Apple A10X CPU

Alright Sup Forums, Apple did it, it's now competing with mobile quad-core i7's.

I wonder if those ARM chips can into "hyperthreading", anyone has info on this?

Remember all that Intel performance comes from the 3.8GHz turbo boost.

Other urls found in this thread:

anandtech.com/show/9766/the-apple-ipad-pro-review/5
9gag.com
twitter.com/AnonBabble

those scores aren't even remotely close

>geekbench

>geekbench

You fucking idiot.

> Implying the A10X doesn't destroy it on ANY benchmark/real world usage.

it's comparable with i3 maybe

anandtech.com/show/9766/the-apple-ipad-pro-review/5


t.droid toddler

> 8MB L2
holy shit

> 2016
> A9X
This thread is from 2017 user

i know just showing how the a9x still behind core i5, nothing on 2017 iPad pro benchmark yet beside geekbench.

show it then

why should I

Why the fuck do you think it's reasonable to compare different architectures (x86 vs ARM), but you're not just comparing different architectures, you're trying to compare them ACROSS operating systems.

I mean, are you ACTUALLY fucking stupid or just pretending to not understand anything?

Geekbench is a multiplatform benchmark that runs tons of different CPU intensive tasks on any OS.

Congrats? No one disagrees with that.

But to imply each operating system will run the benchmarks in an identical fashion is fucking retarded.

Compare hardware on the SAME operating system, or shut the fuck up.

You appear to be severely retarded. Seek help.

>Compare hardware on the SAME operating system
not necessary. Some times we want to know which OS performs better on certain tasks as well.

You appear to have no understanding of how computing actually works, or what a benchmark is actually doing apparently.

Comparing across architectures is already retarded because they're going to handle tasks in different ways, one architecture could be optimized to handle 100+ low power tasks simultaneously but have garbage performance on single high intensity operations.


Now on TOP of these architectural differences, you're going to compare them on different operating systems which handle operation scheduling in different ways, and try to tell me it's going to be an accurate comparison?

Fuck off moron

>no linpack benchmark
As someone who owns apple products I also think OP is a fucking retard.

This is entirely ignoring the fact that the A10X can only sustain max 2.4GHz on all cores for short durations before thermal throttling.

I think maybe it's time you accepted the unarguable truth. Apple had done it. They have somehow made an ARM cpu that stands toe to toe with Intel's mobile best and holds its ground. Without a fan. Without even a real heatsink. And unlike Intel who had got the performance was hard, they show no sign of slowing down. In a couple of years, the A12X will be twice as powerful as coffeelake. There is nothing Intel can do. Then they'll go into MacBooks. Then Intel is kill.

That's fine if you're comparing ARM to ARM, or x86 to x86.

But you can't actually think ARM to x86 is going to give you a fair comparison when you switch up the OS at the same time.

Shut up you fucking retard, you're making us all look like jack asses.

Go put the A10X to encode 2 hours of HEVC video in software mode and compare that to a mobile quad-core i7 doing the same.

There's a reason apple doesn't sell server ARM chips.

This is faulty reasoning. If I want to know the relative strengths and weaknesses of hardware/software combination A versus combination B that is a valid question. It's a higher level question but it is very valid. If I have a file I need decrypted, how fast can A do it? Now try B. I don't care what software is in play on both sides, I just want my data manipulated a certain way. Whichever one is fastest wins. If one is consistently faster then it is just plain faster. You most certainly compare disparate systems and get meaningful conclusions

>Intel babby can't handle the fact that his $1000 ELITE GAYMIN CPU is outone by a mobile ARM chip

If you ACTUALLY think that ARM CPU is anywhere close to performance of a desktop quad core x86 chip you're a fucking retard.

Are you retarded? I have a desktop for that. By far the workload for mobile devices is burst where you need maximum performance for a few seconds at a time. Rendering web pages, starting apps. Shit like that. You need to reevaluate your arguments

Too bad it runs a baby toy virus OS.

Then what the fuck is the point of this thread?

Why even bother comparing it to intel?

But it is though. In every benchmark known, the A10X is within spitting distance of the i7.

Just wait for iPhone 8 and A11.... it will be 2x as fast as A10 and 3-4x as fast as Qualcomm and other Androshit CPUs.

It's ridiculous that Qualcomm, Samsung etc can't even beat a 2 year old iPhone CPU!

>In every benchmark known
show it?

...you actually think a 15w mobile CPU can beat out a 90w+ desktop CPU...?


Holy fuck you ARE retarded. Jesus Christ, do apple fanboys actually believe this?

There are plenty of tablets rocking Intel processors that are used for typical tablet duty. It makes perfect sense to see how the different platforms compare.

No man. We're obviously talking about mobile. Try to keep up

>intel stops trying for 5 years
>get surprised when other companies surpass them.

