Zen is DOA

...

Other urls found in this thread:

ranker.sisoftware.net/show_run.php?q=c2ffcee889e8d5e3d7e6d3e2d0f684b989afcaaf92a284f7caf2&l=en
ranker.sisoftware.net/show_run.php?q=c2ffcee889e8d5e3d5e3d1e2d0f684b989afcaaf92a284f7caf2&l=en
ranker.sisoftware.net/show_run.php?q=c2ffcee889e8d5e3d5e3d1e2d0f684b989afcaaf92a284f7cafa
twitter.com/NSFWRedditGif

whew, lads
>inb4 engineering sample

That 6900K is most likely running at its actual turbo you faglord. So somewhere between 3.7 and 4Ghz.

translation?

numbers are universal, my user

so your point is Ryzen needs to turbo boost ~50%? Good save.

6900k's all core turbo is 3.5ghz, the zen es in this test was also 3.5ghz all core turbo (you can see it in the OPN).

Yes, I believe your poorly mspainted chinese crap.

How much and how fast RAM did the AMD machine have? Do the runes say?

3.5 is the 1 core turbo. 3.3 is the all core turbo
same clocks that French magazine had and that tested within 13% of the 6900k. Would be consistent with the 14.69% in OP's runes, whatever they mean.

Yes! Thanks user, that's a load off my shoulders, I knew that fucking french shit wasn't real, it couldn't be real, this is and proves Zen is garbage like I knew from the start.

I'll wait for the launch. Then I'll be disappointed when it turns out IPC is half of Intel's current gen.

dumb intelfag

>ES
>moonrunes

This is what I just scored with my 6900k clocked at 4.5GHz (AVX -2 multiplier):

MM Int: 967.9MPix/s
MM long-Int: 286MPix/s
MM Quad-Int: 4.32MPix/s

MM Single-Float: 889.47MPix/s
MM Double-Float: 509.1MPix/s
MM Quad-Float: 18MPix/s

4.5GHz/(967.9/724.89) = 3.37GHz

>amd Africa

Why does it say 3.2GHz/4.2GHz for the 6900K? Is it OC'd?

calcs here
put it at ~3.37GHz

Apparent engineering sample running at non-retail clocks, manages to be competitive anyway.

shit bait

SOURCE PLS

>competitive
lolwut? Did you look at the numbers?

>engineering sample
kys

why would anyone root for zen to bomb?

its like you enjoy being assraped on pricing from no competition

Shitty bait, only good enough to make me reply. Enjoy your (You)

>artificial mememarks

Isn't this supposed to be a $500 chip while the 6900k is over 1k?

>doubting this based man

sauce

Zen wil be shit like everything AYYMD in the last 10 years.

nobody knows

nothing so far can be considered solid other than what AMD themselves have shown us

>revisionism
f a n b o y
a
n
b
o
y

>nothing so far can be considered solid other than what AMD themselves have shown us
And that should be taken with a grain of salt.

Let's see some prices before calling it DOA

>people actually believed amd's lies and fudged numbers
>again

does anyone know how much the ISA extensions (say, AVX) affect these tests, and how different are these CPUs in that respect?

Arica

>ryzen
What kind of 12yo came with that name? It's like they asked someone in a twitch stream chat. What the fuck? How can you take amd seriously after this.

>fudged numbers
We saw real-time benchmarks friendo.

>Arica
Africa without a single f given

yeah you saw a video stream of benchmarks without any investigation into the underlying hardware, code for said benchmarks, or any other pertinent data. It's called marketing, boi, and was done for a pump and dump of their stock.

ZEN IS A FLOP
AMDKEKS ON SUICIDE WATCH

So you're telling me when they were having e-sports players playing and streaming at the same time, it was staged all along?
woaaaah...

Oh shit are you saying that its fast enough to play a game?

Sure hope so.

Fast enough to compete with Intel? Nope.

I didn't watch that one - did they show someone remove the cooling solution to show the actual chip packaging? Did they compare that actual game with the i7-6900k benchmark-wise? If you can play games on it for """e-Sports""" you sure as hell can show some real-time third-party-verified benches with it.

any cpu since core2 can encode stream-quality h264 video (720p, 60 fps) in real time

No, trust the numbers goyim

The blender benches straight up couldn't have been faked, they released the versioning and the file they were benching with.

S O U R C E

No, don't you understand? It's impossible for AMD to create a chip that is comparable to Intel's chip because it's AMD! I'm a smart person.

DOA?

prices are everything in competition friendo

checked

only poor people care about price

End your lonely existence.

>not being frugal
>not saving money if at all possible
>"IT'S 2% BETTER! WHO CARES IF IT'S 69 TIMES THE PRICE OF THE NEXT BEST THING?"
We don't even have any solid information yet, and won't until they start selling it. Trying to decide anything before is clearly just baiting or jumping the gun.

Enjoy living paycheck to paycheck, wagecuck.

Never trust Chinese leaks lads.

>Never trust Chinese
ftfy

Ateasl I got money in my pockets, unlike (You).

What does pic try to say?

...

>twice as expensive cpu doesn't perform twice as good
Intel BTFO

Anywhere from "not at all" to "orders of magnitude" depending on the workload you provide. This test appears to be one that mostly tests AVX (my jap isn't good enough to see what other things are tested), which mostly affects applications coded specifically for AVX and those compiled with optimization flags to take advantage of AVX (but even so, the app needs to have an easily parallellizable workload for AVX to have an effect).

