So how about that Threadripper? The demo AMD last night sho

You don't know memory speed, nor how much core scaling is we're looking at.
Furthermore, you don't even know if that's final clockspeeds.

File: 1458626396347.webm (3MB, 1280x720px) Image search: [Google]

3MB, 1280x720px

>>60677082
They literally soldered two raisins together. What a piece of shit. Skylake-X is for me.

>>60677341
Assuming it could be better than what is given on release is never a good idea

>>60677380
>drakeposting
Old meme. Get out.

>>60677391
That happened with Ryzen

File: delid_technologies.jpg (85KB, 600x450px) Image search: [Google]

85KB, 600x450px

>>60677380
>literally soldered two raisins together
at least amd is using solder

>>60677773
until bios updates came out and we also figured out how to use it properly

Intel housefire running 2.8Ghz all-core.
You'll have to pay extra house insurance to buy it.

>>60677857
Not that, I meant AMD sandbagged their "leaked" benchmarks running at far lower clock than we got.

>>60677341
Scene uses less than 300Mb of ram.
Even if you had to read it all for each tile, at 30GB/s that would amount to µs.
Ram is not a factor in this test.

R7 1700 for 300 jewros.
Gonna build new workstation.
Feels gud mang.
Intel turds on suicide watch.
What a time to be alive.

>>60677341
Also
>you don't even know if that's final clockspeeds
Meaning it could get worse for Intel

>>60678060
Higher memory speed helps AMD's arch more internally do more work than simply fetching shit faster.

>>60678123
Supposedly that's less of an issue with Threadripper/Ebyn.

>>60677082
Some AMD dude yesterday said that TDPs are very competitive on these parts.

I expected 180, but is there a chance this will go on 160W? Intel would literally implode

>>60678147
Theoretically they should increase data fabric bandwidth by double over Ryzen to not choke the memory controllers.

Theoretically, we still don't have these things for real testing and infinity fabric is still 90% a pure mystery.

File: AMD-Ryzen-Threadripper-CPU.jpg (35KB, 960x678px) Image search: [Google]

35KB, 960x678px

>>60678149
It's going to be better binned and use better power management, apparently. They have 80% yields, so I don't see why they wouldn't save the best.

The seperate soldered dies and large surface area allow for very efficient heat dissipation. Look at that big sonofabitch.

>>60678123
But if I take my 1600x result @3.7Ghz, and simple math, I come up with 3.4~3.5Ghz for threadripper.
I repeat, that test doesn't even memory speed.
Try it yourself, downclock your memory and witness the nothing.

>>60678177
The reason for 80% yield is that they sell subpar parts as 1600 an 1500x/1400.
It will be the same on TR 10 to 16 cores.
They'll all have 16 cores physically there, but much of them disabled.
Actually, the fact that they're dealing with being ok with unfit cores helps the yield greatly.

>>60677082
>showed it rendering the Ryzen Blender test in 13.04 seconds
Holy shit.
Although from that screenshot the number of samples seems a bit low. Do we know how many they used?

>>60678295
that 80% yield figure was given for fully enabled dies.

>>60678345

150 I think, like their first Ryzen demo

>>60678348
http://www.guru3d.com/news-story/amd-ryzen-14nm-wafer-yields-pass-80-threadripper-cpus-on-track.html

>the yields for fabbing AMD Ryzen processors seem to be reaching strong values and are reaching numbers passing 80% for fully the functional 8-core parts.

File: 1468728335549.png (1MB, 1387x924px) Image search: [Google]

1MB, 1387x924px

This will be a nice upgrade.

>>60678410
Delid Dis.

DELID THIS RIGHT FUCKING NOW REEEEEE

File: Now+it+could+be+trying+to+draw+attention+to+the+_21f6faf63491f26529ec447c24381777.jpg (16KB, 351x329px) Image search: [Google]

16KB, 351x329px

>not buying into the literal housefire skylake-x 18 core
>not buying into the ultimate plausible deniability insurance fraud

Heheh goys you'll never learn how to best (((them))) at their own game

>>60677082
>pre recorded live presentation
had to be the BSOD from last time that got them all worried lmfao

File: toothpaste inside.jpg (31KB, 630x372px) Image search: [Google]

31KB, 630x372px

>>60679336
They were trying to OC some shit back then. Came to this thread just to shit it up with butthurt fanboy tears?

>>60679428
>butthurt
what, about the pre recording that shows literally nothing? Sure kid.

File: 174[1].gif (3MB, 400x288px) Image search: [Google]

3MB, 400x288px

>>60679500

>>60679556
where is your argument I don't speak autism

>>60679500
I've got some pre-recorded footage of your mom I could show you.

>>60679571
kiddo, my mother doesn't fuck niggers but i bet your daddy is open for anything tonight. Just ask him.

>>60677833
/Thread.

>>60677082
i wish they showed these results using linux. especially since they demoed using blender.

>>60679977
>i wish they showed these results using linux
I'd just like to interject for a moment. What you're referring to as Linux,
is in fact, GNU/Linux, or as I've recently taken to calling it, GNU plus Linux.
Linux is not an operating system unto itself, but rather another free component
of a fully functioning GNU system made useful by the GNU corelibs, shell
utilities and vital system components comprising a full OS as defined by POSIX.

Many computer users run a modified version of the GNU system every day,
without realizing it. Through a peculiar turn of events, the version of GNU
which is widely used today is often called "Linux", and many of its users are
not aware that it is basically the GNU system, developed by the GNU Project.

There really is a Linux, and these people are using it, but it is just a
part of the system they use. Linux is the kernel: the program in the system
that allocates the machine's resources to the other programs that you run.
The kernel is an essential part of an operating system, but useless by itself;
it can only function in the context of a complete operating system. Linux is
normally used in combination with the GNU operating system: the whole system
is basically GNU with Linux added, or GNU/Linux. All the so-called "Linux"
distributions are really distributions of GNU/Linux.

>>60679977
Blender's Cycles running on Linux is 3 to 5% faster in most cases, it wouldn't make a huge difference.

File: Screenshot_20170531-182607.png (571KB, 1080x1920px) Image search: [Google]

571KB, 1080x1920px

Is this a fucking joke?

>>60680978
What is? A chip with a heatspreadder and 4 times more transistors is bigger than a naked die?

>>60680978
>THYCC
>THYN

>>60680978
Why would it be ?

