AMD BTFO

AMD BTFO
RYZEN IS FINISHED

Other urls found in this thread:

videocardz.com/65913/how-fast-is-ryzen
twitter.com/AnonBabble

Source?

The results really only make sense if the quad and hex core are closer to 4GHz with XFR while the 1700X is running at base clocks, as far as their relationship to each other Ryzen CPU... but that doesn't make sense with their relationship to Intel CPUs, because we have the example where the 3.4GHz 8-core Ryzen is awfully close to the 6900k (which is actually probably running 3.6Ghz, unless turbo was disabled).

videocardz.com/65913/how-fast-is-ryzen

>Not posting the other pic

Holy shit I'm scared, what does this mean for Intel??

...

Uhh so its IPC is close to Skylake?

Gee thanks for stating the obvious and a Kaby Lake clocked 1.3GHz higher is faster by a significant margin?

>single core benchmarks
what is this? the 90s?

not much.

At the end of the day, the higher turbo speed will mean more to most people than having more cores will.

I'm kind of curious why the per-core performance isn't scaling up with the GHz, but I guess XFR has something to do with it. Though one way to look at it is that this is simply demonstrating diminishing returns with number of cores in the Fire Strike demo.

trash

Intel about to get BTFO really really hard

>mfw waiting for Ryzen is going to pay off

>same IPC
>has 2 times more cores
Okay? What are you trying to prove? That Ryzen will be good?

Intel: 3893
AMD: 2545

AMD BTFO
B T F O
T
F
O

I don't give a fuck faggot

Show me benchmarks next to the 4790k and we shall see if it'd worth an upgrade.

Couldn't care less about meme lake or kameme lake. Cunts

>playing shitty boring version of Supreme Commander
Hoyl fuck.

>6900k and 6950x even lower
Feels extremely good. I can finally ditch this 3570k.

>literally Broadwell IPC
>bad
Per core performance still beats all those i5-2500k @ 4.4Ghz people are waiting to replace.
100% more cores are worth an upgrade. 30% more performance per core was not.

>4.0 turbo chip has same per core performance as a 3.2 turbo chip of the same arch

Yeah, legit as fuck

>3 ryzen chips scored nearly identically regardless of clock speeds
>lowest scoring one had higher clocks than the top 2
what the fuck?
Seems pretty rigged imo.

I already knew its fucking IPC, I want to know how much it can overclock.

amd has been working on this for like 10 years and still cant beat intel .....

>unnamed ryzen chips compared to intel chips all at various clock speeds

aite then

One would expect, that shilling is done in a way, that it can't be identified as shilling, yet the shills even developed a signature

It means that Intel is selling CPUs for 1100 dollars that are outperformed by a 500 dollar chip.

In other words Intel's precious profit margins just got fucked really hard

lmao

I remember some AMD kiddos telling me that the single core performance would only improve from the early downclocked samples.

Guess they lied - again.

Intel's training regiment doesn't teach subtlety.

AYYYMDPOOJEETS ON SUICIDE WATCH

3.3, 3.2, and 4.0 clocked chips have the same score

?????
Is that 4.0 broken or something?

the person who made OPs graph got it by doing "math" on this graph

That's even stranger, the 6 core at the bottom gets 50% more performance from 50% more cores.

That scaling is impossible.
Not even GPUs begin to approach that scaling, much less fucking CPUs

Intel: almost 5 nigghurtz
AMD: barely over 3

This test looks shody at best, clocks and performance don't match at all, its almost as if the 8 core with more cache is significantly slower per clock than a 6 core, which makes about as much sense as Nvidia open sourcing their drivers.
Then there's the linear scaling..

Something's fishy here.

OP sucks cock

>For the above chart, physics scores were divided by the number of cores each CPU has.
This is the shit you get when you trust Videocardz for CPU benchmarks.
OP's chart is literally worthless.

>per core performance comparable to proper i7s
>amd-approced core count, which seems to be four times as much as software developers has already adapted
i'm so buying this

Will there ever be an honest use for graphs and statistics?