To my knowledge, AVX adoption in consumer applications has been pretty poor. Games won't be affected by this, and neither will most other applications. The exception is highly specialized applications, or apps with easily parallellizable workloads, but you'd need to code them accordingly and compile with a compiler with support for the AVX version in use to take advantage.

I assume that "DOA" means " Death Of Amd" (?)

Dead on Arrival

Assuming they wern't using the hardware based solution in the GeForce expereance, that takes a lot more load off the CPU so you could easaly do 1080p @ 60fps on a core2.

Thinking Zen will ever arive, AMD will be bust befor then.

That would explain it. Ryzen is not good on AVX. There was some instruction throughput chart that shoed this.
Guess the reasoing is "GPUs are better at this anyway, why don't you do it on the GPU?"

not all numbers are equal.
you want to tell me that intel is twice as fast as amd after amd has shown real world performance fine, ill believe it, but I need to know if any of this is real world performance of it it's all synthetic best case scenarios.

granted, these best case scenarios are why I honestly believe amd won't charge 500$ or more for the base model.

At least this is how I see it.

GPUs do AVX and SIMD really really well if it's not very branchy code, but some stuff is still just going to be relegated to the CPU because of how it branches.

In this case, AMD is still only about 70-80% of Intel's performance for FPU operations.
They've pushed that up from around 55%, and their integer performance is nearly on par, about 85-95% now, and in rare edge cases they can beat Intel.

the french leak can be considered solid as they are VERY transparent with it and actively answered questions relating to it.

Its still an engineering sample but is a very good bare minimum expected performance.

Integer performance is also low in OP. There is just something wrong with the test setup in OP.
Does Sandra record actual CPU speed?

Performance figures for the i7 6900k in the OP are entirely fake.

An i7 6950X clocked at 4ghz doesn't even score over 600mpixels.
Looks like a desperate little intel fanboy through this together trying to disparage AMD knowing most people are too fucking dumb to look up bench results themselves.

intel shills are terrified of Zen.

>most people are too fucking dumb to look up bench results themselves.

ranker.sisoftware.net/show_run.php?q=c2ffcee889e8d5e3d7e6d3e2d0f684b989afcaaf92a284f7caf2&l=en
ranker.sisoftware.net/show_run.php?q=c2ffcee889e8d5e3d5e3d1e2d0f684b989afcaaf92a284f7caf2&l=en

OPs results are real.

Question is: what did they do to the engineering sample? Not consistent with any other leak.

>OPs results are real.
No, no they are not.

An i7 6900k at 3.2ghz does not perform better than an i7 6950X at 4ghz.
Not all submitted results are valid, you shit eating idiot.

6900k results range from 394.98Mpix/s - 779.50Mpix/s. There are a couple very close together at 394-405 , but the 569 posted in OP is already the 8th lowest score.

How does Sandra get such a wide range of results?

Much like Cinebench its an incredibly easy bench to spoof, and it does a poor job of reporting actual clockspeeds.

Because it will make a lot of people on this board really really assblasted.

Can we just ban subhuman AMD niggers from this board already? Their hardware is inferior every single time, they're literally shills with buyer's remorse that trick retards from Sup Forums into buying their garbage.

Because they're paid by intel.

>N-no this can't be happening! DELET

Fucking kill yourself you AMD nigger.

Sup Forums is designated shitting board for ayymd and MS poojeets for quite some time

Nice shitpost, kid. Stay in Sup Forums where you belong.

Shorting AMD stock

working overtime poojeet?

>OPs results are real.
Wut?
>legit benchmarks
>ssl error
If your benchmark suite isn't willing to spend $50 on an SSL cert why should I trust it?

chinks still run xp and IE6, newfag

>re-enforces my point

ryzen 3.14ghz

nuff said

>take my balls SO deep

SINGLE FLOAT VS INTEL

HAHAHAHAHAHAHAHAHAHAHA

HOLY SHIT THAT CANCER PERFORMANCE

DEAD ON FUCKING ARRIVAL

LITERALLY

BULLDOZER FX REHASH GARBAGE

lel they copied the score from the first column into the third. Can people really be this gullible?

row*

No, that's a result of the processor itself not being on the market; There's only 2 results available for that particular Zen ES, both of those benchmarks gave identical scores, hence the score range being what it is.

ranker.sisoftware.net/show_run.php?q=c2ffcee889e8d5e3d5e3d1e2d0f684b989afcaaf92a284f7cafa

newer SIMT handles coherent branching OK, and future generations seem to be focusing more on wavefront/warp shuffling to keep ALUs occupied.

CPUs will dominate simulation workloads near-term, but that's not necessarily guaranteed forever.

Lmao

If AMD somehow managed to make a CPU that outperforms a 1000$ Intel CPU, why would they sell it for 500$ or less? What would stop them asking for at least 800$?

When AMD had a leading product, as in dominated (almost) all categories, they charged more than Intel top-for-top product.

And they should, and will, price accordingly.

If AMD doesn't push 8 cores in the mainstream market Zen will be DOA.

Because the $1000 CPU will almost certainly drop in price once there's some real competition.

Lmao

LMAO

So there are 2 things that can happen

a) It's shit like originally expected, but reasonably cheap. But that doesn't matter if it's not good enough

b) It's on par with i7 extreme chips, but costs just as much, making it unaffordable (except for enthusiasts who most likely have a 6-10 core intel already)

>Sandra
>a benchmark that only ever gets recompield when Intel puts out a new architectur
>somehow remotely related to AMD

That and (as has been said) its easy to spoof results - much like cinebench.

Saved this.