>>60680978
>amd is literally building all their cpus based on a single die, and just gluing more of them together as needed, and making small modifications as needed

best timeline

File: wew.jpg (476KB, 1080x1242px) Image search: [Google]

476KB, 1080x1242px

>>60680978
they glued 4 chips together to make threadripper

File: AMD-EPYC-CPU-740x416.jpg (26KB, 740x416px) Image search: [Google]

26KB, 740x416px

>>60681806
That's Ryzen Mobile next to Threadripper. The actual inside would look more like Epyc, but with only two dies instead of four.

>>60681806
Yeah, that's not what's happening here.

Neither 4 chips, nor are those smaller chips used in thredripper.
You only got the glue part right

>>60681806
will it only be able to cool with liquid?

i'm guessing it would need a big ass air cooler otherwise, probably to big for normal atx cases

>>60681888
No? You could cool Ryzen 7 on stock coolers, these are just two Ryzen 7's that are slightly more efficient.

Any decent air cooler will do assuming you got the plate to fit the socket.

>>60681806

Isn't that mobile chip an APU though? Seems a bit bigger than a normal Zepplin die. Single CCX+Vega mobile? Huh. . . gonna be pretty good for mobile shit.

>>60681933
It's only 10% bigger than Zeppelin

Probably taken up 70% by the GPU

>>60681888

Ryzen before 3.8-3.9ghz is stupid easy to cool. I have my 1700 at 3.5 all core, stock voltages, never hitting above 65c full load, with the stock CPU fan set on a silent curve.

>>60681933
>Seems a bit bigger than a normal Zepplin die.
Zeppelin is 195mm^2, APU die is ~210mm^2.

>>60681888
These are what threadripper air coolers look like. https://twitter.com/eteknix/status/869063679316295680

The massive surface area plus the fact that they're seperate soldered dies means it's actually fairly easy to cool them.

>>60677833
sick burn .... witch is also what you will get if you buy skylake x

>>60681954
>>60682217
how good are those APUs rumored to be?

>>60682448
According to AMD 50% faster CPU and 40% faster GPU compared to their previous APUs.

What market is this for?

>>60682455
honestly that doesn't sound very good, I was hoping the APUs would now be 4c/4t or 4c/8t, these 2c/4t current ones are shit

>>60682473
enthusiasts/content creators/meme of the week profession

>>60677833
>RFID chip

Lmao these fucking filthy kikes

>>60682492
If they're based on Ryzen I don't see why they wouldn't be 4 cores.

>>60682526
if they really were 4 cores the increase would be way more than 50%

>>60682547
lower clocks for lower tdp

>>60682547
see here:

>>60680592
>>60680681

>>60677082
Will buy day 1 if it's less than $1000

>>60682607
but what about desktop apus? those are the ones that I'm interested in

>>60682645
There are no figures for desktop APUs, only for mobile ones(assuming 15W)

>>60682657
then amd went full retard to put video outputs on the motherboards if they're not going to use them?

>>60682645
they are coming later, I don't know if the specs will be the same

but yeah, the desktop APUs need to be better

>>60682670
AMD doesn't make motherboards and desktop APUs are likely scheduled for January 2018

>>60682671
I'm still rocking a A8-6600k and a HD 6670, I need a good upgrade quickly

>>60677082
Why did AMD drop PGA for Threadripper?

>>60682791
4094 pins would be hell to set up, and would make the package even more massive than it already is.

>>60682791
because it's better to damage a $100~$250 motherboard than a $1000 cpu and >>60682813

>>60682840
Yeah but if you damage your pins on the mobo you'll never get them back straight.

Fixing PGA pins on the CPU is easy assuming you're not a drug addict

>>60677082
2500k +hd 6850 here.

Should I move to zen or wait for +? Also is Vega worth the wait or should I got meme rx580

>>60682900
waiting for zen+ is debatable, but you should wait for vega, it's going to be a lot faster than a 580

>>60682900
If you move to Zen now, you can always upgrade to Zen+ later. AM4 is going to be around for awhile. Vega is only worth waiting for if you care about 1070 or higher level of performance for (presumably) cheaper.

>>60682900
is your computer noticeably slow?
have you tried reinstalling windows if so?
if not wait until its worth it. im running an 8320 and probably wont upgrade until zen+ or a massive deal on a 1600x+mobo+memory

>>60678410
Add more Bingbus ;-&&&

>>60678410
holy shit, I thought it was 80% for everything, not only the successful 8 cores

File: 03_big (1).png (49KB, 849x621px) Image search: [Google]

49KB, 849x621px

>>60678149
>>60678177
Assuming literally two Ryzens, decently binned and running at an acceptable voltage, 16 core Threadripper could be a 140W chip.
And this is actual load power consumption, not TDP.

>>60683077
the rumors said 155W, but those chips probably have god tier overclocking because it's 2 separate dies

I'm curious why /g/ is so up in arms with intel vs amd shitposting when most of the new shit being revealed is super high end server CPU's.

Vidya benchmarks are irrelevant when you're talking about 1000+ buck CPUs. Are there a lot of developer/server hosts on /g/ that actually make use of this stuff?

>>60683119
It won't cock higher than Ryzen, even if you manage to cool it.
The voltage wall at 4GHz is still present

>>60683160
As a Gentoo user, I'm eagerly waiting for that -j33

>>60683160
HEDT is aimed at consumers. Or prosumers. Whatever.

Fact is, everyone has been waiting for Intel to get their comeuppance for sitting on their ass for 5 years and letting AMD get the drop on them with a more efficient architecture that scales better with high yields.

Well, at least I have.

>>60683166
I know it is, but having two 1800Xs on the same package is good enough for me

and it still probably overclocks better than intel's 16 and 18 core, because separate dies

File: 1496168689752.png (327KB, 1508x720px) Image search: [Google]

327KB, 1508x720px

>>60683166
Does it really matter when Intel's 10 core HEDT CPUs can barely get up to 3.3 GHz base? The 16 core probably won't be any higher than 2.9 Ghz. 18 core probably 2.7 GHz.

Turbobutt 2.0 and 3.0 does not effect all cores. All core turbo is seperate, unlisted and much lower, and you won't hit it w/o DELID job anyway.

>>60683160
>Are there a lot of developer/server hosts on /g/ that actually make use of this stuff?
I'm doing 3D renders, video encoding and heavy image processing sometimes. I'm on the fence about getting a threadripper, it'll depend on the price.

>>60683160
it's really just the /v/ side of /g/. they live for X v. Y shitposting

>>60683160
Because /pol/ degenerates drunk on 70 year old propaganda hate Intel because Jews and AMD made itself a meme from 2010-2016 with their trash bulldozer chips and derivatives

>>60680978
HNNNNNNNNNNNNNNNNNNNNNNNGGGGGGGGGGGGGGGGGG
Holy shit I hope RTG gets their shit together and does something similar with Navi. I'm not interested in Vega at all, I'm more interested in small GPUs scaling up to what's needed.