>Show x has bigger number
>Ignore context

"HAHAHA MY CONSUMER CHOICES WERE GOOD, THEREFORE MY TEAM IS WINNING AND THUS I AM WINNING"

OP does indeed suck cocks.
Ryzen at 4ghz across all 8 cores isn't going to be a 95w chip though.
And to make it an honest comparison the all core turbo speeds of the Broadwell-e chips should be listed. The i7 6900k runs at a solid 3.5ghz on all cores, thats its normal all core turbo. Its effectively a 3.5ghz CPU in multithreaded benches and workloads.

>OP's chart is literally worthless.
Yep. Thats why he posted it instead of the actual scores.

Do my eyes fucking deceive me? Is AMD's 2x128bit FPU about the same speed as Intel's 2x256bit FPU?

That's insanity, either Intel's bullshitting somewhere or AMD simply has a far superior branch predictor and scheduler.
There's no way a FPU with half the width is as fast as one double its size.

What is this retarded bullshit? How can the same architecture have the same performance at 3.3 and 4.0GHz?

pls explain in stupid

Doubling the width of an FPU doesn't mean you double throughput in everything.
Zen isn't lacking to any large degree in 64bit or 128bit ops. AMD built it to excel in 128bit ops, just not anything larger.

>4GHz AMD part does worse than the 3.3Ghz part

No drivers.

>Ryzen at 4ghz across all 8 cores isn't going to be a 95w chip though.
That's why he put a plus sign there, you turbofaggot.

Accurate clockspeeds aren't being read by the program.

>I took the liberty of dividing scores across cores from here using """""""""""""""math"""""""""""""""" : the chart

What in the fucking fuck?

No its not. He put that + there because its listed on the page, because its listed by the product. It refers to the cooler needed to utilize XFR.

nice try Goldberg

DOES NOBODY FUCKING SEE THE LAST TWO RYZEN CPUs, HELLO PEOPLE, THERE'S TWO EXTRA CORES ON THE M6 AND ITS 50%+ FASTER THAN A FOUR CORE

THAT'S IMPOSSIBLE, THAT'S 50%+ scaling with 50% MORE HARDWARE
HELLO PEOPLE THAT'S FUCKING IMPOSSIBLE, PHYSICALLY
IN THE FIRST PLACE NO FUCKING APPLICATION ON THIS PLANET SCALES SO LINEARLY


BULLSHIT CHART AND CAPS THANK YOU HAVE A NICE DAY

Nice try, but the topic at hand is whether it's clearly marked in that picture that the chip is going to be 95+ or just 95W -- not "where did the plus come from".

But 512bit ops are everywhere, like in CAD and protein mashing and calculating Pi and. and.. and

what if it's not that the extra 2 cores give it 50%+ performance, but rather the lack of cores gimps the performance by about 50%? it becomes a bottleneck rather than impossibly linear scaling

You guys are so fucking stupid
Can you all just shut the fuck up about AMD BTFO INTEL DEAD IN THE WATER LMAOOOOOOOOOOO AMD WINS INTEL STILL BEAT ROFL CHECK'EM
Filtered

That's nice and all but am I going to have to buy liquid nitrogen to cool off this nuclear reactor of a CPU?

FPU = Floating Point Unit
and the intel one can work on a 256bit floating point number in one go instead of needing to fetch it twice like the amd one
that's why he's surprised that the amd one is just as fast

You buying a 140W Intel? Nah you can do with any cheapo $30 cooler.

You don't understand the point of contention.
"95w+" is a listed AMD has used for their 95w XFR enabled parts. They're still pulling 95w under load, but need more cooling to keep the tcase lower to provide thermal headroom for the XFR frequency boost.
The user in question didn't put 95w+ on that bar to show it was pulling more than 95w, he put it there because he read it on the top of the page on the videocardz article. An overclocked Ryzen chip with 4ghz on all cores would be pulling more power.
Using that in comparison to the TDP of the Broadwell-E part is dishonest so I pointed it out.

