Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

iOS 12 Core ML Benchmarks

heartbeat.fritz.ai

Post reply on HN
Print view View original

21–30 of 59 posts

Re: iOS 12 Core ML Benchmarks

#21

Post by jamesonthecrow »

npunt wrote: post #18
dep_b wrote: post #2

Impressive but with a specialized coprocessor it's not unexpected. It's a bit of a gamble for Apple though, as other ARM SoC's are probably able to bridge the gap in raw single-core and multi-core performance in a year. The whole ML thing is a bit fuzzy for a lot of people even for programmers like me that are still wondering how their regular applications are going to benefit from all this silicon.

Other ARM SoC's haven't been able to bridge the relative gap in raw single-core performance going on 5+ years, and haven't hit multi parity since iDevices went beyond 2 cores. I don't really see a gamble here at all - adding ML blocks is relatively cheap and icing on the cake. Take Geekbench (and any other benchmarks) with a grain of salt of course, but benchmarks & reviews all corroborate the same story. See these G…

Your point about the integration between software and hardware is spot on. Even the Android devices with powerful GPUs or AI accelerators are really difficult to access because Android APIs (even the NNAPI) is really tough to use. Core ML "just works" with the CPU / GPU / Neural Engine.

Re: iOS 12 Core ML Benchmarks

#22

Post by mtgx »

baby_wipe wrote: post #16
overcast wrote: post #13

Earlier quoted context omitted.

The biggest change was ramping up processor instantly, and then back down to maximize performance. So for launching an app / scrolling, CPU reaches highest state, until that portion is over, and immediately get back to efficiency setting for preserving battery life. Previously CPU time would ramp up slowly.

It's probably paranoia, but I feel like my 5S battery is draining faster since the update. That being said, the performance upgrade is well worth it.

Why would it be paranoia? Waking up the CPU faster and maximizing its performance will definitely drain the battery faster.

I assume the iOS12 update includes a combination of more aggressive waking and maximizing of performance along with a reverting of what they did previously to old phones when they lowered their performance to maintain the battery life of the degraded battery. If you remember there was quite a bit of scandal about this earlier this year.

My guess is your old iPhone's battery will also last half as much as it did originally within months. Yes, it's a factor of the new changes, but let's not forget it's primarily a factor of Apple using batteries that lose a significant percentage of their capacity after 2 years in the iPhones, and then hiding much of that by slowing your iPhone down. Maybe Apple should put higher quality batteries in their phones - just saying.

Re: iOS 12 Core ML Benchmarks

#23

Post by fisherjeff »

npunt wrote: post #18
dep_b wrote: post #2

Impressive but with a specialized coprocessor it's not unexpected. It's a bit of a gamble for Apple though, as other ARM SoC's are probably able to bridge the gap in raw single-core and multi-core performance in a year. The whole ML thing is a bit fuzzy for a lot of people even for programmers like me that are still wondering how their regular applications are going to benefit from all this silicon.

Other ARM SoC's haven't been able to bridge the relative gap in raw single-core performance going on 5+ years, and haven't hit multi parity since iDevices went beyond 2 cores. I don't really see a gamble here at all - adding ML blocks is relatively cheap and icing on the cake. Take Geekbench (and any other benchmarks) with a grain of salt of course, but benchmarks & reviews all corroborate the same story. See these G…

I, too, just ended up on the Geekbench browser and hadn't noticed until now that A12 would appear to be roughly at single-core parity with Kaby Lake-U:

i7-8559U: 1901/GHz [0]

A12 Bionic: 1998/GHz [1]

I'm not familiar enough with the actual tests to know how comparable those scores are, plus, as you say, grain of salt, etc., but the rate at which the gap has been closed continues to be remarkable.

[0] https://browser.geekbench.com/macs/433 [1] https://browser.geekbench.com/ios_devices/56

Re: iOS 12 Core ML Benchmarks

#24

Post by rhencke »

dep_b wrote: post #6
spullara wrote: post #3

Earlier quoted context omitted.

Apple doesn't use ARM SoCs. They have their own designs from the PA Semi team and are really just using the ARM instruction set at this point. This is one of the reasons there is such a big performance gap between the top android phones vs the top iPhones.

I guess the same can be said for AMD's x86_64 CPU's? Still it's fair to compare them to Intel x86_64 CPU's. The big difference is that ARM has something like a reference design where Intel doesn't want to share anything if they don't have to.

It's actually the other way around, humorously enough.

x86_64 is also known as AMD64, and for good reason. AMD created the core 64-bit x86 design, but, more importantly, they have patents to key pieces of the ISA required to legally implement it.

So, long story short, it is actually AMD who didn't have to share anything with Intel if they did not want to - Intel's 64-bit x86 CPUs are fundamentally AMD64-compatible CPUs.

