intel x299 vs amd x399. 1950x is 52% faster than the housefi

>>61654206
Jewripper looking good.

>>61654206
post source sauce of pic

>>61654257
curryniggertech


can't you see the logo, fag?

>>61654257
http://www.anandtech.com/show/11678/amd-threadripper-1950x-1920x-august-10th-new-1900x-549-usd-august-31

>>61654206
>5.2ghz
thats not possible tho

>>61654206
>5.2GHz
Are we playing with windows scheduler again?

>>61654287
Under ln2 it is :^)

>>61654323
Noticed that.

>>61654323
Bose-Einstien condensate benchmarks when?

Will threadripper make large custom water loops, chilled water or phase change useful again?

File: 7ytjxYa.jpg (27KB, 306x240px) Image search: [Google]

27KB, 306x240px

>>61654206

File: 1497689967244.jpg (123KB, 638x599px) Image search: [Google]

123KB, 638x599px

>>61654206
Intel si truly finished.

>Threadripper 1950x oc to 5.2g
Threadripper is released in the south pole before the US?

>>61654206
Fuck yes 5ghz here we come 1950x

File: 7c6.png (202KB, 600x700px) Image search: [Google]

202KB, 600x700px

>5.2 GHz
How is that even possible on 16 cores?
Did they disable most cores?

>>61655653
LN2.

File: 1497883390389.jpg (18KB, 267x297px) Image search: [Google]

18KB, 267x297px

>>61654206
SIR PLEASE
DELET THIS RIGHT NOW

File: 1487445614546.png (225KB, 500x358px) Image search: [Google]

225KB, 500x358px

>5.2

File: cinebrunch tr.jpg (76KB, 682x776px) Image search: [Google]

76KB, 682x776px

>>61654206
Even at stock it's 25% faster, that's not even the numbers AMD was showing before (3046)

I imagine with some hardware optimization it could hit the numbers AMD had showed previously.

>>61654206
FUCK AMD
FUCK LISA SU
FUCK JIM KELLER
FUCK RYZEN
FUCK INFINITY FABRIC

>>61655653
No, all cores were going at 5.2GHz. Also liquid nitrogen.
It's only intel's garbage bingbus design that requires cores to be OC'd independently.

>>61654282
>The number given was 4122, representing a 5.2 GHz overclocked (under liquid nitrogen, so not a daily OC)

>>61654206
DELID THIS

File: kiko.png (535KB, 500x824px) Image search: [Google]

535KB, 500x824px

>>61659039
"AND FUCK YOU!"

>>61654206
>that stock score
2133mhz memory?

>>61659114
They math inferred it like retards, not realizing AMD has already shown ~3,050 at stock clocks.

>>61659114
Cinebench is not memory bound.

>>61659172
>Cinebench is not memory bound.
What is infinity fabric.

File: 1b70c8cd-73d1-4cab-bc44-903e46d0351a.png (105KB, 255x345px) Image search: [Google]

105KB, 255x345px

Say my name.

>>61659172
Cinema is not even used by anyone on this board

>>61654287
Remember that the other dipshit reviewers thought it was fair to put the 7900x under a 1k water chiller hooked into a custom loop and show the results as if they were attainable. Not to mention the "standard" of a 5ghz OC on the 7700k that every other benchmark wants to show when you realistically have to delid to get there. It's definitely not the same as running under LN2 but there's still been some very fucky things in reviews lately.

>>61659220
Intel has sold a lot of CPUs on the backs of lies based on extreme edge use cases and golden samples.

File: 1500535353742.jpg (63KB, 592x533px) Image search: [Google]

63KB, 592x533px

>>61654206
>AMD 12 & 16 cores vs 10 core Intel
sure is getting desperate in here

>>61659242
It's easy to see them as competing products when they're the same price, but TR mobo is going to be cheaper too.

File: xeon gold.jpg (349KB, 1060x1047px) Image search: [Google]

349KB, 1060x1047px

>>61659242
Here's your $2000 18 core. Now go fuck yourself.

>>61659272
And in the next post, you will force an clock for clock IPC comparison between Zen and a i7. Hypocrisy at its finest.

File: hunny.png (342KB, 400x400px) Image search: [Google]

342KB, 400x400px

>>61659290
Ohh yeah, Chinese leaks are sooo reliable.

