Using 64bit counter in loops drops the performance by 50% on INTEL processors

Is this board for real? Are you retards actually fanboying over a piece of silicon?

Oh boy, you idiots sure are something.

Also it somehow crashes paper mario and kills a mac.

There is no way I'm going intelaviv on any future build

They haven't reach that part of the post yet. Remember, their CPU is 50% slower.

>reading answers on poopoverflow
nigga, i didnt even read the question, the title is more than enough for this thread

>Intel assumes you're competent LOL
Then why isn't Intel?

How do you write a compiler for the targeted architecture if you don't know ASM?

You transpile to C and then compile the resulting C code.

check out the difference in knowledge depth between the poster and the top answer guy

both are great but the first is just a good coder, the second is a???

Even more sad is that their shills constantly talk about the spikes in fps but not about the 1% or smooth performance. Even in games shintel is falling behind. Only thing that let's them brag about muh gaems is the fact tthat many devs specifically optimise for shintel. But optimisations for amd takes some time.

Attached: skylakex 50% slower.png (728x765, 115K)

>tfw open thread thinking it's a new bug where Jow Forums will discuss CPU architecture
>it's just an old false dependency bug in old compilers
>OP didn't even read his own link
>thread was just meant to be another meaningless flamewar between fanboys
That's just sad, I wish Jow Forums hadn't turned into a /v/ playground

pretty much all spinlock implementations will suffer lol

delid

Attached: 1529144072309.png (682x792, 339K)

Most languages "compile" to C.

Fuck you wofaggot you're destroying our board kill yourself now

Attached: 1494523232402.jpg (960x878, 125K)

/scroll down
> 20gb/s on xeon double precision
> 18gb/s on amd single precision

wow you sure showed intelfags

Using unsigned to do loops, what a fucking brainlet. What if the condition to end the loops is when the variable is less than 0. Unsigned will create an infinite loop cause the condition will never reach.

That's why you take CS classes, so you will program like a professional.

Kill yourself wofaggot

>Using unsigned to do loops, what a fucking brainlet. What if the condition to end the loops is when the variable is less than 0. Unsigned will create an infinite loop cause the condition will never reach.
>That's why you take CS classes, so you will program like a professional.
Are you fucking retarded? Using unsigned is a conscious decision when you know it cannot be a problem.

Kill yourself wofaggot now you disgusting newfag

Fuck your op you embarrassing wofaggot

Post another wofag and I'll kill you

This isn't about me being an Intel fanboy this is about you murdering our board
Fuck of wofag

Mods delete this nigger thread

>discovered in 2016
>still not relevant to this day

>[Deleted] [Deleted] [Deleted] [Deleted] [Deleted] [Deleted] [Deleted]
hitler would approve

OP your wofags are long dead kys

back so soon?

kys?
kiss your sister?

I think he lost his KEYS

hot keys steve

she called me a creep (._. )

go to /b/, they have kinky solutions for this kind of problem

>(._. )

Attached: 1485184748518.jpg (500x860, 219K)

>AMD isn't affected
[citation needed]

can i get a basic gestalt?
havent touched intel since haswell e and only just gone to zen2

>compiler patch fixed it
its fucking nothing

>still faster than poozen

Attached: 1523026947443.png (213x237, 9K)

Anyone know whether either this or the bug in the OP is fixed as of Kaby/Coffee?

>deliberately compromised CPU security for only a marginal advantage in certain areas
lol

>>reading answers on poopoverflow
>nigga, i didnt even read the question, the title is more than enough for this thread

state of Jow Forums in 2018

> Over 3.5 years ago
> Fixed by gcc
AMD is better but stop shilling
gtfo

stupid question

stupid answer no.1
stupid answer no.2

WTF is wrong with you people?
If you don't know ASM, you don't translate to C. That's what happened in the first C++ "compiler" which was something like a library that changed classes to functions and sctructures and then compiled the rest from C.

When you compile your code it goes to many stages before ending in machine code.
First it's the lexical and syntax analysis, which you will do, no matter what.
Then you will translate it to intermediate code, which you will also do.
and then either, GCC or LLVM will take that intermediate code, it will optimize it according to the given or default flags and then it will output it to the requested arch according again to your flags.

REEEEEEE pajeets, you don't even know the basics of compilers. Educate yourselves ffs. Stop wasting time here and sit down and write a simple compiler with your target code and your VM.
REEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEe

tomshardware.com/news/intel-amd-sweepstakes-8086k-threadripper-1950x,37320.html

Top kek

someone post a smug reaction su for me

>My for loop needs to go more than 2 billion times
>AMD fags will defend this
>AMD fags don't know what while loop is

>don't convert it to C
>convert it to C--
wew lad

>story from 2014
>somehow relevant today
nope

for(unsigned i = 100; i != (unsigned)-1; --i)
;

wew lad

this
Stop LARPing NEETs.

Excuse me, we don't mention His name here.