slightly faster/smother than A8. But the dual core A9X is much faster than 3-core A8X. Trust me. Multi-tab on IOS does not work like on Windows, so it's not the same kind of multitasking (one could even argue it's not even true multitasking). But that's another story... Split screen is indeed true multitasking, but that would mean using 2 different browsers (might change in ios 13).
That’s not accurate. You can do side-by-side Split View in Safari right now. That’s been a feature of Safari since iOS 10 I believe.
 
That’s not accurate. You can do side-by-side Split View in Safari right now. That’s been a feature of Safari since iOS 10 I believe.
Yes you are right, you can do that in Safari (I tend to use other browsers too, especially Chrome, that's why I had forgot) but it's the same principle as using 2 browsers. And it's different from multi-tab and it's limited to 2 (not counting picture in picture). And I doubt it's what people use most of the time (contrary to multi-tab). Hopefully with the new windowing in ios 13 we can have a form of multitasking that is closer to Windows. Then we might see a significant difference in chips with more cores, based on how this is implemented in IOS, that is on how may pages can stay active at the same time.
 
I am and I maintain what I said. I have done quite a lot of tests with multiple devices (A8, A8X, A9X, A12 etc.). People that want to know who is more accurate will do their own tests.
Caveat. Not many people actually own or have access to all of those chipsets to be able to make objective comparisons. :p

Mind, one thing that often gets left out when discussing general performance is the GPU. It's not just gaming, GPU has pretty significant impact on normal usage as well.
 
Caveat. Not many people actually own or have access to all of those chipsets to be able to make objective comparisons. :p

Mind, one thing that often gets left out when discussing general performance is the GPU. It's not just gaming, GPU has pretty significant impact on normal usage as well.

True ;-) I have had 8-9 ipads since the first air and still have 5 and use all of them... I use my smartphone very little.
As for GPUs, while I know the subject well for PCs, it's something I know less about on Ipads, other than through benchmarks, as I don't know how to isolate them in real life tests. I have read a lot and there are a lot of conflicting views. Maybe someone more expert on the subject can tell us more (outside gaming).
 
True ;-) I have had 8-9 ipads since the first air and still have 5 and use all of them... I use my smartphone very little.
Ran some informal side by side tests. The Apple A9 (the one on the cheapie 2017 iPad 5th gen, not the X version on the Pro) was actually noticeably faster than A8X in the tasks I tried:
  • displaying PDF pages in GoodReader (technical manuals)
  • web page rendering
  • launching YouTube app and browsing videos
  • searching for text in a 2MB ebook
  • launching Excel and loading a (very simple!) spreadsheet
  • app installation (Gmail)

Last three was where the gap was most noticeable. I could check email already on the 5th gen before Gmail even finishes installing on the Air 2. Excel took more than 10s to launch on A8X vs 4-5s on A9 and opening spreadsheets and switching tabs was noticeably slower on A8X. Sometimes selecting cells, too.

A8X and A9 are evenly matched when it comes to multi-core. It's single-core where A9 pulls ahead (by ~50% on average, iirc). I think A9 is also when they switched to LPDDR4 and PCI express flash storage (previous gen was eMMC, I believe). Really, there's so many moving parts and more likely than not, performance improvement is a symbiosis of all these upgrades (chipset, memory, storage, networking).

Mind, for iOS 7+ UI eye candy, it was easy enough to pinpoint graphics performance as the culprit at least with the following models.

UI lag (1-worst to 4-best)
  1. iPad 3 (A5X/1GB, 2048*1536)
  2. iPad 2 (A5/512MB, 1024*768)
  3. iPad 4 (A6X/1GB, 2048*1536)
  4. iPad Air (A7/1GB, 2048*1536)
Same CPU perf on iPad 2 & 3 but 3 was still slower. Same RAM and resolution on iPad 3, 4 & Air.
 
Ran some informal side by side tests. The Apple A9 (the one on the cheapie 2017 iPad 5th gen, not the X version on the Pro) was actually noticeably faster than A8X in the tasks I tried:
  • displaying PDF pages in GoodReader (technical manuals)
  • web page rendering
  • launching YouTube app and browsing videos
  • searching for text in a 2MB ebook
  • launching Excel and loading a (very simple!) spreadsheet
  • app installation (Gmail)

Last three was where the gap was most noticeable. I could check email already on the 5th gen before Gmail even finishes installing on the Air 2. Excel took more than 10s to launch on A8X vs 4-5s on A9 and opening spreadsheets and switching tabs was noticeably slower on A8X. Sometimes selecting cells, too.