>>61659290
kek, now give it LN2 cooling like the 5.2Ghz 1950x this thread is about.

With LN2 and an unlocked multiplier it could easily hit at least ~4.5Ghz+ and most likely 5ghz+ since the 10 core 7900x can hit ~5.7GHz on LN2.

File: 1501135514673.jpg (38KB, 346x346px) Image search: [Google]

38KB, 346x346px

>>61659297
>you will force an clock

>>61659211
Pussy destroyer!

>>61659211
JEW KILLER

>>61659313
[visible autism]

>>61659310
You realize the non-LN2 results are right there on the chart as well?

>>61659313
>he was an clock

>>61659369
So fucking what?
~3000CB 18 core at 2.7GHz vs ~3000CB 16 core at 3.8Ghz

Wooo, such an AMAZING win from AMD. Seriously, if you give the intel free reign with LN2 it would obviously have more headroom than the AMD chip.

This isn't even to mention, the ACTUAL customers for EPYC and E7 xeons are going to care more about application specific performance, not some generic benchmarks.

File: 1478689619444.jpg (24KB, 174x164px) Image search: [Google]

24KB, 174x164px

>>61659415
>Literally who cares about paying more for less performance.
What about ACTUAL customers, anon?

>>61659469
Lolololol, yeah well lets just keep waiting and see what the data centers and universities etc, decide to go with over the next year.

Sorry, but intel's industry partners are simply too strong to be overcome in a single generation.

If AMD can keep up the pressure for another 2-3 years, yeah we might see a real shift in the market, but until then you're just fantasizing.

File: Intel-CoreX-Specs.jpg (192KB, 1200x729px) Image search: [Google]

192KB, 1200x729px

>>61659415
That is the base clock, the 18 core i9 will be even slower. Prob. will only boost to 2.9 GHz.

Also
>$2000
>2 more cores
You are an unbelievably stupid Intel shill. Like literally on another planet of stupid.

>>61659242
Ok, so here we have a perfect example of an Intel shill.

1) 12C vs 10C = .8333333333 X as many cores
2) 2431 * .8333333333 = 2025.83
3) 4.3 Max Boost (i9 7900X) vs 4.0 (1920X) = .9302325581
4) 2025.83 / .9302325581 = 2177.770833425
5) 2177.770833425 / 2186 = .9962355

TR is within 1% IPC of i9 based on rough calculation assuming that turbo boost wasn't disabled. (unknowable based on graphic)

Intel is getting fucked so hard, especially considering the fact that you are paying ~25% less for more processor.

>>61659492
>intel's industry partners
Jumped onto AMD's train.

>>61659505
yeah, some have, and they don't really have much to show for it since they've barely had more than 6-10 months to work with AMD.

Again, Intel has a decade+ of this with most of their partners.

>>61659499
When you're an actual customer you aren't paying those prices anyway.

Most larger customers wouldnt even pay half of what intel says an E7 xeon should cost.

>>61659415
3062 @3.4ghz. Fixed.

>>61659309
This is uncontrollable damage control!

>>61659505
And Intel just added a while quarter of AMD sales to their yearly forecast. They also just shipped 500k xeons to partners. They aren't scared.

File: 2609288_3b5d4d892302.jpg (990KB, 1622x1667px) Image search: [Google]

990KB, 1622x1667px

I tried so Goddamn hard not to get hyped
I have no self control

>>61659534
We're talking HEDT anyway.

File: 39CPB0.png (625KB, 648x720px) Image search: [Google]

625KB, 648x720px

>>61659492
This is for HEDT, the "DATA CENTERS" you speak of are looking at EPYC vs Xeon and that's an even worse prospect for Intel. At best, AMD only is used for hosting the coprocessors. At worst, it's the whole fucking stack minus some special use cases like large memory sets or massive database work.

>>61659492

Nigger, AMD literally reclaimed 10% of the CPU market in a single fucking quarter, and the price/performance ratio in the server market is even better than the consumer market.

>>61659547
assuming it had the cooling for that, turbo speeds are only supported if you give it enough cooling.