>>60683194
Not even a gentoo user and I would love to make -j32.

>>60682492
>>60682526

It's 4c8t, this was said in the presentation.
It's probably a singe ccx like the R5-1400.

>>60683480
Back into the gas chamber, filth.

>>60683620
1400 is 2 CCX with each CCX lacking 2 cores.

And before you even begin to mentioned caches, that's disabled too, half of it per CCX

>>60684341
1500X is 2+2
1400 is 4 + 0, that's why it has less cache, and no, it's impossible to disable cache within a CCX, you have to disable the entire thing

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

316KB, 882x758px

LMAO

>>60684422
No it's not, look at any latency measurements.
Also caches are split into slices precisely for this reason, both Intel and AMDs

>>60684441
I don't care, nigger

File: helper-contemplating.png (27KB, 741x609px) Image search: [Google]

27KB, 741x609px

>>60677082
>c:\users\amdin

>>60684585
>AMDin
Retarded frogposter.

>>60680978
are those tiny hands or is threadraper really that fuckhuge?

is it the same package/socket as epyc?

>>60684770
It's really that big. It's the same as Epyc.

That's cool and all, but how does it perform in scenarios that actual professionals use, and not some useless FLOSS garbage?

>>60677082
>>60678177
>>60681835
Hol' up
Are you saying that Threadripper will only have 16 cores when there is clearly 4 Zen silicon on a single package? What happened to the other 50% of the cores in each module? Deactivated for Jewery?

>>60685468
The 4 dies on one package is Epyc.

>>60685012
Like Ryzen, so garbage

File: RYZEN7-1800X-46.jpg (108KB, 614x407px) Image search: [Google]

108KB, 614x407px

>>60685012
>>60685482

http://www.hardwarecanucks.com/forum/hardware-canucks-reviews/74814-amd-ryzen-7-1800x-performance-review-10.html

Almost ties with an Intel CPU that costs twice as much, in the industry standard 3D modeling program. Yep, totally garbage

>>60685543
>>60685012
>not some useless FLOSS garbage?
>not some useless FLOSS garbage?
>not some useless FLOSS garbage?

>>60685553
>3ds MAX is "useless FLOSS garbage"

How about you FLOSS your teeth with my dick, fuckface?

>>60685305
I noticed you familia.

>>60684881
If it's the same socket as Epyc, is there anything at all stopping them from making a 3 die, 6 channel, 18-24c ThreadRaper+/Epyc- platform?

Intel is pushing SL-X to the breaking point with their late-to-the-game 14/16/18c models, and it seems like AMD has a hole in their lineup they could trivially plug if the 2/4-die socket compatibility really is true.

>>60685967
Not sure the motherboards are wired up for that.

>>60685981
the point being that a 3-die MCM could just use a subset of the pins the Epyc uses (dropping just the last 2 of 8 DDR4 channels and 32 of 128 PCIe lanes), leaving motherboard design just an exercise in just trimming the fat off an existing Epyc board.

>>60682364
unfortunally intel jewed again, shit tim on hedt

>>60677870
>jews sell you a house fire processor
>jews force you to pay for more house insurance
its like the jews win twice if you buy skylake-x

File: 12114212135.jpg (297KB, 1729x938px) Image search: [Google]

297KB, 1729x938px

>>60678520
7900X
4579mhz
CPU @AVX Frequency 4579mhz
BCLK 106.5mhz
Core Voltage 1.280v
Memory Frequency 4402mhz at 1.408v
XMP DDR4-4338 19-19-19-39

It's not melting.

>>60686277
That is the 10 core, not the 18 core. And it's lacking any other info (was it delidded, what kind of cooling, etc). Any source? System stability test? etc.

>>60678108
pls this. Intel deserves a comeuppance for their price gouging.

>>60677833
Top rekt
Intelfags btfo

>>60682862
Let's face it. If you are dumb enough to bend pins you deserve it. I have never bent a pin in my life. Only people with Parkinsons and clumsy fucking ass syndrome bend pins.

>>60678061
>Intel turds on suicide watch.
>What a time to be alive.
agreed
I'm actually excited about cpus again.
>>60678163
How does fabric work? I'm assuming they run throughout the chip connecting various bits.
>>60678177
What is a usual yield? That seems actually pretty low.

>>60681723
>best timeline
Yes. AMD is back.

Has AMD stolen some kraut space magic lately?

>>60686549
This. You'd have to be a retard using chimp retard strength to bend pins. There is no excuse to bend pins. None

>>60683291
That's just sad Intel. Under 3ghz? Wew lad

>>60680206
> 3-5%
https://www.youtube.com/watch?v=cNeJ3o4mZsk
I don't think so...

>>60686277
>43°c to run the bios
Guess I'm gonna buy a threadripper

File: jimkeller2_qvq6.jpg (365KB, 1772x1750px) Image search: [Google]

365KB, 1772x1750px

>>60686724
Even when just supervising, he can still create miracles. Zen is already ripping holes, Zen+ is going to obliterate Intel. Aren't they going to be behind a process node?

>>60687031
what's zen+ supposed to bring to the table?
But yeah, thank you based jim. Glad they brought him back.
These are exciting times.

>>60687051
Jumping straight to 7nm and general """IPC""" improvements meaning that even if they don't improve their not so high clocks, they will obliterate Intel.

>>60687111
I thought Zen+ was staying on 14nm, but to a newer process that would help increase core frequency and rely on refinement and optimization of the Zen architecture?
Where are you getting the 7nm jump from? That won't happen in another two years, according to GloFo

>>60677833
What am I missing here about Intel?

>>60687175
Intel has been using thermal cum paste and their chips are running hot. Solder has better heat transfer.

What are some benefits of going to 7nm?

>>60687203
>What are some benefits of going to 7nm?
More transistors per mm^2
Theoretically it's 50% of the size of the current 14nm process, but scaling doesn't work that way in the real world.

>>60687314
even a 30% improvement is good. I wonder what size we'll be at in ten years?

Picometer is next, right?

>>60677833
DELET THIS

>>60687382
>I wonder what size we'll be at in ten years?
We would be in 10nm if AMD didn't bring in the Zen.

benches

>>60687655
What do you mean?

>>60687680
Intel delayed their 10nm process AGAIN for the second year in a row. At this rate, we'll see desktop and other non-mobile 10nm processors from Intel in another two years.