>that his $1000 ELITE GAYMIN CPU

Since there are no $1000 mobile CPUs i'd say no we weren't

But you'll probably just cite some retarded mobile Xeon or some shit like a faggot.


i'm done with this thread, you retards are either actually this fucking dumb, or shilling for apple because you have nothing better to do with your life.

Come on man. The onus is on you. Show me a legitimate benchmark where the A10X isn't within a few percentages of a mobile i7
>I'll be waiting

>he A10X can only sustain max 2.4GHz on all cores for short durations before thermal throttling
[citation required]

Dude, common sense says we're not talking about a 7700k. Common fucking sense. I like how you are trying to move the conversation away from the real point though. Must suck to be delidded this hard by Applel.
By the way
>I actually hate Apple but are impressed by what they've done here. You should be too

You gotta admit that Apple and AMD has started digging the Intel's grave long time ago but it's became apparent just recently. Can't wait to see those kikes die out in the following years.

I'm impressed by what they've done with an ARM chip in iOS.

It means literally nothing compared to x86 and trying to frame it as meaningful is just flat fucking wrong.

No they're not. People use intel tablets to get work done and children use ipads to play angry birds and browse 9gag.

Go to any apple store and put something cpu intensive on an ipad pro. You'll notice dramatic performance reduction after only a few minutes.

>within spitting distance
>more than 30% behind
I'm not saying it isn't impressive, but lets not exaggerate.

>Being this hard headed
Let me break it down. You have data. Be it a document, an audio file, video, some html from the network, whatever. Typically, you want something done to this data. You want it transformed in some way. Maybe the video needs to be encoded, the html needs to be rendered, something. The end result is what is important. Whatever app you use is incidental. Just a means to the end. When comparing different platforms meaningfully, what matters is which one transforms your data into whatever form you need. That is what matters. Benchmarks tell you which platform is does it faster. As your data is all that is meaningful, it is ludicrous to insist x86 and ARM cannot be compared. They both ultimately are doing the same thing so the question is, which is faster? I hope you get this through your thick skull and see your error.

does anyone have the version of this where he's fumbling with a noose?

Also who the fuck takes geekbench seriously?

Have you seen the improvements from one generation of Apple's CPUs to the next? 30 percent (it's not really that much) gap will likely be closed in the next round.

Yea TOTALLY reasonable to compare an x86 laptop to a fucking ARM tablet.

Literally kill yourself now, you have no hope of ever finding meaningful employment in the IT field.

We can't compare them because ARM chips can't run x86 code in iOS. Things like full desktop adobe or sony program suites become unavailable.

It's apple's to oranges. Or in this case a children's tablet to an actual computer.

Then find whatever benchmark suits you. The results are pretty consistent in all of them. I find it funny too that the Intel fanboys here are reaching for the i7 when most people buy the i3 and I5. How does the A10X compare to those? I think you know.

Because you cant

Compare it to Zen APUs and then we'll talk.

Your entire argument is "la la la can't hear you" when I am patently explaining in detail why you are wrong and why it is very reasonable to compare ARM and x86. Your really funny in your butthurt.
>Your meming can't hide how much you know I'm right

I don't browse 9gag, ur a fag.

the fact you think that you're makes me laugh, thank you for entertaining me.

It's good to know people actually are this dumb and managing to function just fine in society, it really reassures my own life choices.

>that you're makes
*that you're right makes

> comparing x86 and arm like something equally
LMAO

Just like they're are apps in iOS that won't run on Windows, there are Windows apps that won't run in iOS. You don't just throw your hands up. You use common sense and critical thinking. Again, what are you trying to do with your data? What applications on both platforms manipulate data in the same way? Web browsers being an obvious example. Compare those and look for the performance pattern. It's easy if you try.

>not using a jailbroken Sup Forums app
lmao yes you do

dude stfu and go back to 9gag

>Just use app lmao
I think we both know you are a faggot, there is no need for denial at this point.

I don't even know what the fuck 9 gag is. I know butt hurt when I see it though.

>"hyperthreading"
Just call it what it is, SMT

9gag.com

That's the website address in case you forgot. Please go there and never come back here.

>risc vs cisc
That aside, why is Apple literally the only good ARM designer?

What the fuck is this shit? You should kill yourself for ever making me clicking on that link.

None of the other companies pulled in Keller to get a nice design to coast on.

still cant download youtube videos

Geez OP, whatever will you gain for being this stupid. I get it that you are a faggot and tis your duty to shill hard for applel but atleast make yourself look intelligent by making reasonable arguments...I mean, look at how pathetic your reasoning sound m8....there is a reason why CPU heavy programs like emulators don't run well on iOS devices. Try to play dolphin on iPad and you'll see. Make a start by learning the difference b/w arm and x86. Hint: one is made to support low powered mobile devices with miniscule batteries while the other is made to support strong but high powered devices with batteries as big as a baby's wrist. One requires a fan the other doesn't. There is a reason why tegra X1 still cannot compete with x86 devices.

