Doesn't even compete again a year old locked skylake i5s

>doesn't even compete again a year old locked skylake i5s
HAHAHAHAHAHAHAHA

RYZEN BTFO

AMDDRONES ON SUICIDE WATCH (YET AGAIN LMAO)

Other urls found in this thread:

anandtech.com/show/9483/intel-skylake-review-6700k-6600k-ddr4-ddr3-ipc-6th-generation/3
globalfoundries.com/newsroom/press-releases/2016/09/07/globalfoundries-extends-fdx-roadmap-with-12nm-fd-soi-technology
electronicsweekly.com/news/business/485217-2016-09/
twitter.com/NSFWRedditVideo

>mfw intel shills realize their own 1000$ 6900k cant even into skylake single core performance either.
>mfw i have no face.

>engineering sample at 3.4 ghz keeping up with a released processor at 3.9ghz

>inb4 it's an 8 core
On that list, only BF4 supports eight threads.

>I don't know what an Engineering Sample is

The sample is rated at 3.15GHz, and throttles at 3.4Ghz.

>3.4Ghz vs 3.5Ghz
:^)

DELET

>using the smiley with a carat nose

I like how you blurred out the clockspeeds. Nice troll

>Le ebin maymay with blurred clock

0/10 try harder next time

Where this 3.9 meme came from? I've seen that hour long Zen showcase and their CPU was supposed to run at 3.4 Ghz.

>b-but muh ES! Muh clocks! Muh 8c/16t!
Fucking lol the preemptive denial from amdrones is delicious

Single core AGAIN

6700k reporting.
master race

but too be fair AMD's strong point is multi core.
so its non issue and its just engineer sample and its skylake teir that one is not running at 3.4 GHz its only running at 3.0 GHz

Lmao fan boys on both sides have no idea why ryzen is gonna bring back the cpu wars again.
good to see the i5 6500 get BTFO this is exactly what i expected.

stop posting kid.

Buldozer was also decent in multi cores. Nobody cared.

>testing on obsolete games
>no doom, battlefield 1...etc
>biased intetlshills rewiewing this cpu
>thats not good because lower price for me
>engineering sample

inshits not only on suicide watch, but locked up in ward on antipsychotics due to AMD STRONK

>Video gaming benchmarks for a CPU
What the fuck? Kill yourself fucking illiterate NEET piece of trash. Fuck of back to and never come back

DELET DELET DELET

Sinle core performance is the only relevant metric. You can always add moar coars.

Source?

This CPU is an engineering sample, it is underclocked and doesn't turbo boost

>etc..

Be honest shill, you literally named only two modern somewhat popular titles that favor AMD.

AMDENIAL

Actually the results make me happy: This time a high end desktop AMD CPU won't be beaten by a fucking Pentium.

>this thread for the 10th time

It's a gaming benchmark with an engineering sample at 3.15Ghz when we know the final version will be 3.4Ghz or better.
This shows Zen will have at least Broadwell-E single core performance. If the price rumors are true, Intel will have to cut all their processor prices in half.

Actually it's correct. The 4c/8t Zen was supposed to run at 3.15 with 3.4Ghz turbo.

Bullshit. 4 core will not run slower than the 8 core.

Bullshit. 4 core will run slower than the 8 core.

Don't make stupid assumptions. The guy who did this little ES review isn't even remotely close to a shill. Hes been going back and forth on twitter talking to people, sharing more information, including instruction throughput and latency. Also shared a bunch of details on the state of the mobo, the ES itself. They couldn't get a GTX 1080 to boot at all, and they noted some SMT and cache issues with the ES.
He picked games more or less at random for what he thought would be a good sampling representative of whats available and popular in engine behavior and scaling.

At the same time he explicitly stated that Zen was near Broadwell IPC. He gave it an incredibly fair and honest examination.
If he was a shill he wouldn't be saying any of that. He'd be pushing an entirely different narrative.

No.
3.4ghz is the minimum baseclock for Ryzen CPUs.

this has never happened in human history and you have no sauce.

Looks like my 2500k isn't going anywhere.

>1Ghz difference ES
>still kicks ass

AMDfags be like "But it wasn't supposed to be fast."

>The guy who did this little ES review isn't even remotely close to a shill.
isn't he one of CPU-Z devs?

If this would actually be released cpu performance I would be happy

>hypes up 480 as high end
>It's on par with mid range cards from nvidia

>Hypes up new CPU line as high end
>It's on par with mid range CPUs from Intel

I think I see a pattern here

AMD has always been good at multicore performance. I was hoping that Zen will bring great improvements in power consumption and single thread performance (at least at Haswell performance).

If only op didn't crop other 2 benches that showcased how much better ryzen is than intel at 1ghz lower clocks

How do you know it's not?

>3.1/3.4 GHz CPU keeping up with a 3.5/3.9 GHz CPU in mostly single threaded workload
I'd say it's performing quite well.

>between 6500 and 6600
I'm fine with that.

