AYYMD ABSOLUTELY BTFO HOW WILL THEY EVER RECOVER RYZEN OWNERS

>>60394022
>no THREADRIPPER codename

No thanks

File: threadripandtear.png (33KB, 942x537px) Image search: [Google]

33KB, 942x537px

MEGA-TASKING = Old and busted
THREAD-RIPPING = New hotness

>>60394022
Intel's naming scheme for the processors is so fucking confusing

I'm interested though, I want to see prices.

>>60394090
LEL AYYYMD COMPLETELY AND UTTERLY BTFO

I can't wait until this comes out and then AYYMD shills will do a complete 360 and suddenly say you don't need all the extra cores.

>>60394095

Those 10+ core i9's are going to be in the $2,000 range, I guarantee it.

>>60394090
Can we expect any increase in IPC with threadripper?

>>60394207
I don't think so, that comes with Zen 2.

File: 0b5b20a4462309f7c0a60717780e0cf3d6cad6ff.jpg (43KB, 580x966px) Image search: [Google]

43KB, 580x966px

>>60394022
Watch them be reviewed against 8 core Ryzens instead of HEDT ones

>no prices
inb4 cheapest i9 is $1,400 and the 4c/4t i7 is $750

>>60394223
Ryzen 9 doesn't even exist, you can't review them against an unreleased processor

>>60394223
That looks like BS bro.

>>60394090
>not even a year and they already changed the socket

AYYYYYYYYYYYYYYYYYYYYYYYYY

File: baito.png (38KB, 625x626px) Image search: [Google]

38KB, 625x626px

>>60394261

>>60394223
"threadripper" is computex, Skylake-X is E3

E3 comes after computex

lol Intel fell for the more cores gimmick.

>>60394275
Is that supposed to be a counter-argument...?

The chart literally says they're using a new socket. You idiots insult Intel for changing sockets for new CPUs every year but it's okay for AMD to change the socket not even 6 months after a new release?

>>60394176

Look it's already happening >>60394311

>>60394314
What are you talking about you dumb idiot? You can't physically fit 16 cores on AM4 socket, both pin wise and dimensions.

>>60394369
Delusional AMD fags think AMD is doing anything different, when they're still just pumping out more cores

>>60394359
>Intel releases a new processor that requires more connections to work
>JEWS! JEWS! JEWS! JEWS! They just want to wring every penny out of customers by making you change motherboards to upgrade!

>AMD releases a new processor that requires more connections to work
>M-m-muh AMD does nothing wrong! Look at the tech specs idiot! It's okay for AMD to do it! I support dedicated shitting streets for America!

>>60394194
Intel 12 core $2000
AMD 16 core $1200

File: wwe.jpg (111KB, 800x1000px) Image search: [Google]

111KB, 800x1000px

>>60394420

>>60394022
>7 times the price
>1% better
Ryzen BTFO

>>60394428
Nice try nigger, Skylake-X literally has 50 pins more than Broadwell-E, that's more than enough.

>>60394439
Yes because number of cores is how we measure how good a processor is

>>60394022
>112W quad-core CPU
AHAHAHAHAHHAAHAHAHA

File: 1479885942506.jpg (36KB, 400x399px) Image search: [Google]

36KB, 400x399px

>>60394022
7900 should have I think 10 cores right?
7920 should have 12 I believe

mine now is a 6900 and only has 8 cores.

Comparing intel HEDT processors to ryzen is apples and oranges, designed for different purposes. For home use or gaming I believe in 7th gen that Ryzen will begin to become a HUGE competitor,.

>>60394428
This is AMDs hedt.

Zen 2/3/4 etc will still work on an am4 board.

Can't say the same for delidlake.

File: cirnocore9.jpg (171KB, 620x720px) Image search: [Google]

171KB, 620x720px

>>60394022
finally

>>60394474
Kek. AMD already has the IPC and efficiency advantage.

>>60394458
Maybe if you post enough memes it will become true

>>60394474
We'll we know ryzen 8 core > Intel 6900 8 core

You do the math retard

>>60394494
Of course skylake wouldn't run on an AM4 board. Are you fucking stupid or something? They're from completely different companies

Why does AMD 32 core (server) and 16 core (HEDT) have the same pin count?

File: eca.png (207KB, 651x315px) Image search: [Google]

207KB, 651x315px

>>60394504

>>60394503
You mean AMD is both slower and more expensive

File: 1494714592109.png (2MB, 769x8779px) Image search: [Google]

2MB, 769x8779px

>>60394524
:^)

What would seriously need 12 cores?

Would that even benefit gaming? I often thought 8 cores was an overkill, but 12? Fuck me.

>>60394492
Keep in mind intel is running scared ATM... expect the BIGGEST push ever in intel marketing for the 7th gen - 8th gen trying to quell the upryen.... heh.

Its already happening behind closed doors, the breaking of the tick-tock cycle has already occurred due to ryzen. Intel now has fast tracked HEDT deployment for the first time ever out of step with every other generation including the rollout of the 7 series in general from a tech standpoint.

The new optane will make 7th gen reg processors into powerhouses and intel is banking on this heavily. Want to know more then fly to taiwan and then can talk...