It's your home, stop lying. Now stay there and never come back here please.

Nice selection faggot. ITT OP is comparing A10x to Intel i series CPUs. There is no mention of operating systems. So if we choose IOS as our OS of benchmark and then compare the performance of a laptop running iOS with i7 and iOS with A10x, I bet i7 wins.
Similiarly, if windows is chosen as our reference OS, i7 still wins over A10x. Comparing the combination of iOS/A10x with win10/i5-i7 is downright retarded and doesnt prove that A10x is better than i5-i7. If according to your stupid analogy, the decryption rate is faster in iOS/A10x than win10/i5, there is a high possibility that windows is a crappy OS(No shit!) as compared to iOS.

>She still doesn't get it.
You can lead a horse to water but you can't lead a horticulture. Enjoy your memes lad

7zip and encoders are cross compiled between architectures. You can gauge performance using more programs like that. That's what Geekbench does.

The guy's either an idiot or on the (((payroll))). Either way you're wasting your time. I just hope no impressionable new people get sucked into his brand of 'tism

No you can't you retarded mongoloid. ARM chips make up their abhorrent floating point math performance by having hardware AES encryption which boosts their geekbench scores.

So it's faster. Got it.

Intel chips have had AES acceleration for years.

65 watt CPUs are already beating 140 watt CPUs, the bingbus is nearing it's last destination.

Here's a R5 1400 for comparison (3.9GHz)

>4782 vs 3982
Sure, not within 90%, but more than well within a factor of 2.

But can it run crysis?

>I wonder if those ARM chips can into "hyperthreading", anyone has info on this?
They don't have SMT, no.

The OS isn't going to matter much anyway in purely CPU-bound workloads. Maybe if the multithreaded tests actually used the system libraries' mutex implementations, but I can't even imagine Geekbench being that stupid.

Then I assume Intel is finished when Apple implement that, the IPC improvement and multi-core scaling would be massive.

I was not that guy you were arguing with. Your entire argument/analogy was retarded m8 ,as I pointed out in my reply. Maybe the previous user just plain figured out that it was best not to argue with a block of wood, so he quit. Maybe I should do the same.
>You can lead a horse to water but you can't lead a horticulture. Enjoy your memes lad
Now you again make some retarded analogy(or metaphor, I cant tell because it all sounds like the rambling of a dumb nigger) which is almost unintelligible given the context.
See pic related

>This is entirely ignoring the fact that the A10X can only sustain max 2.4GHz on all cores for short durations before thermal throttling.
This is equally true for the passively cooled Intel CPUs being compared with, though. The ability of the processor to stay cool and not throttle is part of the test. I know it depends on factors that are external to the CPU core itself, but just think of it as "system performance".

>Sure, not within 90%
but IPC is superior on the A10X since it's clocked at 2.4GHz vs 3.8GHz on the i7

Apply has always (and quite reasonably so) focused on primarily single-threaded workloads, so I wouldn't hold my breath for that.

Doesn't really help much if that higher IPC is what causes the chip to clock lower due to requiring more complex wiring and/or outputting more heat, though.

x86 is still performing great on mobile tho, however...
> Snapshit
lmao, no wonder those chips lag everything with it in.

Why do you think the operating system would have an influence at all? Does Geekbench call OS routines to decrypt files?

> A10X
> 15W
> 3-core/3-thread on multithread performance scoring 9275

> i7 4770HQ
> 45W firehouse
> 4-core/8-thread scoring 14941 on multithreaded

With some simple math and even a pessimistic scaling idea, we could say that a quad-core A10X would score around 11.000 - 12.000, thats only 24% slower than the i7, but at 1/3 of the TDP.

Apple won.

call me once the chip can do cinebench/aida64

t. pajeet

i preffer to call myself a linusfag.

>artificial benchmarks

Whoever is talking about geekbench? This guy was talking(not even talking but just giving a hypothetical scenario) about decryption. As for geekbench itself, yeah well, it is not reliable.

Secondly, read all these
There's a good reason why you could simulate an iOS and an android environment on an x86 chipset but not simulate windows apps on iOS or android devices.This is the most retard friendly way I can explain my thoughts.
OK now gotta go, this thread is full of summer fags.

...

The huge problem of geekbench is that it does small workloads that fit into cache memory, one part of the processors is access to main memory, performance with larger loads or the one of multiple workloads in ARM fall.

The multiplication of matrix, is of 512x512 in simple floating point.

...