The most interesting thing is it resulted in a bit of an interesting legal relationship between the two. Intel licenced AMD the required rights to produce x86-compatible processors. AMD, in turn, ended up licensing Intel the IP required for Intel to create x86_64 chips.

Sources:

https://en.m.wikipedia.org/wiki/X86-64#Licensing

https://www.sec.gov/Archives/edgar/data/2488/000119312509236...

Re: iOS 12 Core ML Benchmarks

#25

Post by jeffbax »

dep_b wrote: post #2

Impressive but with a specialized coprocessor it's not unexpected. It's a bit of a gamble for Apple though, as other ARM SoC's are probably able to bridge the gap in raw single-core and multi-core performance in a year. The whole ML thing is a bit fuzzy for a lot of people even for programmers like me that are still wondering how their regular applications are going to benefit from all this silicon.

You clearly haven't really kept up with Apple and ARM for the past 4 years have you. They're cleaning everyone's clock (pun unintended) with no signs of that abating.

Last years X is still a generation faster than the fastest Qualcomm/Samsung chip today let alone what generic designs ARM licenses to the cheaper chip companies.

Re: iOS 12 Core ML Benchmarks

#26

Post by KSS42 »

The article also points out iPad 2018 being faster than iPhone 7/8/X. This is odd.

The article states that iPad 2018 uses A10X, but it actually use the A10.

Is there a mix up with one of the iPad Pros? I can see an A10X (which has more GPU cores) beating the other iPads.

Re: iOS 12 Core ML Benchmarks

#27

Post by micv »

mtgx wrote: post #22
baby_wipe wrote: post #16

Earlier quoted context omitted.

It's probably paranoia, but I feel like my 5S battery is draining faster since the update. That being said, the performance upgrade is well worth it.

Why would it be paranoia? Waking up the CPU faster and maximizing its performance will definitely drain the battery faster. I assume the iOS12 update includes a combination of more aggressive waking and maximizing of performance along with a reverting of what they did previously to old phones when they lowered their performance to maintain the battery life of the degraded battery. If you remember there was quite a bi…

> let's not forget it's primarily a factor of Apple using batteries that lose a significant percentage of their capacity after 2 years in the iPhones

As far as I can tell this simply isn't true. My iPhone SE is at 89% of maximum capacity after 2.5 years of very regular use. It's standing up pretty well.

Re: iOS 12 Core ML Benchmarks

#28

Post by MBCook »

KSS42 wrote: post #26

The article also points out iPad 2018 being faster than iPhone 7/8/X. This is odd. The article states that iPad 2018 uses A10X, but it actually use the A10. Is there a mix up with one of the iPad Pros? I can see an A10X (which has more GPU cores) beating the other iPads.

It may simply be that due to a bigger battery and better heat dissipation it can run faster/longer and thus come out ahead.

That said I think you’re right about there being a mixup. Perhaps they mean the 2017 iPad Pro 10.5 inch.

Re: iOS 12 Core ML Benchmarks

#29

Post by MBCook »

jamesonthecrow wrote: post #21
npunt wrote: post #18

Earlier quoted context omitted.

Other ARM SoC's haven't been able to bridge the relative gap in raw single-core performance going on 5+ years, and haven't hit multi parity since iDevices went beyond 2 cores. I don't really see a gamble here at all - adding ML blocks is relatively cheap and icing on the cake. Take Geekbench (and any other benchmarks) with a grain of salt of course, but benchmarks & reviews all corroborate the same story. See these G…

Your point about the integration between software and hardware is spot on. Even the Android devices with powerful GPUs or AI accelerators are really difficult to access because Android APIs (even the NNAPI) is really tough to use. Core ML "just works" with the CPU / GPU / Neural Engine.

Qualcomm can DESIGN a chip and HOPE it’ll go in 20 million devices in a year. If they take big enough risks it’s possible no one buys the chip and they just eat a huge loss. I do the phone manufacturers don’t want it, or they don’t think it’s worth the extra three dollars compared to a previous chip.

Apple has inly one customer, they can design a trip and KNOW it’s going into 150 million devices within a year. Even if the APIs aren’t given to third parties, it may still be a big win for them.

Re: iOS 12 Core ML Benchmarks

#30

Post by walterbell »

ghop02 wrote: post #10
saagarjha wrote: post #9

It's interesting to see that ML performance is about equal on iPhone X and older iPhones, considering that iPhone X (and 8) has a neural engine and the other phones do not.

It's my understanding that the neural engine API wasn't actually exposed to Core ML until the iPhone XS. The GPU is the important factor for performance of those models

Will it be exposed on the iPhone XR (the "mass market" model)?

Presumably it will be supported on the upcoming iPad Pro and iOS' (mOS?) Arm laptop.

Post reply on HN