File: intel-vs-ryzen-value.png (31KB, 780x371px) Image search: [Google]

31KB, 780x371px

>>60394536
:^)

>>60394022
ok
i'm still getting naples 32 core x2

>>60394524
>>60394536
Damn, I fell for the bait

>>60394514
Compatibility and it ensures future HEDT chips will have enough pins to work with, even the 7nm parts.

>>60394542
Rendering 8k resolution interracial 3D CGI realtime VR sex scenes with your mother or girlfriend scanned in as the woman being gangbanged by niggers

>>60394536
its not just about cores its about the number of threads and the ability to balance load across the CPU which is a art in itself my friend and AMD sadly has never mastered this but does not need to since their not the audience targeted for their CPU line which keeps the cost down while delivering solid performance which is a fact.

>>60394542
>gaming
You need to be over 18 to post on this board.

>>60394022
>new socket a year .... goyim in fear....
>i7 4C/4T
>112W TDP
AyyMD BTFO

File: quick fart.jpg (18KB, 179x164px) Image search: [Google]

18KB, 179x164px

>>60394545
>cost per FPS

>>60394572
>this is a fact
[citation needed]

>>60394545
https://archive.rebeccablacktech.com/g/search/image/p0-QGx1IgmDgyFomlgZTBg/

You really might want to try keeping this retarded chart that constantly gets laughed at to threads where it might have some shred of relevance (Protip: Nobody buys HEDT for muh gaymes)

Aye yo /g/ what do ya see?
Why not sit and listen to me?
Where were you when THREADRIPPER came to town?
Where were you when blue team shills all laid down?
Oh my life changed that day, oh it did in a wonderful way!
My downloads were pure, my code was clean and pumping the wife felt pretty keen!
The shills, no longer dumbfounded, quicky awoke!
With a skip and a pop I ended my poke!
Heavens my dear the jewery is near!
I rushed to /g/ and what did I see?
140 watts a total T-D-P!
I chuckled aloud and said quite proud,
Back to Isreal you jew for you know what's new?
THREADRIPPER is here and your end is now hear!

>>60394584
>i7 no longer have hyperthreading
>literally just a rebranded i5
>still costs the same price as a real i7 would have in previous generations

So... what's the point of buying Intel anymore? They seem pretty desperate to keep people from jumping ship to Ryzen

File: ppdRyzen.jpg (374KB, 1920x1200px) Image search: [Google]

374KB, 1920x1200px

>>60394545
lol.

>>60394624
>>60394545
What the fuck are these retard charts, kill yourself, both of you.

>>60394614
Do you want to tell me you don't play solely 10 year old games at 800x600? Casual

File: HAW HAW HAW.jpg (64KB, 640x640px) Image search: [Google]

64KB, 640x640px

>>60394544
>The new optane will make 7th gen reg processors into powerhouses

>>60394614
Their next mainstream CPUs will all be 6 cores (I think i3 might still be 2 cores, or maybe 3, definitely not 4)

So the i7-7740K is for fags who want to keep the single core performance dream alive for one more year.

File: muh shekels.webm (3MB, 853x480px) Image search: [Google]

3MB, 853x480px

>>60394022
>7640K
Why?
Those MB cost like $300, right? isn't this just a rebadged i5?

Also, not even a week ago the 'mah cores!' was a meme. I'd be nice if fan boys/fuccbois got their story in order.

File: ryzen-5-average-gaming.png (75KB, 600x865px) Image search: [Google]

75KB, 600x865px

>>60394624
>simulated cpu benchmarks

Meanwhile in the real world

>>60394545
>>60394624
>>60394669
Leave gayming out of this and quit shitting up the thread.

>>60394022
>i7 4C/4T

Why the fuck is there a i7 without hyperthreading?

>>60394654
So the i7 will be 6c/12t and the i5 will be 6c/6t? Unless there's a huge jump in IPC from Skylake I still don't see a point in buying that over a Ryzen 5 or a previous gen i7.

>>60394680
The chart I posted wasn't even a gayming benchmark.

>>60394599
>[citation needed]
AMD and its purpose is to compete not with people who render audio/video or do modeling or deep learning.

Its purpose and where it is good is and always will be running vidya and thats about it, not a bad thing but that hasnt changed and is what keeps their processors cheaper.

File: photoshop.png (28KB, 500x450px) Image search: [Google]

28KB, 500x450px

>>60394680
>he thinks it's just gayming

>>60394669
You realize this chart makes Intel look terrible, right?

Base clocks are pathetic.

>>60394720
AMD fags confirmed for illiterate

>>60394715
You realize this chart makes Adobe look terrible, right?

File: 1494627888896.png (1MB, 745x4227px) Image search: [Google]

1MB, 745x4227px

>>60394715
Don't you think it's time to get new image? See
>>60394536

>>60394715
>literally 0.1s difference
I'd rather enjoy my games NOT stuttering than to finish my epic meme exports 0.1s faster, thank you.

File: 1479885990618.jpg (20KB, 225x222px) Image search: [Google]

20KB, 225x222px

>>60394614
any intel processor prior to 6th gen is literally garbage. Only the new 7th gen intel chips per intel will take advantage of win10 features and be able to run optane.