Give it LN2 like this 1950x and an unlocked multiplier and see what happens

File: 673.jpg (35KB, 660x422px) Image search: [Google]

35KB, 660x422px

>>61659492
>>61659526
>actually believing Intel's marketing

File: 1500636486941.jpg (53KB, 702x786px) Image search: [Google]

53KB, 702x786px

CORE I9 IS DOA!
O
R
E

I
9


I
S

D
O
A
!

File: 2017-07-31 11_31_07-Steam Hardware & Software Survey.png (15KB, 876x76px) Image search: [Google]

15KB, 876x76px

>>61659586
>AMD literally reclaimed 10% of the CPU market in a single fucking quarter
among who?
Among gaymers they've lost market share.

File: intel fans enjoying skylake-x.jpg (263KB, 1600x851px) Image search: [Google]

263KB, 1600x851px

>>61659595
The LN2 boils away faster than you can pour it.

File: GntGYw.png (38KB, 499x338px) Image search: [Google]

38KB, 499x338px

>>61659620
>Steam survey is now a better indicator than Intel's own earnings reports.

File: steam intel craptops.png (71KB, 1068x389px) Image search: [Google]

71KB, 1068x389px

>>61659620
Look at the clockspeeds reported. Notice anything? That's right, the vast majority are in "crappy laptop clockspeed" territory.

>>61659620
>Using Steam data
I sincerely hope you're trolling, otherwise you have mental issues

>>61659620
Passmark data.
>inb4 source is attacked by intel shills

File: 2017-07-31 11_36_28-PassMark CPU Benchmarks - AMD vs Intel Market Share.png (40KB, 638x425px) Image search: [Google]

40KB, 638x425px

>>61659640
intels earnings report earlier this week showed 14.8B in revenue and they don't even specifically talk about their market share vs AMD in the consumer space

>>61659211
CERTIFIED

>>61659679
passmark shows ~5% turn around for AMD, great for them, but it's nowhere near 10% kek.

>>61659242
>Compare 1k cpu with 1k cpu
>Waah muh cores
Lol

File: 1281470360227.png (15KB, 730x413px) Image search: [Google]

15KB, 730x413px

>>61659643
Good point. Come Raven Ridge, those numbers may well change.

File: AMD-Intel-Q3-2017-CPU-Market-Share.png (49KB, 749x597px) Image search: [Google]

49KB, 749x597px

>>61659692
Last graph I saw was pic related. I can't see a date on yours, so I don't know if it's before or after. It's entirely possible that a large spike caused a statistical anomaly in the data and it's settled again.

>>61659211

JIM "THE KIKE KILLER" KELLER

For the second time in history AMD processor architecture designed by Jim Keller brought the mega corporation Intel on to its knees!

Just watch, after Zen2 has been released and AMD needs a new architecture again, they are going to call Keller again.

>>61659766
Come Raven Ridge those numbers will definitely change. All of the major OEM's are signing up to lick AMD's asshole. Intel won't be able to use MCP again, regulators are watching this time.

>>61659822
just look at my screenshot, it's timestamped for about 30 seconds before i posted.

>>61659822

>didn't see filename

Looks like it was a spike, but 5% is still monumental. 10% would have been unprecedented.

File: 2017-07-31 11_47_17-PassMark CPU Benchmarks - AMD vs Intel Market Share.png (44KB, 653x511px) Image search: [Google]

44KB, 653x511px

>>61659822

>>61659685
>7% drop in desktop.
>What could have happened?

>>61659964
>calling 5.2% 7%
wew lad

File: 1496276668512.jpg (22KB, 533x477px) Image search: [Google]

22KB, 533x477px

>>61654269
Learn to post source sauce linkage in your OP fekkit

>>61654282
>http://www.anandtech.com/show/11678/amd-threadripper-1950x-1920x-august-10th-new-1900x-549-usd-august-31
>liquid nitrogen overclock
D R O P P E D
U
M
B
A
S
S A G E

>>61659574
> E-atx
> Fan on portion of VRM
> what is that connector beneath the bottom pci-e slot that's facing out and has a white strip on it.
> On-off switches
> Dipswitch
> tall ass M.2 slot riser card directly in the way of airflow to the ram/CPU from the front of the case

Wtf are they doing that the other mobo manufacturers aren't?