>>60685305
Thank you for this.

>>60687699
What does Zen has to do with it? Intel reported terrible yields with 10nm even before the release of Zen.

>>60687846
Different guy, amd was fucking shit and not a real threat so intel had no need to innovate, now that amd actually released something good they not only forced intel to innovate they're making 7nm themselves. The other guy was just basically joking that without zen/ryzen intel would never release anything new.

>>60687699
see
>>60687846

Intel is having Fab production yield issues. AMD has nothing to do with them figuring out a solution.
>>60687959
While I think Intel was holding back a bit I don't think they were when it came to their shitty fail rates. They were certainly complacent with the current lineup and price gouging because amd offered little but they knew they couldn't do that forever and would have liked to have been in a better position and sorted production out so they could have a better answer to zen

>>60686277
>ln2 mode

File: typical modern scandi.jpg (102KB, 350x350px) Image search: [Google]

102KB, 350x350px

>>60686277
>LN2 Mode

>>60686277
>LN2 Mode

>>60686277
>LN2 mode
>43 degrees after getting into BIOS
EBYN :DDDDD

>>60687156

phoronix.com/scan.php?page=news_item&px=AMD-Financial-Day-2017

Two clicks on wiki pedia, you lazy bum.

>>60688332
>>60688730
>>60688813
>>60689394
https://www.youtube.com/watch?v=kyz2gkyIoXI

8 core running just fine at 4.5GHz on a fucking hyper212.

I know you AMDCucks think pooripper will do anything but get wrecked by a shitty TIM-based skl-x, but AMD winning will not be the case.

>>60687655
The normal reaction would have been releasing a stupidly rare/expensive 10nm chip to steal the performance crown.

>>60686277
>Positive 43° on L2N
So this is how the Jews are going to get revenge for the Holocaust

>>60689493
>10nm
>performance crown

Intel have stepped forward and said themselves that 10nm is going to be slower than their current 14nm node for some time.

And that's ignoring the 1.7% yields

>>60683160
>I'm curious why /g/ is so up in arms with intel vs amd shitposting when most of the new shit being revealed is super high end server CPU's.
Been waiting since 2010 for CPUs with more cores. I have had an old AM3 motherboard with an overclocked fx8120 since 2010 or so. The cpu, motherboard and ram must have cost me at most $500 USD. Up until recently intel only offered 6 core CPUs with high clock speeds, If I were to spend over $2000 on an intel HEDT I could probably just barely gain a 2x performance increase in multithreaded tasks.

Paying $2000 and only getting a PC 2x as fast as a $500 one from 2010 is not very enticing. They brought out 8 and 10 core I7s recently but the prices went even higher and performance per $ actually went backwards.

By my rough calculations a 32 core epyc would offer around a 10x performance increase depending on clocks. Even if it cost me $4000 it makes more sense than intel HEDT. As for threadripper vs epyc I will have to wait and see clockspeeds and prices. A thread ripper might be able to manage around an 8x increase if I can get all cores to 3.8ghz or so.

>>60689452
Still DoA.

>>60689627
I'd expect the 32 cores to cost somewhere between 2000 and 3000 burgers

>>60689452
>0.896 V
>4.5GHz
You're retarded

https://www.techpowerup.com/233945/amd-readies-nine-ryzen-threadripper-models

I feel sorry for Intel.
Charlie wasn't joking

>>60686567
>What is a usual yield? That seems actually pretty low.
Compared to the yields on Intel's monolithic dies, it's huge.

>>60690770
Damn, that's some serious firepower.

File: 1496164491711.jpg (43KB, 630x630px) Image search: [Google]

43KB, 630x630px

>>60690770
>desktop ryzen clocks
>155W TDP
>ECC

Is this how dead feels like?

>>60690770
th-that's glorious.

>>60690770
Fuck, guess it's AMD + Nvidia for me, unless Vega is really a Volta competitor but I doubt it, but AMD doesn't have such a good ecosystem for highend stuff like Nvidia.

http://www.anandtech.com/show/11491/acer-asus-unveil-35inch-gsync-hdr-monitors-ultrawide-curved-200hz


As for CPUs, they're delivering way above expectations.
AMD/Nvidia, would have never expected to use this ever since 2004

>>60688332
>>60688730
>>60688813
>>60689394
>>60689505
>LN2 Mode Disabled
come on guys

>>60690770
Is the 1800X going to drop below 499? They have to price the 10 core below the 7820X at 599.

>>60691007
It already dropped under $450 IIRC
It sold a lot less than 1700 and 1700X.

>>60680978
How .. mobile are they planning on being with this Raven Ridge chip, exactly?

It almost looks like they are trying to do what Intel failed miserably at; getting into the phone / tablet space. It may just be me getting a misleading impression but that looks a lot more like something you'd put in a tablet or a phone than a laptop or even a ultrabook.

Related, is Raven Ridge still scheduled for Q4 or more likely Q1 2018?

>>60681806
>they glued 4 chips together

Those "make more cores" meme cartoons we always see posted here

should become

GLUE MORE CHIPS

>>60691058
It's a 215mm^2 chip. That's literally laptop chips, Intel's highest GT4 configuration is the same size while their normal laptop parts are under 130mm^2

In fact this chip is actually bigger than I expected, AMD isn't skimping out on hardware even on 4-30W laptop parts.

>>60690770
unf

Prices will be interesting but specs at least seem to btfo the i9s.

>>60691058
It's not that small. The threadraper to the left is approximately the size of an xbox.

File: smug_lisa_su_face.png (355KB, 777x656px) Image search: [Google]

355KB, 777x656px

>>60684429
Your tears are DELICIOUS :^)

File: 1481744513695.png (349KB, 716x2588px) Image search: [Google]

349KB, 716x2588px

>>60685012
Nice try, but um no sweetie :)

File: 1494519493351.png (2MB, 769x8779px) Image search: [Google]

2MB, 769x8779px

>>60685482
LMAO

>>60683160
>>60683220
>waiting for Intel to get their comeuppance for sitting on their ass for 5 years

Yep. Notice how Intel has been giving us the same pure garbage with a +-1% performance difference for years and years..

.. and suddenly they announce the i9 and a lot more cores?

I'm quite frankly glad we didn't have to wait for Qualcomm and Mediatek to catch up to Intel before they would finally give us more than the same shit in a new wrapping. I really was starting to wonder if ARM or even Risc V would either kill x86-64 or force Intel to make some progress.

Regardless, I'm really happy that AMD is finally doing something that makes them relevant again an I'm also glad it forces Intel's hand.

