ARM is dead

That's stock settings btw

Attached: image_id_2052020.png (1920x1080, 1.61M)

This, we should see 48c96t @ 3 GHz minimum for servers soon.

It doesnt work that way. It doesnt take decades for each ISA. It's all about the implementation around the ISA.
This means that technically any company with money can do a an x86 killer. For this example lets say AMD wants to make a rushed zen equivelant on risc-v. They can do it and have a working prototype in weeks.
The real issue is optimization for the ISA in software like gcc. This is the thing that kills performance for ISAs like power9.

Anyonr knows what's written?

user there is already a 22c/88t solution on power9.
If you get a dual cpu motherboard you can get to 44c/176t.

that's pretty cool, why the outrage?
most of phone's cpu time is probably spent executing jitted js. This is going to save battery life

Nice so how many decades before someone fixes this stillborn freetarded ISA and ports 99% of all commercial and enterprise software to it?

Want to buy DEC stocks? Heard they have a nice RISC architecture.

Cool, how many kW does that consume?

Attached: Dual-AMD-EPYC-7601-in-Supermicro-2U-Ultra-Power-Consumption.jpg (1230x715, 63K)

>half a kilowatt
Intel and AMD to start investing in nuclear power soon?

>it's just faster floating point operations
Why is this a bad thing?

Most of debian has been ported.
Normal android users do not care. Google already wants to support risc-v.
Come to think the majority of android apps are java so it doesnt matter. If samsung for example just says we dont want to pay for arm licencing. A change to risc-v will be quick and painless for their phones.
Servers already use freetard software and any proprietary software ports can be done quickly.(especially if they use some scripting language like python)
Workstations use freetard software unless they want to use windows for something.
Microsoft already wants to get rid of x86 and move to arm slowly so i would not be surprised if they also made a risc-v port if it became famous.
I guess a lot but do companies that buy workstations care about that?

Usually you dont upgrade an ISA for one language.

>ARM is dead
How? The faster things go the better, right? Or are you one of those Jow Forums faggots?

Is it really only specifically made for JS or does JS just happen to benefit a lot from it since it uses float all the time?

Yeah man, who cares if you have to spend thousands of dollars more every month in additional electricity to feed meme CPUs. Efficiency doesn't matter.

he's a rectumravaged boomer who can't cope with the fact that he wasted half his life reading C++ books

power9 would consume at least 10x that

raptorcs.com/content/CP9M08/intro.html
Its 190W per CPU.
That is actually quite comfy for being the high end.

holy shit that's bad, no wonder these meme cpu's have been used for database 99% of the time

What's the ipc like compared go zen, 25%?

How is it bad for workstations to have 190W CPUs.
Threadripper is at that range

Because they have nowhere near the same IPC let alone AVX 256 performance. I was expecting around 95W but I guess I expected too much.

> ARM is dead
It's just JS won.

>Muh cornz!!
>Muh cinecornz!!!

Attached: Mark_White_Logo_T-shirt_Square_1024x1024.jpg (700x700, 29K)

Yeah man, most consumer cpu workloads consist of high i/o database tasks.

Attached: 1538685539419.jpg (480x480, 25K)

>Nice, how many decades before it can compete with the zen 2 x86 ISA processors coming out in a few months?
You don't understand the point of high performance RISC architectures. Here the point is not just IPC, but scalability in *multi-socket* systems.

IBM's year old POWER 9 has 24cores 96threads, and you can have 2 of them on the same board, for 48cores 192threads. And insane RAM amount and bandwidth.

You can get these systems now. What does x86 have to compete with that?

Clusters are not the answer because some workloads require single systems, otherwise latency becomes unusable. If you need those workloads, you *need* a scalable risc processor, even with worse IPC, because it ends up much much faster by keeping it single system vs cluster.

And mind you, POWER is not the most scalable risc architecture.
Sparc M8 for instance is 32cores 256threads. And you can put 8 in a board (the M8-8 server), for a total 256cores 2048threads.
Let that sink for a moment. Nothing x86 can come remotely close.

ARM, another risc architecture, even though not designed with scalability as first priority, is still infinitely more scalable. Proven by the fact that Fujitsu is building their next supercomputer, Post-K, with ARM processors.

In the case of ARM, it has even an IPC advantage over current Intel and AMD x86. Apple's A12, designed for low power consumption, is with 15% in single core *performance* from current x86 chips, at a 2Ghz *disadvantage*. Which indicates crazy higher IPC, in the order of +50-60% compared to x86.

The x86 ship has sailed user. It just lingered on life support due to legacy desktop software.

>In the case of ARM, it has even an IPC advantage over current Intel and AMD x86. Apple's A12, designed for low power consumption, is with 15% in single core *performance* from current x86 chips, at a 2Ghz *disadvantage*. Which indicates crazy higher IPC, in the order of +50-60% compared to x86.
>The x86 ship has sailed user. It just lingered on life support due to legacy desktop software.
Let me guess, you base this off gookbench and not FP64 linpal sustained benchmarks, right?

Attached: i74zmsic4ziggrlkv9c5.gif (800x450, 1.87M)

>JS just happen to benefit a lot from it since it uses float all the time?
This.

So tl;dr, power9 is dogshit outside of database or connecting GPUs together with almost no performance loss as it scales up for supercomputers?

>v8.3
>latest
slowpokenews.org

>Let me guess, you base this off gookbench and not FP64 linpal sustained benchmarks, right?
No, I base it on the industry standard used for servers and databases to measure single thread performance and ipc, SPEC CPU2006.