Intel will not allow and has blocked any future functionality in all prior processors to help accommodate users to making the right decision by upgrading their computers to the 7th gen since its design is so radical that no other processor before it can do what the processors in the 7th gen do hence why Optane and other exclusive win10 features will only function on the 7th gen. ---===* The more you know :D

File: chrome.png (44KB, 650x350px) Image search: [Google]

44KB, 650x350px

>>60394741
Yeah really terrible

Sure are a lot of terrible developers out there

>>60394687
It's a rebranded i5. Don't ask me why. Intel is just throwing whatever crap at the wall to see what sticks.
>>60394690
Yeah, it doesn't seem like it's going to be that great. But they are at a dead end in terms of IPC and their process improvements aren't helping much. http://wccftech.com/intel-coffee-lake-desktop-6-core-4-core-cpu-leaked/

>>60394769
You realize this chart makes Javascript look terrible, right?

>>60394653
>>The new optane will make 7th gen reg processors into powerhouses
Optane basically will run alongside intels 7th gen processors exclusively from Intel like an AI learning the users habits and establishing fetch requests to preload data based upon them. Thats all it is but it will speed up everything a user does exponentially. Intel has singlehandedly revolutionized the way people will use and see computers.

What a time to be alive!

>>60394769
>0.04s difference this time
Now amplify by 50 and reverse Intel with AMD and you get frame-time differences in gayming

File: 1494462880472.png (2MB, 973x6115px) Image search: [Google]

2MB, 973x6115px

>>60394769
Kek. How many non gayming benchmarks have you posted? 2?

File: autocad.png (175KB, 725x548px) Image search: [Google]

175KB, 725x548px

>>60394792
Gotta love AMD fag rationalizations

>Ryzen is horrible at everything. It's the software developers fault!

>shitting up a thread with cherrypicked benchmarks posted a million times

File: ss.png (230KB, 1449x852px) Image search: [Google]

230KB, 1449x852px

>>60394819
Oh boy, 3 this time!

File: excel.png (27KB, 500x450px) Image search: [Google]

27KB, 500x450px

>>60394806
I can post these all day

Ryzen is garbage at anything that isn't rendering/encoding

File: muhsinglecore.png (454KB, 1920x1200px) Image search: [Google]

454KB, 1920x1200px

And here's what happens when you look at the single core performance right next to the multicore performance. Oops, suddenly Intel's huge single core advantage looks like a piddly little joke. lol

>>60394819
>i7-7700k doesn't even appear on the chart because it's so slow

Intel literally BTFO for all eternity. Why even buy above the i5 at this point?

>>60394835
Userbenchmark is shit

File: Excel.png (60KB, 1305x1434px) Image search: [Google]

60KB, 1305x1434px

>>60394838
lol.

>>60394838
Ryzen is also garbage at making your games stutter all the time... Unlike Intel, they are excellent at stuttering, mate.

File: 1494704824963.png (289KB, 596x628px) Image search: [Google]

289KB, 596x628px

>>60394838
>microsoft excel

File: af0ed353cd35fc23a4bb7053737d524dae5b78a22c077490d14cb664bcccefe1.jpg (104KB, 960x640px) Image search: [Google]

104KB, 960x640px

>>60394090

>>60394819
>i5-7600K demolishes an i7-6800K
This chart is beyond pants on head retarded

>>60394838
ryzen is fine for non windows based systems, it does suck at rendering but not everyone sits around rendering shit, do you?

>>60394838
>I can post these all day
We know. You pretty much have no life and have been in every one of these threads for the past three months posting the exact same cherry picked benchmarks non-stop.

>>60394838
Kek. What's his next benchmark, /g/ros?

>>60394924
He'll probably switch back to GAYMAN.

>>60394957
Won't work on this thread :^)

File: here_you_go.jpg (55KB, 1000x1000px) Image search: [Google]

55KB, 1000x1000px

>>60394022
>>60394483
>>60394502
>>60394524

>>60394965
Make a new one.

>>60394903
And he does it for free

>>60394886
Yeah that's what I noticed

>>60394924
It starts with gayman average results, then when he figures out a i7 is 10% faster than i5 and R5 for twice the price switches to one or two images of single threaded "productivity" microbenchmarks(my favorite is that one 0.50 second photoshop filter) that can be run on a pentium and make a 6900k look retarded, when you tell him that he inhales some coke to calm himself and either goes back to gayming benches or quits the thread for a while to shitpost when everyone's gone

>>60394886
wtf i hate intel now

>>60394924
https://warosu.org/g/thread/S60239115#p60248537

According to the board retard, nobody has any reason to own a MCC/HCC Xeon

>>60394545
again this pic, fuck off idiot.

>>60395091
His mom's credit card isn't fat enough to own a 16 core Xeon at decent clocks, of course "nobody needs it"

>>60394838
Oh shit, nigger! I thought you were going to do it all day!

>You now realize each memory channel needs 288 pins
>literally over 50% of the 4000 pin AMD socket is memory

Meanwhile PCIe lanes I think are 3 pins per lane, at 128 that's still a a decent amount.
Summing it up 2688 pins go just for PCIe and memory.

File: intel.png (179KB, 1002x935px) Image search: [Google]

179KB, 1002x935px

what are you, too poor to afford the best?

File: 2017-05-14 23_36_11-Meine Bestellungen.png (23KB, 682x162px) Image search: [Google]

23KB, 682x162px

>>60394022
i recover with price/performance ratio

>>60394022
>new socket
oh boy

>>60395239
>85€ X370 board

????

