Is the future ARM, or are these Benchmarks bullshit?

>>62456053
ofc its the future

Intel dragged their feet for too long and apart from Ryzen nothing exciting is happening on x86

>>62456053
>Geekbench

Geekbench is absolutely retarded and does not reflect performance, whether overall mix or single task, in any way shape or form.

My 1600x shits over it anyway 16000+

single thread is the same who cares its outdated asf

File: Screenshot_20170916-161318.jpg (281KB, 1439x885px) Image search: [Google]

281KB, 1439x885px

>>62456066
An iPhone X has half the muli and 2/3's the single of an i7 7700k according to Geekbench..

>>62456073
Is there a way to accurately benchmark ARM vs. x86?

>>62456089
Is your 1600x 1/4th the size of your thumbnail, passively cooled, and driven by a phone battery?

>>62456053
Geekbench is trash. Truth is arm can do what 2010 CPUs could do.

>>62456096
Probably only SPECint or SPECfp.

Not very good, but better than the Geekbench trash.

>>62456053
a chip that has an ISA that only works on IOS and geekbench has be optimised for IOS
brings out the best result vs fragmented OS

ofc its not the future this is bending reality you simply cant compare them

>>62456053
>ARM @ 24 MHz

File: 1471122072239.jpg (16KB, 351x329px) Image search: [Google]

16KB, 351x329px

>>62456600
Seems legit.

>>62456115
can your applel run actualy games or applications?

ios is dogshit

File: 4424.png (27KB, 857x595px) Image search: [Google]

27KB, 857x595px

>>62456053
A11 beats Ryzen 1200, impressive

>>62456676
applel has no games tho

>>62456676
And the A11 is only running at 24MHz

why is there so much retardation in these threads

>>62456053
RM is the future because Apple has more money than anyone for R&D.

This chip was designed 3 years ago.

Arm is garbage and far from the future.

>>62456676
>People keep explaining why Geekbench is fucking trash and not directly comparable
>Keep posting it anyways
Almost as if someone was paying for these posts!

>>62456120
2010 had some great cpu's like the i7 920.

>>62456844
its 10nm too, imagine the 5nm ones in 2020

>>62456885
its from q4 08 actually

>>62456053
A11 has 4.1b transistors
I7-7700 has 1.75b

Is that due to the GPU in the A11?

>>62456053
Now post the cinebench r15 results.

File: apple a10 soc.png (555KB, 900x600px) Image search: [Google]

555KB, 900x600px

>>62457025
It's a system on a chip.

2 + 4 ARM cores
3 GPU cores
motion coprocessor
neural network processor

Here is the A10 (125 mm2) of which each big core takes only 4mm2.

File: apple a10.jpg (311KB, 1280x960px) Image search: [Google]

311KB, 1280x960px

>>62457199
comparison of the A10 big core to other mobile cores.

Can't wait to see diagrams of the A11.

>>62457199
>neural network
apple fags just eat this shit up, don't they?

>>62456053
>Is the future ARM
No.
>are these Benchmarks bullshit?
Yes.

>>62457268
No, it literally has small ML chip, think of Movidius ones.

RISC is gonna change everything

>>62457647
You mean RISC-V?
Then no, it's so autistically RISC it's unsuitable for any high-performance designs.

So wait, there actually ARE alternative CPU designs to x86?

What the hell has taken them so long?

How the fuck will everyone on x86 adapt if nothing x86 will work on arm or risc?

>>62457702
No one will bother to port anything to ARM though.
That's why AMD shelved K12 you know.

>>62457678
I hope you dont screw like you type

>>62456053
Bullshit. Iirc, geekbench overvalues sha and aplel cpus have hardware sha. They arw much slower than x86 in real eorld performance.

File: 10720.jpg (15KB, 326x262px) Image search: [Google]

15KB, 326x262px

>>62456053
You can't compare different architectures with GeekBench.

>>62457773
Shouldn't the question be why Intlel doesn't have hardware support for the most used algorithm in the entire world.

>>62457025
It's because the ARM is a SoC.
Even the last Atoms have more transistors than an Intel I7-7700 just because it's a SoC and not just a CPU.

File: 1502130844848.png (310KB, 439x387px) Image search: [Google]

310KB, 439x387px

>>62456053
>KikeBench

File: Acorn_computers_logo_rendered_with_text.svg.png (39KB, 1024x327px) Image search: [Google]

39KB, 1024x327px

It's all thanks to Acorn. Long live the ARM - the Acorn RISC Machine

>>62457790
Except you can. All Geekbench versions have the same workload.

>>62457711
>no one will bother
except linaro who specializes in ARM compilers
except arch
except debian
except google for android
except google for chromebooks
except arduino
except raspberry pi
except windows for pi / shitty rt implementations

Except TI, Samsung, Atmel, Apple, Freescale, Rockchip, and pretty much every other major hardware manufacturer.

Intel runs windows gaymin though, gotta love that proprietary philosophy locking software into platforms to keep said platforms alive. Otherwise ps3, xbox, and pc would have died years ago in favor of centralized linux gayming on all devices. This is why you cant have nice things.

>>62458667
No one cares until big enterprise shit like SAP and Oracle DB gets ported. From then on it's smooth sailing.

>>62456053
>iTard believes that a smartphone cpu is better than a desktop/laptop cpu because of geekbench
That's why you buy iShit and I don't

>>62457268
A11 calls A11 bionic neural engine, until moment maybe begins GPU making 600GFlops in 16bits.

>>62458530
All workload almost only exist in cache memory >>62457212

https://www.geekbench.com/doc/geekbench4-cpu-workloads.pdf

>>62456053
They're big as fuck they weren't joking for dekstop classes performance, give it a time we will see it soon on macbook maybe.

>>62458530
The workloads don't represent real world performance.

>>62458924
The absolute state of /g/

>>62459349
>inb4 iphone x > i7 7700k

>>62456053
>are these Benchmarks bullshit?
If you compare ARM to ARM. No they're not
If you compare ARM to x86. Yes

Doing the geekbench my 1700x gets a LZMA score of :

28602
44.7 MB/sec
(geekbench compresses and decompresses a 450KB html ebook)


in 7zip I get
31MB/s compress
500MB/s decompress

Either the implementation is crap, or it's heavily weighted in favor of compressing (Which makes no sense as you will be decompressing much more often) in a arbitrary way (not based on the time each step takes)

>>62459483
Yes Ryzen excels at decompression. This is a fundamental trait of the architecture.

>>62456053
They're bullshit. geekbench is a horrible benchmark even for a synthetic one especially across different CPU ISAa.

>>62456767
>>62456659
is this bait

>>62459855
I know. that should reflect in the the geekbench as well.