What happened to parallel buses 13 years ago is happening to x86 now. Just like parallel buses hit a wall and serial ones eventually caught up, RISC CPU architectures like MIPS and ARM are catching up with x86 now and will soon supplant it.
Nope. Desktops, laptops, and the vast majority of servers still use x86. RISC is for tiny devices and embedded devices.
>>53548713
>x86 is kill
As long as 64bit OSs are useful, it's not going anywhere.
More like x86 is starting to surpass the power efficiency of ARM and Intel is about to start eating Qualcomm's lunch.
ARM processors are nowhere near close and nowhere near catching up.
>>53548713
Unfortunately the millions of x86 applications won't allow x86 be easily supplanted by RISC.
Finally. A whole new architecture to exploit. I can't wait.
>>53549149
not like current hardware isn't back(or bug-)doored left and right
>>53548713
Try playing a youtube video on a rasbperry pi. Not the original, the Pi2 or Pi3. Pi3 plays it but only up to 720p. This is the easiest way to demonstrate the massive processing power difference.
>>53548713
When ARM catches up on Intel/AMD on power efficiency and power, then we can talk about "x86 kill"
>>53549179
None of the Pi have acceleration driver.
There's newly experimental mesa GL driver but still not enabled by default.
I don't know yet.
>>53548962
>implying x86_64 is the only 64-bit architecture
>yfw most modern x86 is RISC anyway
>>53550452
>yfw if you're writing ASM you're still stuck with CISC
https://web.archive.org/web/20140101093853/http://semipublic.comp-arch.net/wiki/RISC_versus_CISC
Take it from Andy Glew, he's the lead architect for MIPS.
>>53548713
RISC / MIPS and x86 are both approaching 40 years old...
>>53548944
>being this wrong
you realize intel passes CISC to RISC internally..
they wish they could switch completely, but it'd take too much time
>>53548713
x86 and AMD64 cannot be killed at this point in desktop computers department, but ARM and AARCH64 might soon overtake them in embedded computing (actually they already have) and headless machines. The reason being, ARM and other RISC CPUs are able to provide upto 70% of the processing power of an x86 counterpart, at magnitudes lower power consumption.
This '70%' figure is growing constantly
>>53549076
There are more ARM CPUs being used everywhere, they're far more in number as compared to x86 and AMD64 CPUs. Probably even your home IR motion detector runs on an ARM chip, so its assured ARM isn't going to die anytime soon
>>53550376
Alternatively:
http://forums.anandtech.com/showthread.php?t=2202232
>>53549210
>efficiency
ARM is many, many times more efficient. >>53550452
this
will meme'd my friend
>>53550510
>they wish they could switch completely, but it'd take too much time
There's no reason to go full RISC. They would lose instruction flexibility through it. They are however working on integrating FPGAs into Xeon chips for dynamic cpu restructuring.
>>53550525
The ISA isn't efficient, it's down to the fact that ARM has always been targeted embedded low power devices and x86 has always targeted higher power devices.
http://research.cs.wisc.edu/vertical/papers/2013/hpca13-isa-power-struggles.pdf
>>53550525
It's not technically more efficient, they just build them in lower power brackets. Instructions per clock cycle are higher on x86.