>>60394458
The funny thing is that Hyper thread cores alone are more powerful than most Bulldozer CPUs

>>60395288
only on amazon.de

there is a high chance of cancelled orders though.

but what if amd sealed a deal with asrock to sell them at cost so people have no choice but to get ryzen

File: AMDRyzen1800XIntel6950X1_0.png (29KB, 617x837px) Image search: [Google]

29KB, 617x837px

>>60394022
Too bad the I-9 7920X is slower than the existing i7-6950x, nevermind the other fucking abortions in that line-up. (See >>60394835) and the 1800x comes close to tying the 6950X already.

Oh, btw, the nature of the Zen architecture means that a 16c/32t Whitehaven should be capable of 3.6ghz base clock and 4.0ghz turbo, as well as a 4.0ghz all-core overclock. The i9-7920X is going to get eaten alive by it, for cheaper.

The second shoah has begun, shekelchasers. Time to get in the oven.

File: 1492587863785.jpg (395KB, 2400x2000px) Image search: [Google]

395KB, 2400x2000px

>>60394838

>>60395336
>Too bad the I-9 7920X is slower than the existing i7-6950x
friend I just got a 6900 could have gotten 6950 but let me say you dont understand processors or their use to make that comment my friend. Speed is irrelevant when your spreading load over more cores.

Useless to you, useful to us who do research and analytics for forecasting.

>>60395256
one socket a gen, keep the goyim at bay

socket change a year keeps the goyim in fear~

>>60395357
Oh, so the 16c/32t AMD HEDT CPU will necessarily be better than the 7920X then, since it has more cores and will be cheaper then?

>>60395343
>post month old prior agesa 1.0.0.4 benchmarks..

how about some 1600 benchmarks instead?

>>60395382
I'll bet my left nut It'll use less power too.

>>60394798
It does have some impressive numbers. Now if it can be used effectively that remains to be seen. I,'ve had spot set aside for some Optane since building in Feb. Just waiting to see how well they impliment it

>>60395382
>Oh, so the 16c/32t AMD HEDT CPU will necessarily be better than the 7920X then, since it has more cores and will be cheaper then?
not for deep learning, only seeing that possibly being useful in custom distros, useless for windows users. Their just made for different things if that makes sense, its about how the cores can share load without overloading. AMD is and never has been great at this, nor were they designed to be, there is a reason the intel ones cost more to do more. You simply get what you pay for. To the average user amd is fine, ive seen the stutter complaints in gaming and that is true due to the pipes available for processing data in combo with a gpu or some shit technical like that if anyone can explain that better.

>>60395437
You're making no sense.

>>60395431
>Now if it can be used effectively that remains to be seen.
It can be for deep learning, would be the same programs they work the same way from 6th to 7th in that aspect for sharing load across cores to divide up tasks.

>I,'ve had spot set aside for some Optane since building in Feb. Just waiting to see how well they impliment it
it will be great for a boost from average users, its not about speeding up the HDD, that is just a gimmick. Its more like a high tech AI virtual cache which may or may not be useful for you. It learns your habits much like a digital fingerprint online tracks how you type pauses etc.. then it preloads programs based off those behavioral routines so its useful to everyone but their marketing sucks saying its an HDD speedup fix, that is literally making what they did sound retarded since its basically a scaled down AI for you.

when we get our digital waifus?

I actually think these new Intel chips will be bretty good, but the trade offs are TDP, and cost. These chips are not going to be easy to cool at those leaked boost clocks so thats extra money for a good cooler. MBs traditionally cost more for Intel, but X399 will have quad channel along with more PCIe so they may end up even. For the 4c/8c chips though you can compare to Ryzen R7 and R5, the mobos for the intel chips in that range will cost more for sure. So yeah Intel will most likely cost more, with maybe +/- 10% perf of Ryzen/Threadripper. Good luck.

>>60395462
which part? the last part I get but you should understand the first part.

a new socket a year
keeps the goyim in fear

>>60395474
The entire post, you seem to lack fundamental hardware knowledge.
Also you sound drunk.

File: 1469845137381.jpg (1MB, 3840x2160px) Image search: [Google]

1MB, 3840x2160px

>>60395437
>Making a series of nonsensical, evidence free claims, vague enough to be untestable.
Go away.

>>60394715
i wish they would bench pen lag instead of this pointless metric

>>60394090
The marketing tards who came up with this junk should be shot.

>using one cpu for cpu tasks
>not spending more than a chinese purebreed for professional, not enthusiast or gaymer components

>>60394022
>i7-7540k
>4c on HEDT platform
>only has 16 lanes
>Dual channel

Why in the fuck is this a thing? I do not understand

>>60395210

No, that's why I bought AMD.

>>60395858
Ryzen was named by engineers, it wouldn't surprise me if Threadripper/Whitehaven came from them as well

>>60394242
Neither do the i9's

>>60395943
Zen was named by the CPU team, not Ryzen.
Threadripper sounds like some 12 year old from marketing came up with

>>60394514
There will be 16c server parts, the HEDT platform seems to be a server rebrand, like every HEDT platform has always been

>>60394769
>i3 performs exactly the same as i7

holy shit this makes intel look bad

>>60395960
>Thread ripper
>Bull dozer
>Pile driver
Hmmm...

