Well, /g/?

>>58920630
>is on borrowed time
You wish.

>autism

>>58920636
Hello Pajeet.

X86 is going no place and the only reason you want it gone is so you can drink gamer tears if you don't want to support it don't

>>58920688
Wow, Intel shareholders out in full force tonight!

>>58920696
Give me one good reason to replace it with anything on the desktop

>>58920704
I don't have to. The desktop is dying.

>>58920728
>This old falsehood

The laptop is dying as phones become more useful the desktop once again isn't going any where

File: impending_death.jpg (74KB, 1153x753px) Image search: [Google]

74KB, 1153x753px

>>58920747
You just keep telling yourself that, pajeet.

>>58920776
PC gaming just had one of its best years for hardware sales and as long as mobile phones can't into good games no one's going to care about them as a gaming platform no one games in their phones anymore that fad is dead and will be till they figure out a better way

Now once again give one reason why x86 needs to die

>>58920776
citation needed

>>58920630
What is Pentium?

>>58920630
ARM

What is x86?

>>58920630
Show me a CPU with better single thread performance. I'll wait.

>>58921614
https://en.wikipedia.org/wiki/DEC_Alpha
This would still be used today if not for Compaq buying DEC. Also the most powerful supercomputer in the world uses processors based on the Alpha.
http://preshing.com/20120208/a-look-back-at-single-threaded-cpu-performance/

>>58920630
What is x86?

>>58926374
>having to ask this question

Go back to >>>/v/

Does anybody have a link to a full history of x86 licensing manufacturing and legal battles? Just out of interest.

Yeah, lets expect people to jump to a new instruction set and a whole new set of software, every time a new CPU releases.

>>58920807
>PC gaming
no one gives a shit

>>58926420
Hey fuck you cancer. I'm not him, but no one should ever be made fun of for asking a question. If you don't know you don't have to resort to shitposting.

>>58920630
What is x86 my man

Is RISC always superior or am I stupid?

>completely backwards compatible with any program written 25 years ago
>"would have been deprecated"
Reminder that IBM still uses the same ISA in their mainframes as they did in the 60s. Another reminder that those mainframes won't go anywhere because muh speed and muh legacy code that runs the entire financial sector.
Literally the only way I could think of IBM being completely fucked is if the government stepped in and deemed all current financial businesses to be monopolies and forced them to give up their business in favor of server farm loving faggot ass startups, like Credit Karma.
Also, x86 ain't going anywhere because of said server farms.

>>58926420
>he doesn't know how to play Jeopardy
I'll take "Shit newfggots don't know jack about" for $500, Alex.

x86 isn't going anywhere for a long time. You may not like it but it's functional and just works.

>>58926503
https://en.wikipedia.org/wiki/X86#History

https://en.wikipedia.org/wiki/List_of_former_IA-32_compatible_processor_manufacturers

https://en.wikipedia.org/wiki/List_of_x86_manufacturers

>>58920688
x86 is already on the way out. x86 is dead on phones and tablets. Wont be long until we start to see ARM laptops. We are starting to see ARM servers. With MS working on x86-on-ARM for Windows it must be a scary time to be Intel.

>>58928728
x86 will be around for a long time just because of all of the legacy applications that will probably never be ported. Running them on ARM would probably be inefficient if not a complete downgrade performance-wise. The average consumer probably would mind a huge loss in performance but on the corporate side all of those losses add up to higher energy consumption and lower performance. Those are going to be the most stubborn customers.

>>58930844
>a long time
January 2038 will kill it unless you use a 64bit time.h

File: .png (79KB, 556x454px) Image search: [Google]

79KB, 556x454px

>>58923713
Not sure what are you trying to tell me(the picture from your very article) other than that old x86 CPUs sucked, today's Intels are literally 8x faster. They wreck DEC even when adjusted for clock speeds.
>supercomputer
We are talking about single thread performance.

File: UjNW8oM.png (74KB, 620x453px) Image search: [Google]

74KB, 620x453px

>>58932039
there were some slightly updated versions of those plots made (still ~3 years old though)

what's up with all the x86 hate and arm shilling lately?

something tells me those shills don't fully understand what they wish for, it's change for the sake of change and exactly zero benefits

File: W0Ooxcd.png (145KB, 1200x806px) Image search: [Google]

145KB, 1200x806px

>>58932967
actually, an even more recent version was posted, and anybody who wants is able to generate new ones:

> https://github.com/preshing/analyze-spec-benchmarks

>>58932998
Seriously, ARM platforms are already a nightmare. They're all closed, locked down bullshit with few exceptions.

>>58932039
>>58932967
I'd love to see two versions of this:
- one that is up to date until 2016
- one that only counts IPC alone. Just slash all performance by the MHz values.

Because most, if any, of the single threaded performance we got ever since Sandy Bridge was down to increasing clock speeds (stock Intel chip is 4.2GHz now today fuckdamnit), and specialized hardware instructions for some tasks which are complex and expensive to do without (AVX for example).

There is too much software for x86 processors to change it to some other ISA.

>>58933020
Funny how the graph gets flat around the Pentium 4s, then does a huge jump at the Core 2, and now it goes flat again for recent cpus.