The 8 core Bulldozer CPUs have been on par with the i7 CPUs for a long time on CPU bound tasks like compiling, encoding and rendering. Ryzen will trump the i7 CPUs again for a long time.
Too bad this board is packed with 16 y/o gamers spreading misinformation.

By the time the FX 8350 rolled around the GCC compiler was pretty well rounded in handling the BD family of chips. Seeing the 8350 compile things faster than the i7 3770K was a pretty good showing for what the arch could do.
The difference now is that AMD isn't targeting intel's mainstream desktop chips, they're going right at the HEDT prosumer chips which is considerably more impressive.

Coming near a $1000~ i7E in a real world workload is fucking impressive for the first iteration of a brand new architecture. They're doing it with an equal number of cores and threads, without needing a huge clock speed advantage, and they're doing it all at a respectable TDP. Fanboy shills are just shitposting because this shows that AMD is legitimately back in the performance market. Such a strong showing from a brand new family of arch is incredible. We can expect future Zen+ to continue the trend.

>3.15GHz engineering sample
gee, I wonder why it runs like shit

>tfw 2600k at 4.4ghz
>thought I might finaly have a reason to upgrade
>see this chart
Nah I'll invest my money in a more powerful graphics card

Looking at performance per clock, and the fact that all but one game there has poor to no scaling beyond 4 threads, performance of the Ryzen ES is incredible.

schadenfreude

>comparing an engineering sample limited to 3.15GHz/3.4GHZ base and boost respectively to products clocking in 3.5GHz/3.9Ghz and 4.0GHz/4.2GHz.

>Is able to keep up despite the significantly lower clocks

That Zen IPC uplift is just fucking nuts. IIRC Zen's 8c/16t products are starting at 3.4GHz as the lowest base clock, so that's pretty exciting.

AMD is total shit, I can't wait until AMD is fucking dead

>can't even compete with Intel's mid-tier of last generation
>I'm fine with that

AMDfags everyone.

in singlethreaded muhgaymes from their benchmarks it seems to have somewhere between broadwell and skylake performance, assuming those skylake chips are running at their base (3.5Ghz) and the ES Ryzen is running at its boost (3.4Ghz).

If the skylake chips are running at their boost (3.5-3.9Ghz) then AMD has a strong chip at very near skylake performance single threaded.

>3.4Ghz chip performing between a 3.5Ghz and 3.9Ghz chip
Yeah I'd be fine with that too.

hey samefag, stop shitposting.

OP's post is wildly out of context at any rate, since they're not even bringing up that those benches were run on an ES and the AMD chip is capped at 3.15GHz/3.4GHz base/boost clocks.

Capped at a baseclock of 3.4GHz without turbo and it was a hair away from the 6900k (which, by the way, was allowed to turbo to 3.7GHz)

And not to mention they are going to have 4c/8t units that are going to clock much higher

All things considered, your response to the product is just blatant intel fanboyisim bullshit, where you have to fling shit for no reason whatsoever.

Did AMD rape your mom or give you a "bad touch"?

get over yourself.

Still lower performance per clock than the i7-6800K which is based on the ancient Broadwell architecture.

Zen has to compete against Kaby Lake not Broadwell.

Only performance for clock matters

Broadwell isn't ancient, it is only a few percent behind Skylake in perf/clock.
Kaby Lake has the same perf/clock as Skylake.
Ryzen isn't competing against quad core mainstream desktop chips.
It is competing against Haswell-E and Broadwell-E.

You shills really do grasp at straws.

Broadwell's HT is much worse than Skylake's.

No, no its not. Stop trying to pull things out of your ass.
anandtech.com/show/9483/intel-skylake-review-6700k-6600k-ddr4-ddr3-ipc-6th-generation/3

The shills are going all out for this one, it has them running scared.

>the truth is shitposting

AMD's new processors are objectively shit, just like always. Get AMD's dick out of your ass, they are fucking you over.

>3.15GHz sample 1% within 3.9GHz i5
>doesn't compete

okay.

>they're over
>ignore their massive leap in performance
>ignore that investors look at trends
>they're over

Nah, they're still not competitive in the high end market, but that hasn't been AMD's goal in a long, long time.

Do you own Intel stocks or why are you shitting yourself? It should be positive that we have competition for the almighty "almost a full percent increase" Intel. No reason to blur out info that would reveal the truth if you are actually objective about CPUs.

>Replying to bait

>gayming

Performance

Power draw

>10 days ago
>muh IPC

>now
>muh clockspeed

Intelfags on life support.

Damn now I don't know what I should get

Wow, it's only 15% slower with a around a 10% clock deficit.
Microcode updates and new mobos would probably net another few %

Neat.

What of these benches involved are heavy on the FPU besides Mental Ray? This is likely where AMD loses big, mixed workloads like Blender are much better.

Other than that, are there any programs out there in the wild that are purely FP that aren't HPC specific?
A fully FP consumer program sounds utterly insane considering the general skill of a normal programmer.

The latencies on AMD's FPU operations are pretty high compared to Intel still but lower than the latencies with bulldozer and derivatives, at least.

They would lose pretty hard to Intel on purely FP workloads, however afaik there's not much if any consumer workloads that would push FP that much, and HPC uses GPUs to a much higher degree for better FP performance.