>>60394022
>140W
>140W
>140W
>140W

>>60394022
Not even 4ghz.
AMD will be just fine.

>LGA2066
>last HEDT was LGA2011
>they want a new motherboard for 50 fucking pins
>implying they couldn't fucking fit Skylake-X in LGA2011

>>60395960
Zen, the arch, was named by the CPU team, possibly due to the project lead name (Suzanne Plummer)
But that's not the marketing name for the lineup, according to Papermaster they couldn't use Zen for reasons, so they ran a survey/competition within the division, and the engineers proposed and voted for Ryzen

>>60396079
That's not as bad as LGA 1150 to 1151.

>>60396079
Meanwhile SPr2 is 4091(why not 4096) and can fit practically anything HEDT coming out in the next 4 years.
>get 16 core next month
>upgrade to 20 core 7nm HEDT chip next year

>>60396142
I'd be euphoric if 7nm chips are out next year.

>>60396175
Well, IBM has already made 7nm test wafers and since GloFo (IIRC) licenses IBM's process nodes, they'll have access to it as well.

7nm by the end of next year is a very real possibility provided everything works out.

>>60396142
I don't care over 16 cores for the foreseeable future, I want them to keep up a 10%+ IPC cadence for a few years, that way they can drop clocks and double their perf/watt

>somewhere out there, someone has got a Zeppelin die running at 0.6V and it uses less power than your dick

>>60394483
>tfw my 2500k at 4.5ghz shows 117w usage when torture testing

>>60396338
The 7700K isn't that much worse.

What happened to Intel's efficiency plan, anons?

>>60396379
Yeah, but unlike the 7700k housefire my 2500k is soldered

>>60396400
Ayyy

>>60396379
There's no efficiency when your silicon is running at absurd clocks, efficiency is measured at saner clocks.

Aye yo /g/ come along an-a tell me what do ya see?
Why not have a sit n listen to me?
Where were you when THREADRIPPER came to town?
Where were you when blue team shills all laid down?
Oh my life changed that day, oh it did in a wonderful way!
My downloads were pure, my code was clean
And pumping the wife felt plenty more keen!
The shills, no longer dumbfounded, quicky awoke!
With a skip and a pop I ended my stroke!
Heavens o heavens my dear the jewery is near!
Come now love Team Red does not fear.
I rushed and waddled to /g/ and what did I see?
One hundred an forty watts a total T-D-P!
I chuckled aloud and said quite proud,
Back to Isreal you jew for you know what's new?
THREADRIPPER is here and your lead you blew

>>60396379
what happened to proper overclocking? compared the 7700k has no headroom

it goes from what? 4.2 to 5(if lucky)
where the 2500k goes from 3.3 to at least 4.5, 5.0 if lucky too

>>60396218
There is a theoretical limit of how fast an x86 processor can be with any architecture. Why are you expecting that 10% performance increase each year is possible?

A socket change next year keeps the goyim in fear.

>>60395529
wtf bro im not going to write you a paper on it...
lol where is all this coming from when you should be able to understand what I typed if not, then clearly this area of tech at least is not of interest to you, dont blame me :P

>>60396456
But you don't know Zen's theoretical limit, and AMD said they'll have it for 4 years before a new one comes so I'm assuming they'll pull all they can from it in 4 years.
As for next year, I'm actually expecting more than 10% IPC due to the current core being both smaller while the cache and IMC improve.
Anything after 2018 I don't know.

>>60396499
>No proof
How are we supposed to believe you, anon?

File: 1474598057623s.jpg (1KB, 125x112px) Image search: [Google]

1KB, 125x112px

>>60394022
>i9
No way. They actually did it. The madmen!

>>60396581
>No proof
>How are we supposed to believe you, anon?
Proof of what are you asking? Proof that Intel has better shared workload capability across cores on a CPU than amd? This is widely known, were not talking some video game benchmark here.... This isnt new and has been widely known but it doesnt matter since noone here does that. Ive never met anyone ever who uses the computational power I do outside of maybe one programmer who does some insane compiling.

Its not that you need proof which is all over and discussed, its more that its irrelevant so outside your periphery for how you compute. For average end users AMD is fine. Its so fine that Intel is deploying earlier than ever this new line in the 7th gen as stated and why optane is being restricted to the same gen.

If want know more just email intel, their quite happy to discuss this.

>>60396501
Damn dude, do you know how much of a 15% IPC increase would mean on a 16 core chip?
That's enormous.

>>60396641
Son, you're barely coherent.

>>60394022
140W tdp ok intel.

>>60396581
No proof? Nevermind proof, he hasn't even bothered trying to articulate ANY sort of rationale that would logically explain why a 6900X would be better than a 6950X and a 16c/32t Zen HEDT CPU for any task whatsoever.

A series of buzzwords about deep learning and how cores "share load" and "you get what you pay for" isn't even an argument. It's literally a non-falsifiable series of buzzwords.

>>60396641
>This is widely known
Meaning you just made it up.

Which is fine since "shared workload capability across cores" doesn't actually mean anything, anyway.