>>60691208
>ARM
Continues to much x86 marketshare. AMD seems to be the only one willing to do something about it, Intel is content to sit on their ass and lose marketshare. By 2021 when Intel's new arch comes out, ARM would've continued taking marketshare to the point Intel would be seriously stressed.

File: 1472994550265.png (371KB, 832x868px) Image search: [Google]

371KB, 832x868px

>>60690770
HNNNNG

>>60691115
They could price the 10 core at $600 and 16 core at $1200 and BTFO Intel on every level. Keep in mind they have crazy high yields on this shit. >>60678410 (You)

File: image_id_1753037[1].jpg (458KB, 1920x1080px) Image search: [Google]

458KB, 1920x1080px

>>60689452
This would put the i9-7900X right next to the i7-6950X at the same clockspeeds, maybe +100 MHz improvement. So basically, no real improvement over the previous CPU, jizz TIM under the heatspreader, hits 80+ degrees (lol you didn't mention that), and $700 cheaper. And this was done by a professional overclocker. Wow, it's fucking nothing. Aside from the obvious overpricing being dropped.

File: Screenshot from 22-04-2017 16:06:05.png (158KB, 2033x909px) Image search: [Google]

158KB, 2033x909px

>>60691908
>no real improvement over the previous CPU,

But we knew that months ago.

File: 1407398745265.png (326KB, 3388x3638px) Image search: [Google]

326KB, 3388x3638px

>>60691908
>Core Voltage 1.624V
this cpu won't live for long

File: perfectlyreasonabletemperature.jpg (296KB, 1920x1200px) Image search: [Google]

296KB, 1920x1200px

>>60689452
I like those temps!

>>60692034
No wonder Intel is telling people to use watercooling, kek.

>>60690989
Those bad boys are going to cost 1 buttcoin at least.

>>60690989
Intel/Nvidia is way better

Intel 18 cores can OC to 4.8GHz

File: amd2.jpg (156KB, 633x758px) Image search: [Google]

156KB, 633x758px

>>60691186
I shall drink your tears forever AYYYMD

File: 1496325637853.jpg (37KB, 600x450px) Image search: [Google]

37KB, 600x450px

>>60692379
>TL note: at 400W and Ln2

>>60692379
>On LN
>On mayonaise
>On 3 cores
Enjoy your housefires

File: 1443516932885.jpg (11KB, 250x241px) Image search: [Google]

11KB, 250x241px

>>60677833
>delid
>lid is literally removed

>>60692409
The average life of a Intelfag is about the length of a Ln2 suicide run

>>60692379
>Source: My Ass

File: lisa_su_smug.png (123KB, 240x286px) Image search: [Google]

123KB, 240x286px

>>60692405
>Being this mad DOAlake-X getting dumpstered in its entirety

File: Power_01.png (72KB, 1302x2013px) Image search: [Google]

72KB, 1302x2013px

>>60692409
Poo in the loo, Poojeet.

File: 1496097659330.jpg (263KB, 1146x1430px) Image search: [Google]

263KB, 1146x1430px

>>60692448
Nice whole system values you stupid fuck. And running Excel? lol.

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

60KB, 1305x1434px

>>60692448
Kek too bad those power numbers are because the Ryzen's finished so much faster :^)

>>60692472
kek
>>60692448
just got btfo

>>60692448
>>60692472
>>60692504
The point of the discussion was not your shitty corelet/threadlet CPUs but 18 fucking cores and their power consumption overclocked.

>>60677082
>the possibility of running dual socket EBYN for 64 cores and 512GB DDR4 in less than 3 weeks
holy fuck
vega mia though

File: keller_amd_great_again.png (418KB, 627x627px) Image search: [Google]

418KB, 627x627px

>>60692537
Oh, I know Threadripper has utterly dumpstered Intel's entire Skylake-X lineup and most of their 1S server line up, but I just had to point out that the mad Intel fanboy BTFO'd himself with that Excel benchmark.

>>60692409
Lets assume Skylake-X is at least 15% more efficient(doubt it)

Then add 80% more cores, you're still around 370W

And that's at 4.3GHz.. at 4.8, that shit would fly well over 500 watts

I imagine you'll have a fun time seeing the socket melt unless Intel predicted this and just throttles the damn thing

>>60692537
I've seen multiple posts about this supposed 18 core Intel running at 4.8 GHz and no proof or links, so what else is there to say? It's bullshit, anyone who knows anything about Intel's architecture and how poorly it scales knows it.

>>60692577
>Lets assume Skylake-X is at least 15% more efficient
It's not.
>>60691908

>>60692577
>just throttles the damn thing
That's exactly what will happen, look at that CPU-z benchmark with a 4.5GHz Skylake-X 10 core, it's barely faster than a stock 6950X with 30% higher clocks.

Obviously it's throttling.

>>60692630
Source on this? Sounds hilarious.

File: Screenshot from 01-06-2017 16:49:55.png (745KB, 1260x673px) Image search: [Google]

745KB, 1260x673px

>>60692715
Don't laugh too much, it's kinda sad at this point.

>>60692715
see >>60689452 and >>60691908

>>60692755
>>60692744
Jesus Christ Intel what the fuck are you doing?

>>60692744
>core voltage
>1.764v

what the fuck lmao

>>60692763
Don't blame them too much, they had to push Skylake-X 2 quarters earlier so AMD doesn't demolish them.

Unfortunately that won't help them much at all.

>>60692771
It'sReadingItWrong(TM)

What does it matter anyway? Not like it's throttling or anything.

File: my sides.png (173KB, 640x400px) Image search: [Google]

173KB, 640x400px

>>60692755
>Intelfag posts some video thinking it'll BTFO AMD
>The exact opposite occurs

It's like pottery, right?

>>60692775
>2 quarters earlier
And HCC parts are not coming this year.

>>60692794
>Intel releases 18 core cpus thinking it will btfo amd
>The exactly opposite occurs
It IS pottery my friend

>>60692808
They are, but probably 2 months after LCC parts.

>>60692771
Divide the total voltage per core and you got real voltage

>>60691908
That broadwell-E is running with 3600mhz 12-12-11-28 281 1T timings against unlisted ram.

>>60692910
RAM doesn't do anything for cinnebench, and Broadwell-E's all core turbo is 3.5

>>60692910
Not only what >>60692949 said but you can literally see the RAM right there, it's some Corsair thing and the guy even put his name on them so I doubt they're cheap junk, come on now >>60692034

>>60692949
4603.33mhz

>>60693334
And it scored.. lower than Broadwell-E?