>CAD
>using AVX3.1
Intel wishes.

I haven't honestly been paying any attention to PCs at all since I built mine. I have an i5-4690K is this gonna be a better bang for my buck CPU or what is the hype exactly?

thanks

Or you can just wait for a bit, get the 6 core, overclock it to the moon and be set for a few years?

Ayy lmao this just goes to show how the Internet isn't reliable.

Guys, I'm scared.
What if AMD make a 12 core die next year?

LOL AMD

>What if AMD make a 12 core die next year?

PC's will develop AI themselves, and wipe out the human race once and for all.
(except cute stocking ''girls'')

A traps world....we are fucked....

OY VEY!!

>hardware is hot again
AMD HOUSEFIRE CONFIRMED

DELETE THIS NOOWWWWWWWWW

what you missed is that the 6950k is a 1400$ cp and its getting btfo and this is only per core test

>and this is only per core test
Except its not even that.
The retard who wrote the article simply divided the scores by number of core each chip has, then put it on a new chart to make it look official.
Its just bullshit.

400$ vs 1400$

This.

Anyone seriously rooting for Intel needs to be flogged until their bodies bleed lard. They've been dry-raping our asses for years because of fatnecks who circlejerk over benchmarks. AMD being even remotely competitive while hitting a VASTLY lower price is going to be a win/win

Even if you're an Intel shit, why wouldn't you want them to push themselves to make better/cheaper chips to compete?

shill btfo have fun buying a slower chip for 1400$

yes its good for the market and good for consumers

>how can u intel bois even compete?

Should I buy Zen or wait for Zen+ or whatever it's called? I don't want to upgrade for another 5-7 years.
I was thinking about getting the 1700(x), but if they release a chip that overclocks better or has more cores for the same price, I'd rather wait a bit longer. Is Zen+ going to be 14nm or 7nm?

Do you think they've banged?

Zen+ is going to be 14nm. The 7nm node won't be entering volume production that soon, and AMD has stated new arch will come out every year.

OOOOOOOOOOOHHHH INTEL BTFO CHERRY PICKING DAMAGE CONTROL OH SNAP

This isn't fair, how can AMD do this to people? They paid $1000 for 8 cores.

please anticipate

People should consider a brain and they won't waste money on obsolete shit.

Props to those who actually needed the throughput of 8 cores, they're the minority.

Doesn't half your CPU get used up by GFX driver overhead when you use AMD?

...

Never played a DX11 game with an AMD card?

there is a strong chance that zen+ will be a drop in replacement.

>memes

pajeet said they're not making a new socket until DDR5

at least it's not 80% like intel/novidya HUEHUEHUHEUHUE

Yes I know, user, just saying that with a bios update there is a chance that it will just work.

>a chance
most likely there wont be AM4+ if they say they're not making a new socket, since making changes like that is technically a new socket

The reason I say most likely is because they haven't explicitly said it.

There is no certainty, regardless of how much sense it makes, therefore, "a strong chance"

true they havent said that their next cpu line will be drop-in compatible with first gen mobos, but it's pretty much implied if they say "we wont make a new socket until DDR5"... they have never said something like this before so I would bet it will be just slap it in and go

How new are you? They say all the time that a certain socket will be around for a while.

Amd has been pretty consistent on this, but sometimes the newer CPUs will be drop in replacements and sometimes they won't.

For instance the original sabertooth 990fx just needed a bios update to support the FX 8350.

I had a previous am3 motherboard that housed my phenom II 955 but even with a bios update wouldn't support a 1100T

>How new are you?
not as new as you apparently cus you dont even understand basic fucking english

they never said the same thing about previous sockets as they did this one.

Funny with that lineup was ryzen againt intel high end cpus. Not a single i5 appear there

Because i5s aren't competitive in anything but very lightly threaded tasks.

>i5
Because something as simple as Battlefield 1 totally annihilates i5 chips

Yes they did child

Will there be an AM4 sabertooth?