Ryzen 7 is a rebranded engineering sample

>>59406231
My dad works for Intel guys, 16 cores will be the new mainstream trust me guys,Intel won't let AMD deface them with a 8 cores, they will go 16

>>59406231
Sounds more like
>EARLY ADOPTERS BTFO

Every chip is a rebranded engineering sample.

You do realize that every CPU ever is a rebranded engineering sample right?

File: 1489192027636.jpg (8KB, 256x256px) Image search: [Google]

8KB, 256x256px

>the city of early adopters

why do they do this to themselves ? will they ever learn ?

probably the same fags that are preordering mass effect andromeda

File: 564564616.jpg (7KB, 232x217px) Image search: [Google]

7KB, 232x217px

>Being this fucking dumb

>>59406231
this is the dumbest thing i've read in a while.

>>59406457
poor ayyymd

>>59406371
>>59406375
>>59406453
>AMDfags in damage control mode

File: 1488205393827.jpg (71KB, 552x661px) Image search: [Google]

71KB, 552x661px

>>59406457
>P-PLS DELETE THIS

right so they're saying an r5, which is cheaper, binned at a lower clock speed and with fewer cores, is actually better than a higher clock speed more expensive 8 core part

some people will believe anything posted online

an overclocked four core ryzen part could perform better in games for the same reason disabling cores on an r7 yields better performance that being ccx latency issues

>>59406466
>>59406493
https://slashdot.org/story/16/09/02/1543234/intel-confuses-rebrands-some-core-m-processors-as-core-i

>>59406403
In another universe somewhere consumers are educated and companies struggle to sell new products since nobody wants to be an early adopting moron.

Source: me

Right.

ryzen is shit in general, binned, non binned, with or without optimization

>>59406231
So it's Phenom 1 with the TLB bug all over again?

>>59406653
I'm sure Intel would be thrilled with 15% perf from a new stepping.

But sadly, no.

>makes extravagent claims
>provides no evidence
yeah sorry this is grade A bullshit without any proof.

That goes for anything, if you want to make a claim you are under the obligation to prove it.

>>59406231
>AMD fags beta testing shitty CPUs for their own money
ahahahahaha

>>59406231
You guys believe every bullshit, don't you. No wonder what's happening behind the pond. And probably on this side of it too, bleh.

>>59406307
Close, actually it's 6.

File: totally legit.png (14KB, 601x283px) Image search: [Google]

14KB, 601x283px

>>59406484
I wonder why intel sells untested chips? Are we supposed to the the QC for Jewtel?

Even better, AMD has realized that every powerful component they have made in the last 10 years had two chips on it. So with R7 they cut out the middleman and just put two dies on the same chip.

>>59406937
kind of like every bit of software you install

File: amdloo.png (316KB, 882x758px) Image search: [Google]

316KB, 882x758px

>>59406231

>tfw all hardware/software/technology/idea/concept is a rebranded version of a work in progress that is completed

Sucks to exist.

>>59407245
Is that Temperature graph?

How will the CCX's of the 6 and 4 core parts work? will the 6 core have 3 on each or some have 3 3 some 4 2 and will the 4 core be worse with possible 1 3, 2 2 or all on one?

Or is it entirely different?

File: 1487577679181.png (197KB, 415x402px) Image search: [Google]

197KB, 415x402px

>>59407689
>two dies on the same chip

Why yes hello, intel is very good yes

>>59407985
3+3 and 2+2, which is honestly the best for most workloads, assuming that OS schedulers know what the fuck they're doing.

More cache is better than lower inter-core latency in nearly every use case except HPC simulation stuff and HFT, and not always even the latter.

>>59407992
R7 is literally is two modules on the same chip tho.

>>59407985
6 cores will be 3+3 but 4 cores will be one CCX. I read it in a french magazine AMD plan making quad core APUs with ryzen and the GPU will replace the other CCX.

File: 5b5a711e1cad26c75252a280c662f8f9.jpg (192KB, 1278x1678px) Image search: [Google]

192KB, 1278x1678px

WHEN WILL AYYMDPOOJEETS EVER LEARN

>>59408144
Dud you read it in the same French mag that said
>5ghz on air!1!!
?

File: 1479291323954.jpg (93KB, 978x546px) Image search: [Google]

93KB, 978x546px

>>59408172
Here : http://www.comptoir-hardware.com/actus/processeurs/32270-quatre-coeurs-zen-creent-un-cpu-complex-ccx.html

File: lolpng.png (149KB, 1120x977px) Image search: [Google]

149KB, 1120x977px

>>59408172
>5Ghz on air
>WC ryzen crashes on stage barely pushed to 4.1

Jesus fucking Christ, you faggots need to stop.
Reminder that even this shitty board notices a good amount of reports.

>>59408232
>momy! People are making fun of team red again!

>>59407946
uh, no. It's totally legit, real benchmarks from my ass... Can you not read?

