So, who here is excited for AMD to Make x86 Great Again? >LOW

>>58862720
if this is true then intel is truly finished, but i'll wait for the benchmarks.

>>58862758
>if this is true then intel is truly finished

Yeah maybe they'll drop to an 85% market share. Truly the end

Remember when AMDrones were saying the RX 480 would be cheaper, more power efficient GTX 980?

File: 1481763072137.jpg (56KB, 907x661px) Image search: [Google]

56KB, 907x661px

>>58862819
What retard was saying that

It was clearly an RX 380X replacement

>x86
It should be a crime to even make Dedicated Desktop CPUs just for that architecture.

>>58862720

All of the IntelShills at work be like:

>About time AyyyMD!!
>Muh competition!!

>>58862819
8 core ryzen at 65w meanwhile intel offers a 140w housefire.

>>58864085
It's x86_64

>>58864327
Good, good, I don't mind supporting it because of backwards compatibility.
I still absolutely hate anyone that makes dedicated OS for x86.
I think that completely disappeared in Win10, you can't install Win10 on a dedicated x86 CPU because there is no x86 OS version of Win10.

Talking out of my ass right now, didn't fact check if this about Win10 is actually true, read it somewhere once and never actually tested this to see if it's true.

>>58864386
when was the last time there was a cpu that was x86 only?

File: 42-6900K-Power-Consumption-Torture.png (20KB, 600x450px) Image search: [Google]

20KB, 600x450px

>>58864258
oh.. wow..
It actually does draw 140W on just heavy FPU loads...
I though it was just the AVX extensions that warranted the 140W headroom.

>>58864417
>213 watts
JUST

>>58862720
yeah but i got my 6700K for under $150

>>58864492
>4.3 GHz OC
AND HOW

File: meme677889.jpg (35KB, 1280x720px) Image search: [Google]

35KB, 1280x720px

>>58862720
>mfw when the guy who built the AMD64 built the Zen
Lets fucking gooooo

>>58864386
Nice post number

>>58862720
Nobody is going to buy an AMD with that price no matter the performance and cores

>>58864584
If I can get my hands on an 1800X for 500USD, I will definitely buy one.

>>58864404
Just checked it, quite some time ago.
The last one was an Intel Atom CPU from 2008.

Retard here that never OC'd a CPU before, and quite honestly this question belongs to a SQT.

How worth it is it actually to OC a CPU?
Isn't the Turbo Boosting feature of Intel the same shit, just that it's variable depending on what you're actually doing?
Because so far as I see it, doing OC just sets a static clock instead, making it always perform at the highest possible clockspeed someone set.

Feel free to call me a retard if I'm wrong.

HAHHAHA Intel niggers on suicide watch and full damage control. Not the reasonable consumers fault that you love to get shafted by anti consumer practices.

>>58864712
Screen caped so you can keep that suicide watch feel for years to come when Ryzen turns out to be shit right next to Vega.

>>58864728
Keep talking shit and paying bricks of cash for garbage that ain't worth it. Meanwhile I'm going to drown myself in the Finewine and money I saved.

>>58864728
Screen capped this shit for when you kill yourself and Jim Keller teabags your corpse

>>58862819
/thread

>>58864673
Turbo has a maximum that it will go to, and there are algorithms that make your CPU drop down from it even when it doesn't need to. It's not perfect.
OC'ing not only lets you set a fixed clock, that the CPU will not drop down from, you can set it higher than the max turbo by increasing voltage at the cost of excess heat output and power consumption.

>>58864936
>OC'ing not only lets you set a fixed clock, that the CPU will not drop down from
>Not ocing while leaving C1E and C-states on

This isn't 2007 anymore anon.

>>58864947
2/10. Got me to reply.

Sorry for those that don't know, I neglected to specify "does not lower clocks UNDER LOAD". Turbo will still have the clock rate frequently dropping down from max turbo under load.