Ah. I kept saying I was pretty certain they'd have 12 core, and everyone kept telling me it was only 10 based on currytech leaks.
But I doubt that 12 core is going to be those clocks. There is no way 4.5GHz all cores can be 140w TDP even with Intel's usual lie about TDP. That would pull 275+.
It's also messed up how they cut down the L3 cache when the i9-7800X up to i9-1920X

And worst of all, it's clear that Whitehaven will poop on these. AMD's SMT is so much better than HT. 16 Ryzen cores are like 19 Intel cores when it comes to multithreading. A 12 core Intel even if it was 4.3GHz all core wouldn't compete with 16 at 3.75 in anything threaded well which is EVERYTHING these CPUs are bought for.

>>60394061
>No WHITEhaven

>>60396696
Maybe he wants to say something about load balancing but is too English-retarded to know the word?

The OS does load balancing anyway, and if he's talking about interrupt handling while all cores are maxed out, AMD seems to do that better too as you can actually use your system while you're running all cores full tilt, Intel's stutters completely.

>>60396696
>>60396664
>Which is fine since "shared workload capability across cores" doesn't actually mean anything, anyway.
Im not the tech person, I am an end user, please consider this and try to straighten out what I say so it makes sense if anyone can technically.

Shared workload means running more tasks that can be divided up and run at the same time which means running more jobs and having them evenly spread out. Intel is capable of this and designed for this. the 6th line was a shit line and fucked everyone due to it literally removing abilities to function better in some ways than previous generations since the whole line was revamped to meet wider application and not meet the needs of all users as previous gens did. 7th is a hybrid mix which gives regular users the ability to use optane to raise their efficiency and HEDT users will have yet more cores but basically the same processing power as 6th if that makes sense.

>>60396786
Go home, you're drunk.

Jesus Christ it's almost like I'm listening to some kid from youtube.

>>60396722

They are only claiming 4.5 Ghz with TB3, which I believe is supposed to boost a single core for single threaded applications. I'm guessing it's going to be around 3.7 for all 12 cores.

What's the point of Kaby Lake-X quadcores if Coffee Lake will have six cores and most Broadwell-E users will just go for the 7800X?

>>60396808
im a regular user.... I just have very high tech equipment. I am trying to explain from what I know why we use it. Your asking me the same way you would ask a carpenter to explain how the drill he uses down to the engineering level of design and voltage capacities.

Sorry not a tech guru but we do spend millions researching what to use and so do many others there are tons of open discussions on google in forums around the world discussing these results. Thats how we describe them as end users.

>>60394428
Kys /v/tard

>>60396859

Probably for the same reasons they made the 3820 and the 4820. It's a very niche product, but apparently some people want more PCIe lanes.

Either that, or to further cement their single threaded performance lead.

>>60396696
>Which is fine since "shared workload capability across cores" doesn't actually mean anything, anyway.
The fuck?
Yes it does.
It means I can run a simulation and compile at the same time, ALL WHILE my IDE works perfectly.

AMD's superior SMT also helps a lot in this. With a 6900k if you have something rendering in the background, using all 16 threads obviously, and then try to use photoshop at the same time... Photoshop fucking lags like shit.
Do the same thing on Ryzen? Something about its SMT is smarter to give better context to photoshop and let it run smoothly. There is more delay in context switching with HT.
It's probably the same reason why background processes cause games to dip and stutter on Intel CPUs when they stay smooth with Ryzen.

Intel's architecture just sucks no matter how many bibahurz they have or how much they try to keep up with coars.

>>60396864
>I just have very high tech equipment
>but we do spend millions researching what to use
:^)

>>60396912
just google deep learning thats basically why we use HEDT processors and why they exist, cloud computing is not legal to use or allowed for many types of research due to privatization laws and proprietary safeguards.

File: 515[1].gif (1022KB, 640x480px) Image search: [Google]

1022KB, 640x480px

>4c/4t 112W
>140W
>i7 without HT
>i9

>>60394494
well, i mean you kinda can

my LGA1151 board currently running skylake can/will run kaby lake, coffee lake, and potentially even cannonlake

>>60396909
See, the difference between you and the guy I was replying to is that you were able to articulate a definition for what you understand the phrase "shared workload capability across cores to mean", and then articulate a rationale for why you believed certain processors are better at it (AMD's SMT being better than Intel's HT and not stuttering when asked to do additional tasks at full load).

The guy I was replying to was trying to simultaneously claim that a 6900X was better at whatever he meant by "shared workload capability across cores" than a 6950X and an 1800x and a 16c/32t Whiteraven. As I'm sure you will quickly realize, that makes no sense whatsoever.

>>60396984
Too bad those are all the same core.

>>60396909
In simple terms, AMD handles interrupts better, it was already mentioned some posts up.

>>60396966
Okay, but why the hell would a 6900X be BETTER than a 6950X at "deep learning", and why would either of them be better than a 16c/32t AMD HEDT CPU?

>>60394022
>4c/4t
>112 watts
christ, somebody call the delidding brigade

>>60396991
I think it means it's better at farcry primal before the patch broke it.

>>60396905
But they only offer 16+4on chipset lanes, it's the same as the normal desktop boards.

In a similar vein the 7800 and 7820 only having 28lanes puts them at a severe disadvantage locking the full pcie capability behind what is expected to be $1000 tier processors. This lane shit heap is probably the weakest aspect of the x299 platform, the powerdraw and performance are merely potentially not quite as good.

If whiteraven comes out with it's cheapest cpu with all 64 (52?) lanes enabled it's going to be a big hit. There's probably a market for an 8core (2cores per ccx) threadripper and all pcie lanes.