>>60687382
Afaik 4nm is the theoretical limit for silicon chips

>>60693365
Fucking hell.

https://www.pcgamesn.com/intel/intel-skylake-x-solder

> 4.3ghz is what they managed. They also mention 1.3v being the limit without "serious cooling", whatever that's supposed to mean. These things will be overheating, over priced and will, according to this Asus source, match broadwell-e in terms of OC due to thermal limits. That's an artificial limitation, not a hard one like broadwell's. And yeah, anything that isn't soldered is overpriced because it isn't worth the shoe polish TIM inside it. This was a shady, low, disgusting move by intel. Massive middle finger to their enthusiast customers who dared ask for a better priced product. We got that, and we also got TIM and less PCI lanes.

>>60693501
Sounds like even the Intel faithful are getting sick of this shit. Good. Hopefully a paradigm shift happens sooner than later. x86 can't survive with Intel at the helm screwing everyone over and not making any progress. New arch in 2021 my ass. Should be right fucking now.

>>60693501
Skylake-X looks like a fucking mess.

Same IPC as Broadwell, thermally limited to Broadwell's frequencies, less lanes, no ECC again.

Threadripper will eat Intel for lunch I fear.

File: good goy.png (604KB, 800x523px) Image search: [Google]

604KB, 800x523px

>>60693548
But hey, at least the prices are now "assrape with lube" instead of "assrape without lube".

>>60693548
broadwell was soldered if i'm not mistaken, right?
so this means skylake-x will have less overclock capability out of the box

>The 18 core CPUs are not scheduled until next year. Won't have them for a while. Either way, unless you're using the rig for rendering or encoding to make a living, no need.

t. Raja@ASUS

https://rog.asus.com/forum/showthread.php?93632-Late-June&s=f08ca0f5d8f6d1883b1530987df74d21&p=653561&viewfull=1#post653561


AHAHAHAHAH

>>60693624
Yep, Broadwell-E was soldered. Skylake-X is using toothpaste.

>They've been seeing a 30°C drop in temperatures by popping the top on the chip and using a liquid metal thermal interface.

Jesus fucking CHRIST Intel what the fuck kind of garbage are you putting in there?

>>60693641
so when intel releases their 18 core HEDT AMD will be close to completing their zen 2 lineup that will be at 7 FUCKING NM

HAHAHAHAHAHAHAHAHAHAH

>>60693641
So all we're getting this year is Broadwell-E rehash with jizz TIM and non-terrible pricing?

>>60693641
Any word on the 12, 14 and 16 core models?

>>60693641
Jesus Intel, is it gonna take you half a year to get some decent clocks out of this thing?

>>60693704
probably late Q3/ mid Q4

>>60693704
This doesn't look good, at best I estimate a September launch at earliest.
Might as well easily fall to November

>>60693711
They're probably expecting less than 600 people worldwide to buy one of the things.

>>60691120
+1 Internets to you, my good sir.

>>60693659

AFAIK a significant part of the cooling improvement with delidding is that you decrease the gap between the heat spreader and the die. When you use toothpaste, you have to glue the heat spreader in place and the glue raises the heat spreader slightly, which has a significant impact on the heat transfer.

t. first gen DELID THIS i7 owner

>>60677082
Serious question. Why does anyone need 16 or even 8 cores? What exactly do they do?

>>60693775
VMs, 3D rendering, video encoding, compiling, high end gaming while streaming off the CPU at the same time.

Just Cores Stuff(tm).

>>60693775
virtual machines, video editting/encoding/transcoding, compiling, real time encoding (so basically streaming)

tl:dr; if you don't know why this is useful you probably don't need it

>>60693819
So for normal people it's useless?

>>60693824
Normies don't need anything more than a low powered ARM chip. You're grasping at straws here.

>>60693824
Normal people don't buy $2000 semi-server processors.

>>60693824
for now, yes
some games are already making use of more than 4 cores, and this amount will increase in the next few years

>>60693839
I bet you think Ryzen is a good gaming cpu huh

>>60693889
>300 vs 400 fps
>oh no!

>>60693824
Define normal people. The majority of the PC market aren't gamers, they're OEM normie soccer moms and office spaces that could literally survive off of a thin client or a chromebook. People who need more cores include scientists who need to run a variety of programs in HEDT or clusters/supercomputers, high end developers, server/datacenters, virtual machine and thin client providers, massive multimedia. These guys are about as profitable if not more so than OEMs and will be more than happy to oblige themselves to running these processors.

If writing some 15yo's research paper, playing CSGO, and watching youtube videos were the end goal we wouldn't ever move beyond 4 cores and a powerful IGP.

>>60683160
I'm actually a department head and run a local data center for one of my companies headquarters. There are multiple data centers across the globe we use. AMD is the talk of the shop here. Depending on our leasing agreements it's likely we might be switching to AMD based platforms for some things. It's going to hinge on the enterprise market as a whole.

>>60693770
this is well known thing and one of the main rason amd cpus seems display much better temps than intels

>>60693978
Epyc is apparently selling out .. everyone wants in on the new arch.

File: 1439125308100.jpg (15KB, 250x250px) Image search: [Google]

15KB, 250x250px

>>60693775
>What the fuck would you even need 16 cores for?
>12 cores? Why do you want 10 cores? >Help! A madman is trying to kill me for 8 cores!
HAHAHAHAHAHAHAH

File: 1488678336902.png (408KB, 882x1418px) Image search: [Google]

408KB, 882x1418px

>>60693824
Don't worry you shekelchasing corelet, since it has the 2:1 infinity fabric and better binning the 10-core Kikeripper will have better single-core performance than an 1800x while also having better multi-core :^)

>>60693993
I don't doubt it. If this goes AMD's way this will all trickle down into the consumer market within a year or two at most. You'll start to see more optimizations for AMD than Intel in some cases etc. This is how it starts and happens with tech. I use AMD at work so I know it's good (or bad) therefore I tell mom and pop or a cousin to pickup that new Ryzen system and ignore that Intel.

>>60694050
Yeah, Intel should really have a new arch out by now. Instead it's out in 2020 at the earliest, 2021 probably. RIP.

>>60692744
Someone with an 1800x, fire up CPUz and run the bench for us?

>>60694081
CPUz recently gimped AMD's (and older Intel) performance because they were doing some function far too efficiently, it's become a joke.

It's only good for comparisons between CPU of the same vendor/arch

>>60694081
>>60694088
Yeah it's trash, don't bother. http://www.guru3d.com/news-story/new-cpu-z-upgrade-lowers-ryzen-performance.html

File: homulewd.jpg (63KB, 475x513px) Image search: [Google]

63KB, 475x513px

>>60690770
It's over, Intel is finished.