>>58864417
They tend to suck up a lot of voltage very quickly past a certain point, and that point is pretty low. If you overclock Broadwell-E to 4.2ghz, god forbid any higher, you should probably invest in a water loop.
Anandtech had their Broadwell-E chips run into stability issues under load from a lack of voltage. It was starting to get pretty hot under their tower cooler at just 4ghz, and seeing how it failed the test run at those settings, it needed more voltage than they were providing it, meaning it would have ran even hotter. 200w~ and 80c load
http://www.anandtech.com/show/10337/the-intel-broadwell-e-review-core-i7-6950x-6900k-6850k-and-6800k-tested-up-to-10-cores/10

This is the cooler they were using
http://www.coolermaster.com/cooling/nepton/nepton-140xl/
And reviews show its at least on par with a big Noctua NH D14, but it still wasn't enough.
https://www.overclock3d.net/reviews/cases_cooling/cooler_master_nepton_140xl_review/5

>>58864936
I see, I'll look into it.
I bought an unlocked CPU but never OC'd it (4790K).

>>58864975
It surprised me that the 6900K was hitting its TDP ceiling on pure FPU load, not even hitting the hottest, power consuming extensions to the core.
It surprised me because the 6950X has the same TDP rating, but higher clocks, so that 6900K must just be leaky as all hell.

If the 6 core overclocks well it could be the bargain of the century.

>>58865080
Theres the thing, no FinFET parts are leaky. There are more characteristics to getting stable transistor switching at a certain input voltage and frequency than leakage. The variability in the physical height and width of a bulk fin determine a big part of required voltage for a whole chip, thats a big part of clock binning, the consistency of the structures themselves.

>>58864960
1/10 got two sentences out of me. You don't have to back off after typing a retarded statement.

>>58865159
0/10, you're just trolling yourself now.

File: windows 10 1607 msdn.png (49KB, 711x415px) Image search: [Google]

49KB, 711x415px

>>58864386
Microsoft does actually have x86 only versions of Windows 10.

File: 1245094465859.png (23KB, 400x400px) Image search: [Google]

23KB, 400x400px

>>58862720
>LITERALLY DOUBLE A 7700K

>$500

AYYMDPOORFAGS CONFIRMED ON SUICIDE WATCH

>>58862819
980 was going for $400 wasn't it?
980 TDP was about 160w

So AMDrones were right?

>>58862720
Too fucking late to give a shit. Been hearing about fucking Zen for over a damned year now...no cpu to buy.

Fuck it. I got an fx 8320...I just don't have it in me to wait for the miracle chip that never fucking shows up.

Nobody cares anymore.

>>58865505
There's no "than". There were speculations like "RX 480 will be more powerful than GTX 480 for half the price".

>>58865526
> GTX 480
Damn, I don't see what I type. Whatever.
RX 480 is clearly more powerful than GTX 480, but I was talking about 980, of course.

>>58862720
>>LITERALLY DOUBLE A 7700K FOR THE SAME PRICE
W h e r e a r e t h e b e n c h m a r k s
Faggot

File: 1449948441928.png (12KB, 700x400px) Image search: [Google]

12KB, 700x400px

GO AMD GO

File: 6900k.png (72KB, 1452x530px) Image search: [Google]

72KB, 1452x530px

>>58862758
here are your benchmarks
about the same as 7700k in quadcore loads
50-70% faster in multicore.
much more overclocking headroom because half the power consumption per core and no shitty TIM

File: skeleton.jpg (27KB, 300x209px) Image search: [Google]

27KB, 300x209px

We wait for Zen....We wait...waiting.

Wait forever...wait alwAys...we Wait....

Dear God...so much time...never ending....waiting

We wwwwaaaaaaaiiiiiiitttttttt.

>>58865697
>20% slower in >95% of the applications
>60% faster in <5% of the applications

I will never understand why neo-/g/ touts 8c/16t Ryzens as a reasonable choice for a normal desktop "because it's the same as Broadwell-E for half the money!"
You don't tell people to buy 5960X or 6900K either, unless their use actually warrants it, which for 99.9% of /g/ posters it doesn't.

>>58865946
>single core benchmakrs
>relevant
nothing runs on just 1 core anymore. even videogames utilize at least 4 cores. quadcore benchmark is indicative of lightly threaded workloads. singlecore benchmark is pointless.

love u amd

ty for 64bit and apus<3

amd marketing, go home

This is RX 480 all over again.