Thankfully most game engines out there are much more partial to integers than fp, I'd wager around a 80:20 due to code maintenance and generally FP operations are more power hungry and aren't usually needed for game engines.

...

The ES is at 3.15GHz actually. It's slower than retail.

The FPU was one of the things rebuilt with Zen

Yeah, but it's still quite weaker than Intel's massive FPU they've been refining over the last 6 years.
Actually the FPU is the only real thing Intel has put in effort since Sandy Bridge.
That and power.

Eh, if it's cheaper than kabylake that could actually be a really good result.

That said, if it's not Intel CPU + AMD GPU will continue to be the best combination indefinitely.

FPU is useless because of GPUs

Don't be fucking retarded, GPUs can't do branchy code at all.

The entire core arch is brand new, they're not using the FlexFPU anymore. Regardless, the FPU isn't nearly as big as intel's, and it doesn't have the same throughput.
This has been known for about a year.

kek

How come x86 didn't stagnate for AMD which promised 15% IPC for Zen+

15% over Zen is 10% over Skylake.

Because the Core series is based off a heavily reworked Pentium Pro core, it's got nowhere left to go.

By the time Zen+ comes out Intel wil have cannonlake and coffeelake which will obliterate AMD

Neither Coffee Lake nor Cannonlake are bringing large IPC uplift.
Coffee Lake just brings two more cores to the mainstream desktop line(obviously trying to compete with Zen) and they're making Iris Pro IGPs mainstream.

Cannonlake is just perf/watt uplift from their 10nm process, if it ever materializes. Yields will be even worse than their troubled 14nm node.

7nm SOI FinFET Zen+ will have near parity with intel in everything.

Besides actually selling like it should, we can't have that or some Intel lobbyist might get very angry.

Core architecture is a dead end. Intel knows and is planning a replacement for 2020.

you're right about Intel but rather optimistic about AMD/GloFo IMO.

I don't even think 7nm will be SOI, just a separate 12nm FDSOI node last I heard.

What's the difference between SOI and FDSOI and what does the 'fully depleted' part actually mean

which is an even bigger deal due to the 95w tdp of the 8core vesus 165 of intel in the same SKU. If Zen yields that kind of performance at lower clocks along with being lower TDP at each clock rating then they have a monstrously efficient low power part that with slap the shit out of the U"series an Y"series intel i5/i7's

jebus imagine if those maniacs crammed a 6-core/12-thread zen chip into a 15-25watt envelope? or a 4/8 C/T chip into a >5watt TDP to go after the y-series parts in tablets?

They absorbed IBM's foundry business. They appended their WSA agreement to have access to IBM's 7nm node. They're going to use it.
They're not going to use 22FDX, or any smaller variation of it.

FD-SOI and PD-SOI are just different kinds of SOI.
Pic related explains what you're asking.

No one said Zen is going beat Intel. They said it was going to be skylake i5 to i7 performance for half the price. So far that seems to be accurate.

FD = the channel is so thin that the entire depth can be effectively void of mobile charge carriers.
All-around better performance (especially lower leakage) but apparently easy to fuck up the manufacturing of.

I don't want to rat on you, but honestly, this shit gets on my nerves.

Are you doing it just to fuck with people or are you getting a slice of the pie?

this sounds like a whole new set of bullshit.

7nm. just say 7nm.

12FDX wasn't announced all that long ago, but presumably somebody will have to want to use it.

I don't have enough of a sense of what it actually is (16nm FF + FD-SOI without any real shrink maybe?) to speculate whether it's for mobile, GPUs, or other applications.

> globalfoundries.com/newsroom/press-releases/2016/09/07/globalfoundries-extends-fdx-roadmap-with-12nm-fd-soi-technology

Everything in the FDX line is going to the be same, just at a different node. Its planar, UTBB, double gate effect leveraging body biasing.
The lack of IP for certain blocks is what limits its usability. As has been the case, and will likely continue to be, this process will probably be used for analog, RF, and other specialized circuits. We won't see a high performance CPU use it.

What are you even trying to say here?

> electronicsweekly.com/news/business/485217-2016-09/

> uses double patterning and 40% fewer masks than Finfet.
> 15 percent performance boost over today’s FinFET technologies and as much as 50% lower power consumption.
> delivers the same performance as 14/16nm Finfet and with back bias it is equivalent to 10nm Finfet
> The process development is six months ahead of schedule and the 12nm process is slated for production in early 2019.

basically unshrunk planar transistors that you can get "10-16nm FF equivalent" performance if you're willing to fuck around with active bias control

>biased intetlshills rewiewing this cpu
>those guy literally made an article about why intel is screwing up everything they are doing lately
>take care to tell the reader about why performances are this low so they shouldn't consider it as a real and definitive benchmark
yeah no

The reviewer plainly stated that Zen had near Broadwell IPC, and praised it throughout.

The level of rationale here has really dropped over the years. People just want to scream shill at everything if they think it doesn't fit their narrative. Even when a source is being totally neutral and objective. Its ridiculous.