A8X and A9 are evenly matched when it comes to multi-core. It's single-core where A9 pulls ahead (by ~50% on average, iirc). I think A9 is also when they switched to LPDDR4 and PCI express flash storage (previous gen was eMMC, I believe). Really, there's so many moving parts and more likely than not, performance improvement is a symbiosis of all these upgrades (chipset, memory, storage, networking).

Mind, for iOS 7+ UI eye candy, it was easy enough to pinpoint graphics performance as the culprit at least with the following models.

UI lag (1-worst to 4-best)
  1. iPad 3 (A5X/1GB, 2048*1536)
  2. iPad 2 (A5/512MB, 1024*768)
  3. iPad 4 (A6X/1GB, 2048*1536)
  4. iPad Air (A7/1GB, 2048*1536)
Same CPU perf on iPad 2 & 3 but 3 was still slower. Same RAM and resolution on iPad 3, 4 & Air.
I have given up trying to figure out what’s going on with Excel. It usually launches faster than that on my iPad Air 2 but then again sometimes it has taken about that long on my 2017 Core i5-7600 iMac with 1 TB SSD and 24 GB RAM.

My 2009 8 GB Core 2 Duo 2.26 GHz MacBook Pro with an older version of Office launches Excel faster than that 2017 iMac.

According to Geekbench, the i5-7600 averages at 5093 single-core 14858 multi-core whereas the P7550 is 1541 single-core 2561 multi-core.

The fastest though by a long shot is my Windows desktop. It’s always lightning fast on that one. 8 GB RAM and SSD with 6-core Phenom II X6 1055T, ~2500 single-core and ~10000 multi-core.
 
Last edited:
I have given up trying to figure out what’s going on with Excel. It usually launches faster than that on my iPad Air 2 but then again sometimes it has taken about that long on my 2017 Core i5-7600 iMac with 1 TB SSD and 24 GB RAM.

My 2009 8 GB Core 2 Duo 2.26 GHz MacBook Pro with an older version of Office launches Excel faster than that 2017 iMac.

According to Geekbench, the i5-7600 averages at 5093 single-core 14858 multi-core whereas the P7550 is 1541 single-core 2561 multi-core.

The fastest though by a long shot is my Windows desktop. It’s always lightning fast on that one. 8 GB RAM and SSD with 6-core Phenom II X6 1055T, ~2500 single-core and ~10000 multi-core.
Load time on the iMac seems unusually slow.

My office PC has i5-2500, 4GB RAM, Samsung 850 Pro 256GB SSD and Windows 7 32-bit, and it launches Excel 2010 in <2 seconds and Excel 2003 in 1 second. Geekbench 4 for i5-2500: ST 3300, MT 9200.

Mind, iirc after switching to SSD (circa 2009-10), launching Excel was always fast on my PCs (all Windows household). Excel's usually up and running after 1-2 seconds on Core 2 E8400 or newer.
 
Load time on the iMac seems unusually slow.

My office PC has i5-2500, 4GB RAM, Samsung 850 Pro 256GB SSD and Windows 7 32-bit, and it launches Excel 2010 in <2 seconds and Excel 2003 in 1 second. Geekbench 4 for i5-2500: ST 3300, MT 9200.

Mind, iirc after switching to SSD (circa 2009-10), launching Excel was always fast on my PCs (all Windows household). Excel's usually up and running after 1-2 seconds on Core 2 E8400 or newer.
Excel loading is slow not just only on my 2017 iMac. It’s very slow to load on my 2017 12” MacBook too. Excel version is 16.16.9. Legal licence.

I just tried it on my iMac after a fresh reboot, loading a relatively simple 87 KB Excel file off my NAS (which is connected via a Gigabit connection and which has SSD), and it took 17.5 seconds! Rebooted and tried it again, and it took 18 s. :rolleyes:

I then exited the application and then relaunched it with the same file off the NAS. This time it was 1-2 seconds, so it’s not the Ethernet connection speed, the NAS speed, or the complexity of the file that’s to blame. BTW, it’s good to have more than enough memory on a Mac. Caching helps a lot. I have 24 GB on my iMac and 16 GB on my MacBook.
 