>all this people literally born in 2009 that can't remember Bulldozer much less Nehalem
And that's not even going back to Netburst, K8, or the first Pentium
Software never supports new hardware decently on day 1, specially MS software
Also regarding the random bugs from mobo vendors do this retards completely forgot the mess that was X99 at launch? Or the plethora of random Skylake bugs that weren't fixed until recent steppings?
What the fuck is up with all this astroturfers, this is just retarded
>>59406653
No, Ryzen is performing pretty close to as well as it would even under the most optimum circumstances
That's why AMD isn't pretending Microsoft fixing their shit will bring in massive improvements all around
>>59408107
It's not, that's Snowy Owl
Saying the R7 is two modules on the same chip is like saying that Yorkstown is 4 modules on the same chip
The R7 is a single die
>>59408172
You can do that by disabling all the other cores, which is exactly what they said and did

How the fuck did they shat that thread? It was pretty decent
Now you have retards pretending that GloFo's/Samsung 14nm LPP is new, for fuck sake it has been in production for more than a year and half and has already yielded high performance parts in massive numbers, see Polaris
Fucking video game niggers

>>59408232
the fuck are you even talking about?

>>59408889
He and half of /g/ also said R7 could get to 5ghz on all cores with better VRMs and the final production chip.

>>59406937
>tfw R5 4c/8t master race

>>59406231
>>59406403
>tfw been waitfagging longer for asrock boards to come into stock
>tfw saving me from a lifetime of regret buying ryzen on launch

>>59406403

>missing the chance to invite girls back to your place to see your Ryzen

more for me I guess?

File: 1472356270080.png (169KB, 725x364px) Image search: [Google]

169KB, 725x364px

i5 is dead. Utterly destroyed.

>>59409370
Don't disturb the new amdrone narrative.

>>59411339
I wonder if these will come with B2 stepping

>>59406231
> Paid Intel shills blaming AMD for Microsoft's OS not being ready for Ryzen despite Microsoft knowing what they needed to fix for over a year.

>>59411527
>no L3

>>59406484
> Don't look at me, I'm totally not a paid Intel shill

>>59411543
That also happened with Preshot and Nehalem.

>>59411682
If people did that, I'm not one of them.

File: 1489152808881.gif (2MB, 320x240px) Image search: [Google]

2MB, 320x240px

>>59411339
>mfw cant decide if a 1600x run better in games then a 1700

>>59411767
it will, it clocked higher and basically will be 1800x in games

>>59411907
Also moar L3 per core.

>>59411339
pretty low clocks though

>>59411339
>3+3 cores
great, so the chance of hitting the penalty from switching CCX is even great with that.
boy, are the benchmarks going to suck for that one.

>>59412288
How so? Scheduler will be fixed with W10 Creators Update.

>>59412302
no, it won't.
AMD took the blame yesterday and said the win10-scheduler is fine, although it isn't.

obviously MS told them to take a hike, and now they're trying to save face until the first ryzen-revisions are available.

>>59412451
>win10-scheduler is fine
It handles SMT fine, so there's no Nehalem-like problem, it still failes to recognize different CCXs.

File: Screenshot_20170315_113004.png (84KB, 726x790px) Image search: [Google]

84KB, 726x790px

>>59412458
>>59412451
There's issues with load balancing, sure, the CPU itself could to fill up on one CCX before moving to another to avoid latency issues since Windows is incapable of that, but some performance can be gained by turning on High Performance power mode and disabled HPET and that seems fine for now, AMD is working with developers to fix issues on the major games that show fuckups (bethesda and fallout), so Microsoft doesn't have to lift a finger because Intel doesn't want them to.

Good thing for industrial/datacenter workloads this is a non-issue as they fill up the cores and stay there, and since nobody will be running Naples on Windows it's even more of a non-issue.

>>59412937
What the fuck is this? I distinctly remember AMD telling reviewers to turn off HPET and balanced mode for optimal performance and people only realizing this 2 weeks later?

Nobody actually BOTHERED to test with two different tickboxes?

>>59412956
Eh, no, better add some more clickbait titles.

>>59412956
You mustn’t click random boxes in Windows! It'll break!
Better leave it all alone.

>>59412937
This is all AMD's fault.

>>59412937
>AMD is working with developers to fix issues on the major games that show fuckups
yeah, no, they'll be laughed out of the building.

>>59406231
Hahahaha thanks beta-testers

>>59412451
nah, what they said is that scheduler is fine but something else is definitely wrong

>THEY FELL FOR THE RYPOO MEME

>>59412956
journalists, especially reviewers, are the laziest and most incompetent people around.

>>59406231
let's all thank OrangeKhrush of the Anandtech forums for this reliable non-speculative fact-checked information

>>59406231
So it is basicaly Nehalem? At least they beat the shit out of X99.

>>59412937
Well, thank god AMD went to bethesda to fix their games, hope they patch fallout 4, game runs like shit on a 7700K.

tries to shill

doesnt realise that every chip we have on this world is an evolution of an ES sample

intel fags are running out of things to shill

>>59413303
oh yeah because fuck you progress

lets all create a single unified cpu with super pipelines and rings along with core

because if you create something new that no one is supporting is BAD

and to think there are idiots that believe that jim kevin sorbo keller will actually fail

>>59407985
Just know that it will be shit for gaming.

I'm buying the 6 core cannon lake next year because of this

File: 1488919463324.gif (547KB, 256x196px) Image search: [Google]

547KB, 256x196px

>>59415508
>cannon lake
>10nm
>next year