>>60694098
God these kikes are filthy, I've seen people actually defend the jewcum + mayonnaise and RFID chip in Skylake-X during the last two days.

No shame whatsoever.

File: ElPQdzp[1].jpg (159KB, 1002x1480px) Image search: [Google]

159KB, 1002x1480px

>>60694098
>whosbehindthispost.jpg
What the actual shit is going on.

>>60694098
how do they know this new benchmark isnt bugged on ryzen now? its obvious they talking shit inteljews paid them big buck

>>60694143
They intentionally state in the patch notes that they "corrected" ryzen scores, because they were too high.

>>60694143
that benchmark made very good use of the L3 cache, and ryzens L3 cache shits on intel's

File: Screenshot from 2015-03-16 01:57:47.png (372KB, 628x474px) Image search: [Google]

372KB, 628x474px

>>60694122
>RFID chip in Skylake-X


What the fuck is Intel even doing at this point

>>60694150
exactly what im saying, how do they know its correct now ? :) its obv kikes paid dem big buk

>>60694156
L2, not L3.

>>60694169
yes, L3, intel's L2 cache is better than AMD's by a wide margin

>>60694156
Not to mention AMD has put some serious effort into their branch prediction, so that shit just eats up simple instruction paths, like the ones that used in a lot of "modern" software, including the C++ benchmark that CPUz uses. So it's a case of
>Y-YOU'RE N-NOT ALLOWED TO USE FEATURES, THAT'S CHEATING!

File: 1489122339460.png (495KB, 1070x601px) Image search: [Google]

495KB, 1070x601px

>>60690770
NOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

>>60694181
AMD'3 L3 is higher bandwidth and lower latency per core than Intels(18mb AMD vs 20mb Intel, clear AMD win here)
AMD's L2 is slightly lower latency but much higher bandiwdth (AMD has twice the L2 per core)
AMD only loses in L1, both bandwidth and latency, especially bandwidth which is a small difference, but bandwith Intel's L1 is almost 60% faster. That's mostly for AVX pipes though, so you're not seeing it used much in normal x86

This is all seen in AIDA tests.

>>60694202
>slightly higher latency in AMD l2

>AMD only loses in L1, both bandwidth and latency, especially bandwidth which is a small difference,

Fixed that.

>>60690989
>nvidia + amd
enjoy gimped drivers

>>60694267
>Nvidia recommending AMD CPU

https://blogs.nvidia.com/blog/2017/05/30/amd-ryzen-threadripper-intel-core-i9-battlebox/


I really doubt it unless Nvidia is into the business of making their GPU perform worse

>>60694267
They fixed Rise of the Tombraider. I think the only game left that has AMD+Nvidia DX12 issues is Rocket League? Maybe?

There's literally no reason anymore to buy Intel.
Muh 144Hz gaming meme is for retards.
It's nice, but it's retarded.
If you're looking for image quality, you'll be buying a 1440p or 4K monitor.
Once you're there, CPU power is the last of your worries.
Believe me guys, I was quite perplexed about 4K, then I bought a cheap ass samsung 4k monitor for 400€.
I'll never go back.
My own brother couldn't even believe how good Tomb Raider looked on it, even though he had completed the game on his 1080p Gsync piece of shit.
Sure it requires a lot of horsepower, (2x rx480 or 1080Ti) but it's fucking amazing.
Now, what's going to happen is, all those guyz that have kept their 2600k,2500k,3770k,3670k for years are gonna start seeing their config die of old age, as mine did.
What do you think they'll buy anon?

>>60694334
Intel dual core becuase their best buy seller told them to

File: 1489880038374.jpg (18KB, 267x297px) Image search: [Google]

18KB, 267x297px

>>60677833
Y-y-yeah? Well fuck you

>>60694159
What can they do? New architecture isn't even close, and new process is somehow worse then existing.
This is what they get for sitting on their asses all those years.

Gigabyte ITX

http://www.guru3d.com/news-story/computex-2017-gigabyte-shows-ryzen-mini-itx-motherboard.html

So now we have ASROCK, Biostar and Gigabyte.
ANUS when?

File: 1430236296498.gif (112KB, 255x231px) Image search: [Google]

112KB, 255x231px

>intel toothpaste 18c 2.7ghz
>intel i9 holocaust furnace edition
DOA.

>>60694334
Intel because brand loyalty))))))

>>60694344
Personaly, I bought a 1600x from my deceased 3770k.
It's like the exact same thing, performance wise.
It just has 2 more cores, so faster winzips is always a win.
It was also cheap as fuck.

>>60694418
Why is it so small?

>>60694455
Someone correct me if I'm wrong, but I think most of the motherboard logic is integrated into the processor.
No south bridge required if you stick to basic functionalities.

File: Eye of the tiger.jpg (114KB, 928x488px) Image search: [Google]

114KB, 928x488px

>>60690770
>While you were partying, I studied the blade
Intel grew too lazy and comfortable at the top.

>>60694494
The thing bout Intel is, they could see it coming ages ago.
It was all public information that Zen was gonna be 8 cores a while before release.
Did they really think AMD was gonna another Bulldozer?
Even though they publicly said they worked on IPC very hard?
They had a fucking year to prepare, but all we got was Skylake2.

>>60694418
No chipset? About time.

Personally I think the chipset on Ryzen boards are useless, CPU already provides both SATA, ethernet and USB, why the fucking chipset?

File: High impact sexual violence...jpg (56KB, 400x489px) Image search: [Google]

56KB, 400x489px

>>60690770
Holy fucking shit. I'm tempted in getting one of these instead of the 1700.

>>60694543
Intel anticipated 40% IPC increase, not 55%

Big fucking difference, about the same difference between Sandybridge and Skylake lol

>>60694543
I don't think they expected this good scaling that allows AMD to challenge Intel at their server stronghold.

File: amcap.png (520KB, 525x482px) Image search: [Google]

520KB, 525x482px

reminds me of this

>>60694561
>implying there won't be a 12 or 16 core nonX version at $550/700 which you can overclock to stock X version clocks

I'm fucking doing it.
This shit is faster than a 24 core Broadwell Xeon, which runs at like 2.3GHz , that thing cost $7000

>>60694418
Biostar ITX boards are still sold out, guess people really want those 1600/1700 in a small case.