Last edited:
Excel loading is slow not just only on my 2017 iMac. It’s very slow to load on my 2017 12” MacBook too. Excel version is 16.16.9. Legal licence.

I just tried it on my iMac after a fresh reboot, loading a relatively simple 87 KB Excel file off my NAS (which is connected via a Gigabit connection and which has SSD), and it took 17.5 seconds! Rebooted and tried it again, and it took 18 s. :rolleyes:

I then exited the application and then relaunched it with the same file off the NAS. This time it was 1-2 seconds, so it’s not the Ethernet connection or the complexity of the file that’s to blame. BTW, it’s good to have more than enough memory on a Mac. Caching helps a lot. I have 24 GB on my iMac and 16 GB on my MacBook.
Wonder if it's an Office 16 issue or because the new Mac versions aren't as optimized as Windows. We have paid full versions of Office 2010 installed on our home desktops and there hasn't been a reason to update to newer. I do have an Office 365 subscription because of the iPad 12.9, though. Might try installing the latest Office version on one of the Windows laptops to see how that performs.

Lol, motherboard permitting, I've used 8GB RAM standard on my builds (or aftermarket laptop upgrades) since Nehalem (2009), 16GB since Sandy Bridge (2011) and 32GB on Haswell (2013). I always max out on RAM. :p

Macs are too expensive and I like being able to fix my PCs myself. :D
 
Wonder if it's an Office 16 issue or because the new Mac versions aren't as optimized as Windows. We have paid full versions of Office 2010 installed on our home desktops and there hasn't been a reason to update to newer. I do have an Office 365 subscription because of the iPad 12.9, though. Might try installing the latest Office version on one of the Windows laptops to see how that performs.

Lol, motherboard permitting, I've used 8GB RAM standard on my builds (or aftermarket laptop upgrades) since Nehalem (2009), 16GB since Sandy Bridge (2011) and 32GB on Haswell (2013). I always max out on RAM. :p

Macs are too expensive and I like being able to fix my PCs myself. :D
It's an Office 16 issue, since all their apps are slow to load. Office 2011 is much faster to load.

But neither are anywhere near as fast to load as Office on Windows. My Phenom II X6 1055T machine used to be an Athlon II X3 435 machine at almost exactly half the speed. The Athlon II was starting to lag, but the Phenom II makes it feel like a modern machine, despite the fact the CPU is almost a decade old, partially because Office is so fast on it. (Picked up the Phenom on eBay for cheap.) Mind you, I'm only using Office 2007 on my PCs, so of course it's fast.
 
It's an Office 16 issue, since all their apps are slow to load. Office 2011 is much faster to load.

But neither are anywhere near as fast to load as Office on Windows.
Not necessarily. You haven't tested Office 16 performance on Windows yet so we don't know if it's an issue with Office 16 in general or if it's because the Mac version of Office 16 isn't as optimized as the Windows version.

My Phenom II X6 1055T machine used to be an Athlon II X3 435 machine at almost exactly half the speed. The Athlon II was starting to lag, but the Phenom II makes it feel like a modern machine, despite the fact the CPU is almost a decade old, partially because Office is so fast on it. (Picked up the Phenom on eBay for cheap.) Mind you, I'm only using Office 2007 on my PCs, so of course it's fast.
I had a 1055T build before. To be honest, it was slower than the quad-core i7-860 for most tasks other than video encoding while guzzling more power.
 
Not necessarily. You haven't tested Office 16 performance on Windows yet so we don't know if it's an issue with Office 16 in general or if it's because the Mac version of Office 16 isn't as optimized as the Windows version.
I mean it’s an Office 2016 for Mac issue. I wasn’t meaning to comment on the Windows version.


I had a 1055T build before. To be honest, it was slower than the quad-core i7-860 for most tasks other than video encoding while guzzling more power.
Yes it’s a bit slower than my i7-870 in some regards but it’s more than adequate, and it makes it feel like a modern machine, in the sense my 12” MacBook also feels like a modern machine.