>Tens of overhyping AMD shitposters invade /g/
>5 threads at any given time
>Product comes in
>Turns out to be anything between decent to utter shit
>No way to meet the expectations.
>Annoyed Intel/Nvidia fanboys that were bombarded with shilling for months start to latch at AMD fans who suffer cause of those few severely autistic shills.
>AMD fags suddenly pretend nothing happened and nobody actually expected AMD to perform as well as Intel or Nvidia.
>Week later they latch onto next AMD soon to be released product.

I wish we had an Nvidia fan autistic enough to screencap all those retarded marketing threads we had prior to release of 480 so we could end those threads the moment they pop up.

>>58862758
>if this is true then intel is truly finished
you'd hope so but dumb goyim are willing to pay the intel tax

>>58865540
Too late the damage is already done and your opinion marked invalid.

>>58864258
I read somewhere that Intel partnered with Nvidia on making the best house fire.

>>58864417
Even my completely stock i5 4670 with 84W TDP draws around 115W during heavy AVX loads.

>>58866269
Those that claimed the 480 will beat the Titan X for 100 DORRA!!!! were absolutely retarded and nothing official or leaked at all pointed to that being true.
With Zen on the other hand, we have solid evidence that it will actually deliver and be on par with Intel in most workloads.
The only remaining question right now is the clock frequency of the final 4 core chips and Zen's general overclockability.

>make x86 great again
>x86
>great
>again

>>58864553
Noticed

File: Shitwrecker.png (105KB, 255x345px) Image search: [Google]

105KB, 255x345px

>>58866519
You called?

File: 1485013521718.jpg (53KB, 512x384px) Image search: [Google]

53KB, 512x384px

>>58866269

File: 1467649778879.jpg (35KB, 600x338px) Image search: [Google]

35KB, 600x338px

>>58866269
Anon I...

File: 1485309897300.gif (2MB, 500x500px) Image search: [Google]

2MB, 500x500px

>>58866269

>>58862720
>$320
>for x86
>in 2017

>>58865246
I don't understand why, how big is the x86 market?
Did they do it because of Intel Atom?
How many PCs actually deploy Intel Atom?

It blows my mind that almost 2 decades later we still have x86 only Operative Systems.

>>58867604
Microsoft is just incompetent.
OS X for the longest time supported x86 and x86_64 in the same install, deciding which kernel to load at runtime. (press 64 during boot and it loads the 64bit, press 32 and it loads the 32bit)

>>58862720
>just one image
>no source
>still $300 for what /g/ spent months saying it would be $200

Fuck I have to start saving

>>58862720
>no Windows 7 support

They blew it. 40% of potential sales wiped off right there. Microsoft must have bribed them handsomely to reject those still on Windows 7.

>>58868194
literally nobody was saying the 8c was going to be 200$.
Reasonable people were hoping the 4c might go as low as 200$.

>>58868194
>still $300 for what /g/ spent months saying it would be $200
You haven't been here senpai. Every mouth breather in here has been spouting for months now that AMD can't sell the chip that costs them $50 to manufacture for anything less than $800 because intel sells their chip that costs them $50 to manufacture for $1100

>>58862720
Coffee Lake is more exciting desu

>>58869378
Coffee Lake will be playing catch-up. Just six cores, 1 year late.
The advantage will be that you can use it as LGA 1151 upgrade - unless Intel ties that with new chipsets/motherboards.

>>58869399
>unless Intel ties that with new chipsets/motherboards
>implying they will absolutely not

>>58869399
>>58869630
I'm inclined to agree here. I remember seeing an article claiming that intel was prepping something along the lines of a 7740 and 7640 that would help to compete with ryzen that would both be on lga-2066, which would imply that coffee lake would be too.
side note, the proposed i5-7640 is rumored to have hyperthreading enabled, if it actually exists, so that's interesting to think too. Maybe intel will actually bring hyper threading to i5s now that AMD is bringing it to lower end chips as well.

>1700x = i7-7700k
>Intel is $40 cheaper
>available now
>has 10x more research done on it so its almost guaranteed to be more reliable

Again why should I buy this piece of shit?

>>58869694
> more research done
What the fuck are you even talking about anon.

>>58869710

Y'know, the researches done by the scientician people in white coats.

>>58869718
Hopefully this is bait.

>>58869710
Sorry I meant Intel is more reliable, this isn't even disputable.

