Editorial x86 Lacks Innovation, Arm is Catching up. Enough to Replace the Giant?

[Aquinus]

Goes without saying, you can find those cases even when the geometric mean is not that far off.

They tend to occur more frequently when it is though, but you're right. It can still occur when it's a lot closer. Just not usually like this.

 

[R-T-B]

Arm always wanted to get into High Performance computing, whereas x86 manufacturers always wanted to get into ultra low power devices. They never quite made it, because they develop optimal tools for completely different scenarios.

Thing is, ARM has made it into HPC, and several times now. The issue is more compatability at this point than raw ability.

my problem with ARM is that it isn't part of the familiar Windows ecosystem.

Technically with UWP and such, it is now.

If ARM was so clearly more energy efficient, all servers would be running ARM today

We are already seeing signs of a migration, it's mainly fear of the unknown and recompiling certain legacy apps holding us back.

Before AMD Ryzen
- My birth. 1986.

After AMD Ryzen
- My daughter's birth, 2018.

Thanks, AMD

Hey, same birthyear. Let's party like it's 1986!

 

[mtcn77]

one of RISC's advantages was that executing any instruction within the same timeframe/cycles dramatically simplifies scheduling. By contrast, ever since Intel went pipelined (Pentium, iirc), they essentially have a sizeable silicon chunk breaking complex instructions down into simple ones, emulating what RISC does

I like favors of this kind, you see. There is something visceral about conducting research. Applied sciences rock.
I wonder what Intel will do, they seem to have the best chances of deploying a breakthrough superconductor and what not, since they are more predisposed to search outlier hardware performance cases.

 

[ARF]

This 64-core Graviton2 is (much) slower than the 80-core Ampere Altra.

 

[ARF]

Excellent.

ARM-based Japanese supercomputer is now the fastest in the world

A Japanese supercomputer has taken the top spot in the biannual Top500 supercomputer speed ranking. Fugaku, a computer in Kobe co-developed by Riken and Fujitsu, makes use of Fujitsu’s 48-core A64FX system-on-chip. It’s the first time a computer based on ARM processors has topped the list.

Fugaku turned in a Top500 HPL result of 415.5 petaflops, 2.8 times as fast as IBM’s Summit, the nearest competitor. Fugaku also attained top spots in other rankings that test computers on different workloads, including Graph 500, HPL-AI, and HPCG. No previous supercomputer has ever led all four rankings at once.

ARM-based Japanese supercomputer is now the fastest in the world

Fugaku is being used in COVID-19 research

www.theverge.com

 

[bug]

Excellent.

ARM-based Japanese supercomputer is now the fastest in the world

ARM-based Japanese supercomputer is now the fastest in the world

Fugaku is being used in COVID-19 research

www.theverge.com

Doesn't mean squat in this context.
The workloads super-computers handle are very different from what we run at home/work.

 

[ARF]

And yet ARM-based home systems should be perfectly viable and we must start using them and start replacing the old x86.

 

[Vya Domus]

And yet ARM-based home systems should be perfectly viable and we must start using them and start replacing the old x86.

Why ? Just because ?

 

[ARF]

Why ? Just because ?

First - because I don't like my x86 experience with AMD Ryzen - it lags in loading in many cases - for example when opening the Start menu - some parts of it take longer to load and it gets ugly how one part of the Start menu is visible, and the other needs another second to appear.

Second - because power consumption - imagine a smartphone-sized tiny computer cases/boxes integrated in the monitors or attached to the monitors with the latest Snapdragon ARM-CPU.
That will be perfect for things like - email processing, 4K viewing, YouTube, Facebook, video playback of any type, lite smartphone-type gaming, etc.

 

[Vya Domus]

First - because I don't like my x86 experience with AMD Ryzen - it lags in loading in many cases - for example when opening the Start menu - some parts of it take longer to load and it gets ugly how one part of the Start menu is visible, and the other needs another second to appear.

Nonsense, come on, I have never experinced any lag in the Windows interface (unless obviously I am running something extremely demanding in the background). But let's assume that's the case, what makes you think an ARM chip wouldn't "lag" on the start menu as well ?

Second - because power consumption - imagine a smartphone-sized tiny computer cases/boxes integrated in the monitors or attached to the monitors with the latest Snapdragon ARM-CPU.
That will be perfect for things like - email processing, 4K viewing, YouTube, Facebook, video playback of any type, lite smartphone-type gaming, etc.

That's what you ARM advocates do not understand. ARM SoCs are designed to be power efficient at all costs, desktop chips aren't, they are focused around performance. A desktop ARM replacement will be "just as bad" as a x86 counter part if you want the same performance. ARM chips are no magic, they are just optimized for something else, looking at smartphone SoCs and extrapolating that to desktops is exceedingly dumb.

And if you don't want the performance, just use your damn phone or tablet. There is absolutely no point in having the exact same limited performance as you do with a phone without it's only advatage, which is mobility. No one wants that, you would be getting the worst of both worlds.

 

[medi01]

This. Once ARM follows much or all of what x86 CPUs have been doing, they are more than likely to end up in a very similar place in terms of all parameters.

Or not.

There must be reasons why RISC architecture evaporated from the server market.

 

[TheoneandonlyMrK]

First - because I don't like my x86 experience with AMD Ryzen - it lags in loading in many cases - for example when opening the Start menu - some parts of it take longer to load and it gets ugly how one part of the Start menu is visible, and the other needs another second to appear.

Second - because power consumption - imagine a smartphone-sized tiny computer cases/boxes integrated in the monitors or attached to the monitors with the latest Snapdragon ARM-CPU.
That will be perfect for things like - email processing, 4K viewing, YouTube, Facebook, video playback of any type, lite smartphone-type gaming, etc.