We still don't know final clocks for AMD's HEDT, lets assume they're gonna stick with 180W

Then there should be no problem getting a 3.5-3.6 base clock 16 core.
Especially when these HEDT chips might be using B2 silicon(confirmed) and lesser leakage dies(unconfirmed)

>>60396966
>deep learning
>with CPUs
>not GPUs
wut

>>60396991
>The guy I was replying to was trying to simultaneously claim that a 6900X was better at whatever he meant by "shared workload capability across cores" than a 6950X
Never claimed that, your confusing the part where I said I just got a 6900 instead of a 6950. but that isnt because the 6950 is better, its more that at this point there is no justificating for allocating the difference in resources economically when were all sitting on ass pissed and waiting for the 7th gen because all the previous gens have capabilities locked out to us to use that are unlocked in the 7th gen this is a first ever for intel, its like how win10 has all this stupid exclusive shit that could easily exist on win7 like dx12 but doesnt because fuck you thats why is what M$ says and now intel is on that bandwagon beating their drum to get consumers to buy the 7th gen due to their fear of ryzen.

>>60397191
Okay, fine, but none of that explains why you would think a 6900k would be better at anything than a Ryzen 1800x, much less why it would be better than a 16c/32t AMD HEDT CPU?

>>60397162
It's probably going to be around 3.6 all-core and 4.1 for 4 cores if the others are idle.

I'm certain these leaks aren't accurate.
3.6GHz max single core boost, with the all-core probably lower, would only be like 120-140w TDP.
They're probably fake leaks from AMD to throw Intel off.
Or AMD gives fake information to various departments to find leakers.

>>60397232
>Okay, fine, but none of that explains why you would think a 6900k would be better at anything than a Ryzen 1800x
Hmmmm I only said it was better for what I do not everything, keep in mind I did say AMD has a good reputation known for working with full functionality in gnu/linux based systems where intel can sometimes be crippled a bit. For deep learning in a windows OS environment Intel has been leader of the pack which is all I meant and is widely known. better is subjective but there are pros and cons to each, just like with vehicles a truck and car are good for different things and worse at others.

>>60397162
There was an ES with 3.1, things are looking good, the first ES leaks we had were at like 2.9 or something.

Besides that, base clock is just base clock, there's no stopping AMD from making a 3.1/4.0 that turbos all cores to 3.6 and a 3.6/4.0 CPU that doesn't turbo all cores.
There's no actual difference.

>>60397270
Okay but WHY do you believe that Intel is better for deep learning in a Windows OS environment? What technical advantage does a 6900k have over the AMD alternatives for this applicaiton?

I'm not even disagreeing with you here, at least not yet anyway, I'm still trying to figure out WHY you think this.

>>60397276
Besides that the latter one looks better (bigger numbers) to normalfags and braindead middle managers

Has no one actually considered for a moment to think that the HEDT chips will have slightly higher IPC and ST than mainstream Ryzens due to higher bandwidth IF, 4 channel memory and higher TDP leading to higher XFR/single core turbo ranges?
The 100GB/s Infinity fabric should help the most here in certain software that likes to bounce the main thread around.

Yeah I know nobody gives a shit about ST on 16 cores, but it's rather interesting to think about.

>>60397253
The leaks were from ES no?

Anyway clocks similar to 1700x are likely and in line with the previous rumour of them trying to get it to match desktop speeds. So base 3.4 turbo 3.8 and xfr can do what ever the fuck it wants, probably 4.1 as you said.

>>60397303
>Okay but WHY do you believe that Intel is better for deep learning in a Windows OS environment? What technical advantage does a 6900k have over the AMD alternatives for this applicaiton?
From what I am told and from discussing it online with others who research this same thing it has to do apparently with balancing that load evenly across cores with separate divided tasks when running many jobs. from what I know and this is all over the internet so feel free to look up these discussions to get the tech facts behind it, but from what I know intel has always specialized in this and spends the money to do it. AMD saves money since they dont spend much if any on this so they dont care and I dont think their users care since they dont run those routines or modeling like i said.

were talking one very specific use here where one outdoes the other, it is in no way saying AMD or Intel is best at everything just better at some things and as previously stated from my same work people say AMD is better for distros all of this is available in millions of topics online to learn more if you care, I am in no way tech literate enough to explain it.

Intel thinks that by rebranding an i7 and then selling it for more will somehow play catchup.

Their inner jew is coming full forth revealing that they're immediate reaction to AMD is to make their products more expensive.

>>60397387
ES was 3.1/3.7 with unknown all core turbo.

Mind you, this is still an ES, not even a QS, god knows at what stage of development.
I think 3.5 is very possible, especially at 180W
If it was at 155W like some leaks suggest, there's no saving Intel's datacenter from Naples.

>>60397416
>it has to do apparently with balancing that load evenly across cores with separate divided tasks when running many jobs
Okay, what do you mean by that? Because I cannot come up with any interpretation of this phrase where AMD's Zen architecture wouldn't be better at it in virtually every case.

The 1800x beats the 6900k at virtually every single highly multi-threaded workload, precisely because AMD's Zen architecture has a better implementation of SMT (simultaneous multi-threading) than Intel's Hyper Threading. As a result, an 1800x should be better in the exact scenario you seem to be describing.

This is why I am confused.