Ironically, I’m just using my iMac i7-870 as an external monitor right now, for my i5-7600.
 
Yes it’s a bit slower than my i7-870 in some regards but it’s more than adequate, and it makes it feel like a modern machine, in the sense my 12” MacBook also feels like a modern machine.
In fairness, after Core 2 Duo Wolfdale, the biggest upgrade one can make to their PC is a good quality SSD (even a SATA 2.0 one helps a lot).

Apple chipsets have improved exponentially.

A7 - Atom or worse (still, the fact that they're nipping at Intel's heels was a major achievement that no other ARM chipset could match at the time)
A8X - comparable to Core 2 Duo mobile
A9 - comparable to Core 2 Duo desktop

The gains beyond these three are even more impressive and GPU-wise, I believe Apple's chipsets have always been ahead of Intel's integrated graphics since at least A7. I'm really quite amazed with what Apple has done with their chipsets in such a short period of time.
 
In fairness, after Core 2 Duo Wolfdale, the biggest upgrade one can make to their PC is a good quality SSD (even a SATA 2.0 one helps a lot).

Apple chipsets have improved exponentially.

A7 - Atom or worse (still, the fact that they're nipping at Intel's heels was a major achievement that no other ARM chipset could match at the time)
A8X - comparable to Core 2 Duo mobile
A9 - comparable to Core 2 Duo desktop

The gains beyond these three are even more impressive and GPU-wise, I believe Apple's chipsets have always been ahead of Intel's integrated graphics since at least A7. I'm really quite amazed with what Apple has done with their chipsets in such a short period of time.
My A8X iPad Air 2 with 2 GB RAM feels considerably faster than my 2.26 GHz Core 2 Duo P8400 MacBook Pro with 8 GB RAM.

This shouldn't come as a big surprise though since A8X in Geekbench 4 gets 1800/4550 on iOS, whereas P8400 gets 1600/2700 on macOS.

Note that they are close for single-core, but nonetheless I think one of the biggest differences in perceivable performance here is with surfing. The A8X can lag at times, but overall generally it's quite decent. The experience is much, much worse on the Core 2 Duo MacBook Pro. It's often a struggle in 2019 to surf on the P8400 because complex pages just bog it down so much.

I know you can't necessarily compare Geekbench cross-platform like this but that delta nonetheless is significant, and partially explains my experiences.

---

I had said to myself that if I was going to get a new iPad, it should have a Geekbench 4 score of 10000 or higher, not because I really needed that for mainstream usage right now, but because it would last a while. The iPad Pro 10.5" I ended up buying didn't quite make that arbitrary grade, but close enough. ;) I tend to like to double my performance with each purchase when possible, so the next time around I'm aiming for 20000. :) They're almost there already though, with the 2018 iPad Pro at 18000.

For my iMac, I went from an i7-870 (10000) to an i7-7700K (24000), but then dropped back to the i5-7600 (16000) due to fan noise with the 7700K.
 
Last edited:
I know this is a bit off topic, but since there is no "Waiting for the 2019 iPad Pro" thread, I'd like to jump into speculation with the next iPads... and the A13 chip. I mean... I'm starting to feel a bit like with the MacBook Pro eternally waiting for the better version. And I think that's unhealthy.

I know the 7nm+ EUV process, and the A13 chip will be better than the A12, but how much better it will be the A13X compared with the A12X? More RAM? We'll see the jump to 6GB of RAM on all the lineup? What if Apple waits until 2020 to release the next iteration of iPad Pro?

Right now I'm debating between the iPad Pro 11" or waiting for the next one. But it is impossible for it to come until next October, or even after that. So... Maybe getting the current one will be a good idea.

I'm sure iOS 13 will run well on the current 2018 iPad Pro. And that's what matters.
If iOS 13 brings up external disk support, it will surely be at least for USB-C iPads.
[doublepost=1555700302][/doublepost]Also, I was tempted on waiting because of the known as "touch screen disease", which should be fixed with newer models. Although it seems less spread among 11 inch devices rather than 12,9 inch ones.
 
Last edited:
Mouse support
Setting default apps
Desktop safari
Improved file management, support for external drives, usb sticks etc.
Multiple instances of the same app side by side
Support for non-MFI game controllers
Ability to resize picture in picture window
Dark mode
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.