How do you have your pc setup, are you using a potato HDD, cos that's not right.
Both of mine are snappy.

I have my phone for light surfing, communication etc, my pc does things arm does poorly, all day, soo good luck with the x86 demise banter, to me that's all it is banter because anything worth doing well, I don't scrimp on the doing equipment or performance.

I'll check back in ten years see how it's going.

Oh and. Point one no one cares What you personally hate, especially if it's nonesense.

Point two, smart TVs, we already have them, office and all game's are shit on them.

 

[dragontamer5788]

Or not.

There must be reasons why RISC architecture evaporated from the server market.

The commonly cited reason is that Desktop chips provide a "mass production" target, subsidizing the lower-volume server market. We're seeing it again: TSMC has become the #1 fab in the world, because Apple, Qualcomm, NVidia, and AMD are all funneling their R&D money together. Cell phone chips subsidize the development of desktop, GPU, and more.

Its less a technical reason, and more of an economic one. Intel is in trouble, because TSMC probably has more high-end volume than Intel now. Remember: chip designs only cost pennies to mass produce, but hundreds of millions of dollars for R&D (fab-lab costs, mask costs, engineers, etc. etc.). The jump to 5nm probably will be on the order of $1 Billion to $10 Billion.

 

[TheoneandonlyMrK]

Pennies my arse, your exaggerating it costs more than pennies just to package them.

 

[ARF]

I'll check back in ten years see how it's going.

Yeah, in 10 years it will be Intel's 15th generation 14nm process

 

[TheoneandonlyMrK]

Yeah, in 10 years it will be Intel's 15th generation 14nm process

They won't be around if they're still on 10nm then, it's worth remembering Intel and AMD can and are adapting acceleration IP for incorporation into x86 design's while also continuously improving both efficiency and just as important , scheduling and power gating technology.
X86 will be down in arms park long before arm ever gets up into x86s performance turf in general compute.
With an efficiency and performance bat as a tool to f arm up.

 

[mtcn77]

And if you don't want the performance, just use your damn phone or tablet. There is absolutely no point in having the exact same limited performance as you do with a phone without it's only ad_vatage, which is mobility.

In a climactic turn of events, your overzealous appeal to pc enthusiasts letting down the tablet community has struck a cord with the banter squad that will now rain down upon your argument. Behold.

Your overconfidence has taken its shape in the grammatic oversight on the 'advatage' error and is just as faulty as the sum of all your pc mob mentality.

That's what you ARM advocates do not understand. ARM SoCs are designed to be power efficient at all costs, desktop chips aren't, they are focused around performance. A desktop ARM replacement will be "just as bad" as a x86 counter part if you want the same performance. ARM chips are no magic, they are just optimized for something else, looking at smartphone SoCs and extrapolating that to desktops is exceedingly dumb.

Well, arm is showing development. It might not be much, but it is where it is needed the most. I have 2 tablets, one a smartphone actually. Both, have hardware acceleration errors; however the quality of troubleshooting in them are worlds apart. The smartphone which is harboring a later chip, soft boots. The tablet however goes clunky and loses fluid animation requiring a hard reboot. The issue is definitely a failed base speed bin execution, however both are faster than pcs. You might state they come with flash drives which tampers the comparison, but it never was about just the chip. Arm devices are well put together. They are in SoC arrangement, you should look into that sometime. I like picture in picture mode which is crazy coming from such a small device.
Also, I like not having to manage memory usage. Enough said.

 

[bug]

And if you don't want the performance, just use your damn phone or tablet. There is absolutely no point in having the exact same limited performance as you do with a phone without it's only advatage, which is mobility. No one wants that, you would be getting the worst of both worlds.

And yet there's a good chance Apple will sell exactly that.

 

[lexluthermiester]

First - because I don't like my x86 experience with AMD Ryzen - it lags in loading in many cases - for example when opening the Start menu - some parts of it take longer to load and it gets ugly how one part of the Start menu is visible, and the other needs another second to appear.

Second - because power consumption - imagine a smartphone-sized tiny computer cases/boxes integrated in the monitors or attached to the monitors with the latest Snapdragon ARM-CPU.
That will be perfect for things like - email processing, 4K viewing, YouTube, Facebook, video playback of any type, lite smartphone-type gaming, etc.

Most people aren't having your problems. The world does not revolve around you...

 

[mtcn77]

Most people aren't having your problems. The world does not revolve around you...

But, he didn't make an assessment. Your word is as good as mine on him. It is just an opinion. Not general, he didn't extrapolate.

 

[ARF]

Most people aren't having your problems. The world does not revolve around you...

That's not true. Most people are having much deeper problems with Pentium-class systems and HDDs.

 

[bug]

And yet ARM-based home systems should be perfectly viable and we must start using them and start replacing the old x86.

Because <insert ARF's problems here>

But, he didn't make an assessment. Your word is as good as mine on him. It is just an opinion. Not general, he didn't extrapolate.

Right, he didn't extrapolate one bit

 

[ARF]

bug, I reported your content for trolling.

 

[mtcn77]

Right, he didn't extrapolate one bit

It is contextual and within a frame of reference. It is thus reproducible. It is personal, but not outside the frame of a contextual remark.
I mean it is grounded in real world examples.

 

[bug]

It is contextual and within a frame of reference. It is thus reproducible. It is personal, but not outside the frame of a contextual remark.

"I have a problem, the world must change" is "contextual and within a frame of reference"? You must be a lawyer, methinks.