>>58932998
>and exactly zero benefits
Excuse me?
In what world "not being completely locked to two vendors" is not a benefit?

I'm not even shilling for arm, I would rather see risc-v, openrisc (no chance because muh gpl) or powerX but point still stands.
Competition drives invention and brings down prices.

If it has to be ARM that takes it down, let it fucking be ARM.

>>58933032
>I'd love to see two versions of this:
>- one that is up to date until 2016
>- one that only counts IPC alone. Just slash all performance by the MHz values.

how about you make and post one?
>>58933020
>> https://github.com/preshing/analyze-spec-benchmarks

>>58920776
>charts on 4Chan
Go ahead and post the source, champ.

>>58933092
ARM would eventually do the same thing. Someone just has to buy out ARM Holdings and then not care about licensing out the IP and then we're stuck with whoever is currently in the market. Then over time the individual companies are just going to get bought out or leave.

Everyone switching to ARM is not a proper solution it just puts it off for a little while longer.

>>58933092
>In what world "not being completely locked to two vendors" is not a benefit?
In a world where every vendor tries to lock you in with denying you the ability to install custom software on your own hardware.
In a world without standards, arm ecosystem has probably a decade to go for it to be at least comparable to x86/PC in modularity.
In a world of SoCs and custom unmodular hardware, with every system designed to be a single unserviceable monolith, subduing the competition.
In a world of app stores, walled gardens, planned obsolescence, treating devs as slaves

that is our reality, that is ARM "ecosystem"

you truly don't understand what you wish for
similar concept has been tried in laptops, and the result is the trainwreck in an ocean of crap which is the laptop market today
who benefited of that? users? nah, only corporations

> risc-v, openrisc
ISA is not the point, architectures also aren't, ecosystem is.

>>58920776
Now show a graph in terms of actual units sold

>>58933206
So an already established AMD-Intel monopoly is better than other vendors trying to do the same?

Got it senpai

>>58933092
ARM today is literally stuck in 80s, it's the wild west of incompatible home computers all over again, except they are all locked down

>>58920776
where's the bar for servers?

>>58920655
>Getting mad about the only CISC instruction set
Great job retard.

>>58933235
Not the same but much worse. Change for the sake of change, as I said. It will only benefit the companies, users will only get a huge uroboros of crap and not the competition.

>>58920776
>Thinking Desktop V Mobile matters
There's a reason Intel is switching new features to their Xeon line. Dumbass.

>>58933235
>AMD-Intel monopoly
It's called duopoly.

>>58920630
wait but when AMD brought 64b to home, intel already had 64bit servers, itanium right?

isnt itanium in the same family of ibm-pc-at architechture?

whats so complex about x86 x86-64 that its so much more expensive than ARM 64bit?

mobile ARM is just slow because it doesnt have as much paralel floating points and to-RAM-bandwidth and cache as x86 usually do right?, but what makes it so 100x more transistor dense and therefore expensive and heat inneficient?

is it all those MMX, SSE1/2 SSSE1/2/3 legacy sets that are kept till today and also some sets like virtualizing, that most of the time a home user could do without?


what if GNU, AMD and torvalds were to release a x64 CPU, compiler and kernel using just one set of 64bit ready instructions in a super efficient CPU?

pls explain why am i a retard

>>58933235
>>58933334
Triopoly? They may not be trying but Via has an x86 license.

>>58933350
>isnt itanium in the same family of ibm-pc-at architechture?
No.

>>58933235
Starting everything from scratch for dubious reasons is worse than keeping what we already have

>>58933360
VIA has a very narrow segment of the market,basically all they make are embedded processors and even there there's not much going on with them. I actually don't even know what they do for money at this point.

>>58920630
>loaded OP
fuck off phonetard
>>58920776
market contraction because there's no reason to upgrade as often anymore != total death

products also do not have to sell more than another product in a totally different market to be profitable or otherwise successful
>>58923713
the alpha was really only good for floating point maths and was starting to lag behind by the time DEC was purchased by Compaq, it was also expensive as fuck and nobody bought it for desktop use even though it still ran Windows

sunway is not based on the Alpha, merely inspired by it
>>58928091
you have no idea what you're talking about really
>>58928728
there isn't a single actual ARM server out there that isn't vaporware

>>58933360
there's also some company which makes radiation-hardened x86 processors, forgot the name

libc done in hardware when.

>>58933447
It's one of the defense contractors, probably Raytheon, Northrup Grumman, or similar.

>>58920630
Intel Itanium

File: 1484960018771.jpg (250KB, 924x1200px) Image search: [Google]

250KB, 924x1200px

>>58933350
pls stop

>>58933350
>is it all those MMX, SSE1/2 SSSE1/2/3 legacy sets that are kept till today and also some sets like virtualizing, that most of the time a home user could do without?
Instruction set extensions do not need to be implemented in hardware. x86 manufacturers will implement little used and/or obsolete instructions in microcode rather than in hardware to save die-space.

>>58933314
>implying x86 is even still a thing on servers

>>58920776
i see that as general populace have become so informed about building their own pc instead relying on a branded already built pc sold at rip-off price. People are getting smarter today about buying things with the most bang for the bucks.

look at the sales of gaming computers component, its booming and nothing stopping it