File: last shred of hope intel wojak.png (84KB, 653x726px) Image search: [Google]

84KB, 653x726px

>o-obviously they were holding back just to give AMD a chance
>t-they're going to break out their top secret tech they've been hiding all this time

>>60694601
>mfw first time in the history of forever Biostar held a monopoly on a popular platform form factor

Asrock and Gigabyte say nope

>>60694585
Just being curious, what use case do you have that requires that much threads?
I mean it's always good seeing Intel getting it's ass handed over, but I just don't think high end is gonna be a major revenue for them, even if they're not jews and will fine a fair price for it.

>>60689508
>1.7% yields
WEW LAD

>>60694646
Wait, what?
Source on that shit?

>>60683505
Especially seeing that Nvidia is going FULL INTEL with massive dies.

File: Girls.png (490KB, 449x401px) Image search: [Google]

490KB, 449x401px

>>60694639
Tabulating your delicious tears :^)

>>60694695
AMD should just give up on compute and go full VLIW5 500Bn transistors gaming monster.

File: 1496150690319.jpg (434KB, 1500x1557px) Image search: [Google]

434KB, 1500x1557px

That'll be $400

>>60694639
High end is where the money is, big margins and they have 80% yields so everything they make is pure profit.

>>60694714
Literally what?
I'm a frenchtard, so I just don't understand.
Let's be clear, I don't care who owns the crown, as long as it gives me better prices.

>>60694757
Okay? You know there's going to be more options than the RGB laser light shows they show off at trade shows, right?

>>60694695
600mm is maximum die size, reason for 800mm die size nvidia made is because it has 4 hbm memory sticks glued on.

>>60694714
You really need to stop attacking everyone who asks about what they need more cores for.

>>60694639
Solidworks, FireWorks, KVM, LAMMPS and ANSYS

While playing games

>>60694759
I'm not sure, man.
Volume is important.
It made money to Intel because it was so overpriced.
But I wouldn't put it past AMD to sell their 16 core shit at barely a 1000$, linear to 1800x pricing.
Maybe 1100$ because LGA and shit.
It's just how they are, if you've been following their shit for 20 years as I have.

>>60694609
>break out their top secret tech
Yea, yet another Skylake re-release.

>>60694772
cheapeast x99 in my country is around 220euros right now, i bet my ass cheapest x299 will be around 300euros at launch.
And there will be bunch of good goys buying 400-500mobos with 6-8cores, while you call literally 6-8cpu+mobo amd combo for price of x299 mobo

>>60694639
Threadripper is a workstation chip so rendering, CAD, Photoshop, programmers that compile a lot, scientific computing in labs, home/small business servers, etc.

>>60694809
released there overpriced piece of shit cpu for thousand bucks, same thing with that fury x duo pro piece of trash

>>60694824
Don't you just use GPUs nowadays for such tasks? (besides compilation)

>>60694757

Don't forget quad m.2 drives in raid 10 on a single x16 Dimm.2 card.

>>60693775
>Why does anyone need 16 or even 8 cores
I've got 16 cores in my main server (2 Xeon E5-2687), and I end up using pretty much all 32 threads. 2 of those 12 threads are used for encoding 2 separate streams of 1080p 25/30/60FPS 8-bit color 4:2:0 near-lossless h.264 raw footage to a much lower bitrate h.264 output at the same time. The other 8 threads are used in 10 different VMs attached to my network.

>>60694845
No.
GPUs are only good for non-branching parallel stuff, most x86 code is not non-branching.

>>60694845
Something has to feed the numbers to the GPU fast enough.
Also, not all computing can be done on GPU and some things (like video encoding) have to be done on CPU for best quality.

>>60694790
Why? Every single one of them repeats the exact same shilling bullet points in the same order every time (except the few days between Skylake-X spec leaking and Threadripper specs leaking of course):

>1) Why do you need so many cores?
Multitude of use cases are given

>2) But aren't GPUs better for all those things?
No

>3) Aren't most desktops just for gaymen?
No

>4) LMAO 7700k gets moar FPS at 640x480 in DX11 gaymen! MUH ARMA3 MUH DCS!

Rinse and repeat.

>>60694926
Grab a beer and stop taking these threads seriously.
The beer helps tremendously.

>>60694926
Turn off the PC and go outside for awhile.

>>60694881
>>60694901
Well I guess you're working on fucking insane amount of data then.
Also,branching can be avoided a lot of times by just doing blank calculations.
I'm kind of doing my own thing with OpenCl, and I'm still wondering how I could ever reach the 8 Gb limit. Even then, I think I could somehow manage streaming the data on the PCI-E bus to prevent it from choking.

>>60691950
F

>>60692034
85 c? Jesus. Put a frying pan on top of it and crack some eggs. You can cook breakfast while you work, brilliant idea Intel!

>>60695156
It's reading t wrong

>>60692585
>It's bullshit, anyone who knows anything about Intel's architecture and how poorly it scales knows it.
This
>>60692744
> core voltage
JUST

>>60695286
lel. Yep the monitor directly off the motherboard is wrong, thanks Obama.

>>60693376
Oh wow. What's the next material? Or will they stop and just focus on chip design?

>>60692744
>1.764V
WEW LADDY

>>60691019
1700x just got a $50 price drop 1700 $10 I wouldn't be surprised to see the 1800x drop another $50

File: muffled Eurobeat.png (240KB, 500x297px) Image search: [Google]

240KB, 500x297px

>>60694757
Did they glue two 1800x or is one 1800x a 19**X split in two? Either way, I love that AMD is just scaling their CCX anyway they want and it isn't shit like what Intel tried before.

>>60695286
Oh yeah? Then what's it supposed to be, wise guy?

>>60695848
tMeltdown

>>60695847
What did Intel try before?
>eurobeat
/o/ fag?

>>60695874
The same thing AMD is doing by "gluing" for for more cores. I think the Pentium D or the Core2 quad was done that way.
No, not from /o/. I might have grabbed it from there though.

>>60695847
It's two 1800X on one CPU.

>>60686277
>>60689452

AMD SO BLOWN THE FUCK OUT

HOLY SHIT HAHAHAHAHAAHAH

>>60682900
Rumor has it the next desktop zen (not zen2 but zen+) will be the brother of the ironed out b2 stepping EPYC dies with the fabled full speed 2/1 ratio memory interconnect, ~10% higher clocks, same power use.

Rumor also has it these wont be out for months and months if not til next year.

File: pHrnT.jpg (184KB, 980x602px) Image search: [Google]

184KB, 980x602px

>>60696773
Too bad it wasn't any faster than a 6950X at that OC :^)

>>60677833
>Flawless Victory
/thread