---
Everyone else has :
> ATX
> No fan on vrm
> no on-off switches like this is the 90s
> no dipswitches
> M.2 slots on mobo

This board looks like its all over the place.
They have 16/16/8/8/4/4 - 56 PCI-E lanes
But they have 2 M.2 slots off the riser = x4/x4 ? 64
But the chipset takes x4

And this is the most expensive board. Someone please explain
Wtf gives.

>>61660009

Shill harder. On air it still beats a $4000 intel part.

>>61659242
1800x released amd says its a direct competitor of 6900k shills just went on with 7700k
1600x came out shills realised it destroyed everything and they suddenly felt the need to "tdp doesnt matter"
TR came priced at the same amount as a housefire providing more perf more core and less tdp
"MUH CORES"

File: THE SHITWRECKER.png (80KB, 1700x1890px) Image search: [Google]

80KB, 1700x1890px

>>61659211

>>61659211
Jimbo Kellogg

>>61654206
now wait for actual software support
Ryzen 7 already gained 15-30% from applications optimizing.
...
not that many applications are ever going to use 18 cores/36 threads or whatever they bring now.

I'm looking to buy in some time between november and february. Looking forward to what they're going to squeeze out until then.

>>61654206
jews btfo

>>61660630
That's because Intelfags are mentally ill.

>Multi-core doesn't matter!
>Price/performance doesn't matter!
>Power usage doesn't matter!
>Temperatures don't matter!
>Soldered dies don't matter!
>Stutters don't matter!
>Streaming doesn't matter!
>Data centers don't matter!
>OEMs don't matter!
>Hyperscalers don't matter!
>Upgradeability doesn't matter!
>*NEW* Locked platform features don't matter!
>*NEW* Synthetic loads don't matter!
>*NEW* Burnt pins don't matter!

>>61659220
stop being anti-semitic. intel is just still trying to recover from anudda shoah.

>>61654950
Is this a warframe reference? you cuck?

So Coffeelake is 14nm but still hasn't been released. Probably get a full release right at the end of 2017 or early 2018. Cannonlake is 10nm but will probably be another year behind. So everyone has fully caught up to Intel in terms of process now? Other than the fact that Intel's 10nm is still behind the other fabs' 7nm process everything is gonna be released 2H 2018 or later.

File: nm process details.png (31KB, 707x797px) Image search: [Google]

31KB, 707x797px

>>61661279
Pretty much caught up or beaten in some areas.

>>61661325
>Intel
>1Q-17
That has to be wrong unless Intel is going to do simultaneous Coffeelake and Cannonlake release.

>>61661279
>Coffeelake is 14nm
process->architecture->optimization->rebrand
amarite?

>>61661591
Well at least this time it seems like they're going to up the core count.

>>61661553
The initial release of 10nm is going to be laptop CPUs only. And by Intel's own admission will be slower and have smaller die sizes.

>>61661616
you're right
let me correct that:
process->architecture->rebrand->MOARCOARS

>>61661660
more like this desu
lower clocks -> toothpaste -> mayonnaise -> more watts

>>61654206
NOOOOOO DELET DIS
PAJEET START ANOTHER VEGA SLIDE THREAD

File: 1499018882146.jpg (582KB, 2000x1000px) Image search: [Google]

582KB, 2000x1000px

>>61659685
>>61659822
>>61659860
There was a thread about this. Considering intel's bad business practices i wouldn't be surprised if they actually paid for modifying the statistics.

>>61661765
oy vey

>>61661765
yeah totally not some anon photoshopping

File: 1476847592312.jpg (44KB, 199x349px) Image search: [Google]

44KB, 199x349px

>>61662999
Are you fucking retarded? Did you just start browsing /g/ this week?

Okay, I'll spoonfeed you. Go look at the archives. Set the date to July 1st, search "passmark" or "BTFO" or "bankrupt" or whatever, and you'll see the first chart. Set the date to July 2-5th, and you'll see the massive shitstorm over the (((corrected))) chart.

File: 1500260377912.jpg (96KB, 638x710px) Image search: [Google]

96KB, 638x710px

>>61661765
DELETE THIS YOU STUPID GOY

>>61663153
>"BTFO" or "bankrupt"
to be fair, that should give you around a thousand results

>>61661765
D E L I D I L E D