Do you have a benchmark proving otherwise?

>database or connecting GPUs together with almost no performance loss as it scales up for supercomputers?
This one is true.

>power9 is dogshit outside of
This one is half true. There's nothing wrong with the architecture or the implementation, just nobody bothers to write software optimized for it, and even if they tried, the "free" compilers are hugely optimized for x86, while ignoring POWER. They support POWER in the most basic level, almost as if they're emulating it and not natively supporting it.

When people bothered to write, say, PPC optimised software (like Photoshop for PPC macs), it ended up faster even when the chips were lagging behind Intel's.

But the opposite is true as well. Try using IBM compilers, and the binary is blazing fast on POWER and molasses on x86.

POWER chips don't have an IPC disadvantage, or it's impossible to tell if the do.
Because everything is overshadowed by the abysmal performance of gcc and clang when building for POWER targets.

>"In the SpecInt2006 benchmarks, the 462.libquantum benchmark is highly vectorizable. The baseline computer for all benchmarks is a 1997 Sun Ultrasparc server computer. Whereas most of the spec sub-benchmarks turn in a performance improvement of about 5x to 80x times faster than the Ultrasparc, the particular 462.libquantum sub-benchmark turns in a result that is up to 4082 times faster than the Sun Ultrasparc.[3]This suggests that for this sub-benchmark, most of the improvements over the Ultrasparc are due to vectorizing compiler improvements, NOT due to CPU hardware improvements, since 1997."

>SPEC cpu2006
not this meme again

Attached: cpu2006-20120307-19618.gif (799x606, 9K)

bruh

Attached: 6597_07_microquill_smartheap_overview_improve_your_spec_cpu2006_scores.png (589x952, 207K)

>5.4GHz
>stock settings
You're supposed to inject your weed, not smoke it.

So?
We're talking about how *current* chips *compare* between each other, no compared to the reference.

The optimizations apply for both, so in that *sub-benchmark* they'll both have inflated scores.

Can you explain why ARM would have more inflated score than x86, unless it's a superior architecture?

Also, even excluding that sub-benchmark, can you provide proof that an Intel or AMD chip performs better?

If you want to know more, here is a fun document
static.docs.arm.com/ddi0487/ca/DDI0487C_a_armv8_arm.pdf

Look at again you stupid motherfucker.

I don't see the problem here. People use ARM phones. Those phones use websites with JS. ARM isn't doing anything wrong. They're just providing a better product for the customer. If anything, the real issue is people accept all these bloated JS sites to begin with.

>"show me a benchmark where an x86 chip beats teh A12"
>"here's one where an x86 chip with optimizations performs better than the same chip without optimization"

I'm arguing with a child blessed with one two many chromosomes.

You might of as well used gookbench. specint cpu2006 isn't even measuring FP performance which in case of AVX loads is what actually matters.

Read the post again and then the one he's quoting then think about what you just said you dense fuck.

Maybe I don't feel the need to read every linked post you assblasted shit goblin.

Attached: 1523548280522.gif (351x251, 40K)

>They're now trying to improve the bloated web at the hardware level because web devs can't make lean pages

Attached: 1481746843394.png (400x389, 48K)

perro caca

boludo

>specint cpu2006 isn't even measuring FP performance
Since when?!
Half of the benchmarks (more than half actually) are floating point:
>spec.org/cpu2006/Docs/readme1st.html#Q11

>AVX loads is what actually matters.
And? You seem to think ARM has large vectors too, they call them SVE (scalable vector extensions).
AMD goes up to 256-bits, Intel's AVX-512 is 512-bits, and ARM's SVE is *scalable* up to *2048-bits*, 4 times as wide as AVX-512. Also remember than Intel halves the frequency when using AVX-512 heavily.

Have you even read anything about the architectures apart from memes?

Apple dumped a bunch of money in it for PDAs in the '90s and subsequent low-power ARM designs could be pretty decent in high-end embedded applications against competing ISAs like MIPS and SuperH. After it pretty much disposed of all other mainstream competition in the early 2000s it settled and became the x86 of embedded devices.
That was somewhat true 30 years ago, nowadays the RISC vs. CISC label is pretty much meaningless and it's a matter of development and brand loyalty more than anything.

javascript core is written in c++

>The x86 ship has sailed user. It just lingered on life support due to legacy desktop software.
I'm really amused and baffled at the same time when I see morons write things like this with a straight face after evangelizing dinosaurs like SPARC (dead) and POWER (dying) that are very rarely seen in the wild outside of the handful of niche applications they won't shut the fuck up about and (more typically) vendor locked shops that are desperate to get away from them and the incredibly overpriced and underperforming and systems they ship in.

The big RISCs are dead, dying and have been so since the dotcom crash, it's time for you to move on, not us.

Attached: top-500-chip-technology-intel.jpg (770x471, 92K)

>R-R-R-RISC FIVE WILL KILL 8086

>Apple's A12, designed for low power consumption, is with 15% in single core *performance* from current x86 chips
Fucking tech illiterate retard, stop comparing architectures in mobile "benchmarks".
Standart A-53 cores have crazy scores in them too, but phoronix test suite on a 4xA53 based reference board says otherwise

> RISC / CISC
wtf are you serious have you written any x86 or ARM assembly lately? notice how different they are or how the machine works... doesn't that reflect how different they still are??

Are you for real, idiot?

>x86 is totally better than power and sparc when you exclude use cases where power and sparc are better

Attached: 863287623.jpg (1280x2880, 271K)