>>58869744
Provide evidence.

>>58869756
Dude Intel, come on bro. You're comparing a fucking Sony to a Vizio or Westinghouse. Also this is a brand new processor that is slightly slower than the Intel equivalent while costing $40 more MSRP. How do you idiots feel that somehow you are not getting fucked by AMD?

>>58869744
>Sorry I meant Intel is more reliable, this isn't even disputable.

Ah I remember this meme going all teh way back to the late 90s. It turned out it was anti-AMD propaganda that was parroted by stores stocking Intel pcs.

>>58869786
>slightly slower
>competing with the 7700k bigger, badder brother the 6900k
you fucking what

>>58869786
>>58869694
>double the cores
>double the threads
>very similar IPC (within 5-10%)
>7700K doesn't clock twice as high
7700K is actually just garbage now.

>>58869786
>slightly slower
Are you literally retarded?

File: GDyOS.png (224KB, 778x960px) Image search: [Google]

224KB, 778x960px

>moar cores
>faster
>witness actual bench
>its only slightly slower guise
>AMD even said the naming is based on competing with specific Intel processors

>>58869855
>>58869786 (You)

>>58869855
>damage control

>>58869756

Not that guy, but i've been working in datacenters for over a decade now. In my experience Intel is definitely more reliable, but again I say that is experience. Past AMD CPU's were also housefires and if Ryzen runs at good temperatures they'll probably be on par in terms of reliability.

I think one of the biggest barriers AMD will have with breaking into the server/infrastructure/"cloud" market is RMA. Intel has a badass RMA program for corporate clients. In some of the really large Google and Amazon datacenters they have Intel employees on-site to conduct troubleshooting and immediate replacement. I do not see AMD really doing this unless they start making good money.

PRAISE THEE, AMD, FOR I HAVE NO FORSAKEN YOU!

I HAVE WALKED THOUGH THE FX PURGATORY, FROM BULLDOZER ALL THE WAY UP TO VISHERA.

LO', I HAVE TRUDGED THOUGH THIS MISERY, THIS 8350. I HAVE DEALT WITH SUBPAR PERFORMANCE, BUT FEAR NOT FOR I HAVE NEVER WAVERED! MY LOVE FOR YOU IS TRUE AND COMPLETE!

ALL HAIL AMD! DOWN WITH THE DEVIL INTEL! THE CORPORATE OVERLORD WHO TRIED TO SNUFF YOU OUT; NO! WE WILL NEVER STOP!

ALL HAIL AMD! ALL HAIL AMD! ALL HAIL AMD!

>>58869694
>more reliable
You can almost smell the Intel marketing team sweating as they come up with this bullshit

>>58869744
>Sorry I meant Intel is more reliable
https://www.theregister.co.uk/2017/02/06/cisco_intel_decline_to_link_product_warning_to_faulty_chip/
https://www.theregister.co.uk/2017/02/07/intel_atom_failures_go_back_18_months/
http://www.anandtech.com/show/11110/semi-critical-intel-atom-c2000-flaw-discovered

>>58869255
>taking the bait
They know and you know people were just shit posting about ayymd.

does anyone know the difference in single core performance between broadwell and kaby? I want to know clock for clock, single core improvements
>inb4 theres no improvement
im not trying to start a fucking flame war, I'm just interested because i want a better comparison of ryzen to kaby, so hell some kind of review showing the improvements would be great

>>58869946
Check AnandTech's Skylake review, then add literally nothing to that. There you go.

>>58869255
>4 cores
>$200
Basically the same price as Intel which actually has a proven track record. Yea AMD is finished.

>>58862720
>2x the cores
>0.75x the performance
It's AMD, all right.

File: Intel IPC.jpg (259KB, 600x2221px) Image search: [Google]

259KB, 600x2221px

>>58869946

I just want AMD to do decent so intel prices come down

and maybe intel will actually start improving too

>>58869982
>proven track record
how long before you run out of marketing buzzwords?

>>58869694
>>58869744
>>58869786
>>58869855
>>58869982
>>58869990
((((((((((You))))))))))

>>58866115
Fucking bullshit you are trying to spread.

Most games don't profit much from 4 cores at all.

Increasing frequency does more than cores.