>>60397387
I'd expect the same few-core boost clocks as the 1600X/1800X, but possibly lower base and all-core boost clocks than 3.4.
You put 2 1800X under one IHS and 180w TDP is reasonable. There's absolutely no reason they can't or wouldn't make and sell that. There's no reason they'd make the 16 core clocks that low when there is definitely the demand for a 180w TDP HEDT CPU if the performance justifies the power usage.

Also remember, the first leaks of Ryzen7 and Ryzen5 clocks were also off. The very first ones especially, which were 3.15GHz of the 8 core.
It's sandbagging. AMD likes sandbagging with leaks now days.

>>60397431
Snowy Owl is going to be <100w TDP 16 cores.
2.2GHz base, 3.4GHz boost(not all-core, obviously), I believe.

>>60397424
>rebranding an i7
More like a Xeon

>>60394598
That's almost how I look at it and it's a pretty good way of looking at it.

I just look at price/performance or to put it another way, what CPU gives the best bang for the buck.

I don't play games so I don't care about that but cost/performance is a good way to look at it.

>>60397386
If anything Raven Ridge will have the highest IPC due to one CCX design, you can see this in current Ryzens by running 4+0 core configuration, there's a 3-10% IPC boost.
The 10% is in rare cases like RoTR, it's generally around 4-5%

>>60394261
>>not even a year and they already changed the socket
a socket change a year keeps the goyim in fear

>>60394572
hahaha what a load of hogwash.

make -j8 will distribute the compilation jobs across 8 threads. It creates 8 threads on any Intel CPU. It creates 8 threads on any AMD CPU.

ffmpeg will automatically detect the number of CPU cores and CPU features and distribute encoder jobs accordingly - using the CPU instructions that are available.

Load balancing across CPU cores is done by the OS and software that runs on it.

I know the american shills working for that low-wage marketing corporation Intel has a contract with are stupid, but please try to make a minimal effort. I mean, they are paying you. The least you can do is to learn the basics of how a computer works.

>>60397538
>The 1800x beats the 6900k at virtually every single highly multi-threaded workload, precisely because AMD's Zen architecture has a better implementation of SMT (simultaneous multi-threading) than Intel's Hyper Threading. As a result, an 1800x should be better in the exact scenario you seem to be describing.
Hmmm well we do pay people to research these things since we spend millions on this stuff and I get free computers due to the work so they dont give them out until they know what is right that we can request...

that being said maybe those are tests of some kind but not actual jobs with results being completed might be why? I dont think unless you can find some official AMD source that they ever wanted to or said they specialized in this have they? Im sure they wont say their bad it when asked but I dont think its been their focus to get us niche deep learning users. were the only ones who use that capability except like I said that one compiler guy I know but again he is niche like my use :P

You are right that ZEN has intel running scared with their architecture. this is new which I talked about. Intel is deploying optane and locking down 7th gen specific features in a bid to compete with zen. so while I dont disagree with you it may also be that if AMD is doing right and they are focused on this niche its the first time which is why its not being widely adopted yet like anything else thats a first.

Intel is violating their own tick tock cycle of CPU release their so scared of Zen with those lockdown features mentioned. I have an aunt who works at intel used to be a VP at hewlett packard I think was the name, shes old and does paper shuffling of some kind there but I can email her questions if you have them? she always sends me answers but during the week when she is there. Very upset this gen at Intel and it fucked apple too in the 6th gen just so you know hence that whole armature debacle that they cant figure out there at apple.

>>60397676
thats great I dont do encoding.... your comparing different things to different things. noone buys those overpriced intel CPUS to friggin render or encode unless their like professionals not the average end user.

Okay guys, stop replying to him, he's obviously in his own world.
He should come discuss this stuff when he can make heads and tails of the basics of OS and hardware.
Preferably with intermediate knowledge of English too.

>>60397720
tard I speak english and have been, its YOU who doesnt understand since you dont live in that world so you are right about the own world thing, using jargon and all that. it is what it is.

File: single socket holocaust.jpg (28KB, 550x550px) Image search: [Google]

28KB, 550x550px

>>60394022
RETURN OF THE KING.

>>60397778
Babelfish is more coherent than you.

>>60397819
I think the lingo used to explain it is the confusing part sadly.

>>60397431
Eh, that's fucking amazing for an ES.

That turbo practically guarantees a minimum of 3.4 all core, and you said it's not even a quality sample yet?
I wouldn't actually be surprised if the end chip will have 3.3/4.0 with a ACT of 3.6.
The base clock is left low for AVX* workloads, base clock is always minimum guaranteed clockspeed.

>>60394428
>1156
>1155
>1150
>1151
all those are Intel CONSUMER grade and not HDET kys jewtell shill

>>60397983
Speaking of turbo, the mobile 15W parts will probably have crazy turbos like 2.2/3.8
The lower TDP you go the crazier turbo and base clocks differ.

File: 1490751939451.jpg (168KB, 900x1200px) Image search: [Google]

168KB, 900x1200px

>>60394022
Coming soon to an outlet near you.

>>60397687
>Hmmm well we do pay people to research these things since we spend millions on this stuff
Those people just order Intel chips because "Nobody gets fired for buying Intel" aka ass covering.

>>60398113
Underclocked (for efficiency) processors have more oc headroom. What's crazy about that?