>>58869203
not even close, most sales will be prebuilts running Win10

File: 1467055276405.jpg (26KB, 251x242px) Image search: [Google]

26KB, 251x242px

I cannot for the reviews and read the comments of apologists tards here.

>>58865525
GDC is the last AMD event in Q1, expect to see it launched there
>inb4 /v/

>Zambezi
>its over Intel is finished
>Vishera
>its over Intel is finished
>Zacate
>its over Intel and Nvidia is finished
>Steamroller
>its over Intel is finished
>Excavator, piledriver, other Fischer Price construction toys
>I..its over Intel is finished
>Zen
>Uhh uhh I.. i.. its over guise Intel is finally finished this time
Said increasingly nervous chink cock gobblers.

>>58862720
First, I'd like to clarify that OP is a faggot. Ryzen chips still don't have fixed TDP, as 95 W has proved itself to be every unstable at 3.6 GHz 8c/16t. The info you have is for the review samples sent to select people, that have already proved themselves to be failed. And also , more reliable sources confirm higher MSRP of 700-800$ for the top chip. Stop falling for this low quality bait from AMD shill.

>>58870556
you got some links on the 95w claims things? I'm interested to see, as I thought that was a really interesting draw behind the chips

>>58869982
Please point me to a source where Intel sold 4c/8t chips for 200$ to the public.

Also:
>reliability
See
>>58869937

>>58870556
Please source all your claims.

>>58869691
LGA 2066 is socket for Skylake-X, the new HEDT replacing X99.
Basically Intel had the silly idea to repackage Kaby Lake so that it can be used in that platform too - despite it only havin g dualchannel DDR4 and PCIE 3.0 ×16, so half of stuff on LGA 2066 motherboards won't work with such chips.

Coffee Lake will likely still come into a separate mainstream socket - LGA 1151 or its like.

>>58869694
Buy i7-7700K, don't wait three weeks to see, be may guest.

But enjoy if your CPU will be 50-75% slower in MT apps and games, for the same price.

>>58865526
nobody cares faggot
amd just wins

>all those people wanting AMD to fail
Do you really want intel to continue assraping you on high prices for only tiny 5% improvements?

>>58866269
>latch at AMD fans

oh no, please don't latch at me

fucking retard

>>58870475
afraid, my child?

THANK YOU X HOW CAN X EVEN COMPETE X IS FINISHED AND BANKRUPTED


:-)

>>58866314
Not true. During Athlon 64 days, Intel only had the OEM industry to keep them afloat. The OEM industry is DEAD in 2017. PCs have become pretty much premium enthusiast grade products, and Intel was practically dead in that space when they were behind. Same as AMD is dead today.

ARM cpu's when??

>>58871067
/g/ in a nutshell

>>58871067
X IS GUNNA GIVIT TO YA!

>>58871196
about, 20years ago.

>>58870708
1151 is confirmed to last through canonlake, so they can introduce as many useless and retarded quarter generation re releases between now and then but it'll be supported there

I kinda wish I didn't make a new build for Skylake... I doubted ryzen. At least I can get canonlake i5 with hyperthreading in a couple years for cheap

Maybe I'll get a Vega 11 GPU then too

File: skeleton-in-cobweb-covered-chair.jpg (19KB, 311x423px) Image search: [Google]

19KB, 311x423px

The wait....the eternal wait for Zen...We wait always and forever. So much time has passed...

We wait so long...we never stop waiting...and Zen never comes...so we wait some more...

The endless wait...do we even remember what we wait for it why we wait? So much time passes...

We only know the endless waiting now...waiting is life

>>58871679
Really don't know what you're talking about, but your memes are stale

>>58871277
>At least I can get canonlake i5 with hyperthreading in a couple years for cheap
If not Cannonlake, at least a Coffee Lake 4c / 8t for cheap. Cannonlake will be on 10nm, so I doubt that it will be on LGA1151

>>58870062
Quite a few recent games double the minimum FPS when switching from an i3 to an i7 (approximately double theoretical processing power) provided the gfx card is powerful enough

>>58869982
He meant 4c / 8t

>>58869378
Depends on the 4c / 8t i7 price

>>58868194
Wrong, retard

>>58865525
Good god, you are even more retarded than the memers in this thread