HAH AHAH AAHA HAHAHAH A HAH HAH AHAH AAHA HAHAHAH A HAH HAH AHAH AAHA HAHAHAH A HAH HAH AHAH AAHA HAHAHAH A HAH HAH...

he didn't say which CES ;^)

sometimes i have the feeling im gonna die
and im ok with that

>what is i7-7960x

>I don't get the point of this thread
>can I get a quick rundown

Some C-list e-celeb made came out with a "leaked" lineup from AMD that was supposed to be announced (if not necessarily launched) earlier this month.
AMD announced fuck-all for actual products, but they teased an 8c early sample running neck-and-neck with an i9-9900K at much lower power but undisclosed clock speeds.

This pic:
is a closeup of the teased chip's substrate showing room for (and subsurface electrical wiring) a 2nd 8c chiplet to fit in the package, basically evidencing 12c if not 16c Zen 3 parts on the AM4 platform.

AMD is going to be the undisputed king of MOAR COARS this summer, but fuck-all is still known now about performance, power, and pricing of the final products, so everyone can basically shitpost freely for the next ~5 months.

OP thought this was the Wccftech comment section.

Attached: 1546404431152.png (384x380, 31K)

an overpriced housefire

>fuck-all is still known now about performance, power

except that amd dabs on intel at half the power draw and only desperate incels honestly believe a 65 watt part is going to be their flagship

prerelease beats 9900k at lower power. final will fucking demolish intel at stock and leave nothing but a smoking crater overclocked.

A CPU that costs over triple the eventual price of the 16-core Zen2 CPUs that will outperform it at far lower power draw and temperatures.

I just want a good desktop APU. I guess I'm going to be stuck with a 2400g or a 2200g for at least a year.

Excuse me sir, please delete this immediately. It endangers our narrative.

Attached: 1514927526947.jpg (267x297, 16K)

amd ceo basicly confirms 16 cores
amd cto basicly confirms 16 cores

some randomg coping Jow Forums autist says otherwise

well i guess i should trust Jow Forums

So does the Ryzen 5 3600 exist or not? I want to upgrade from this 2500K to something decent.

Jow Forums also said in 2016 zen wouldn't release 8 cores for less than 750 burgers. Jow Forums said zen would be in the 2ghz range for clocks. Jow Forums said zen would have IPC around sandy bridge level.

In short, Jow Forums is always wrong.

What we've seen so far strongly suggests that 8 core ryzen 5s are a reality, but we don't actually know at this point.

So it's the same as always.

NOOOOOOOOOOOOOOOOOOOOO

Attached: 1506977173618.jpg (882x758, 324K)

7NM IS NOT REAL
ROME IS NOT REAL
CHIPLETS ARE NOT REAL

Attached: 1527629778452.jpg (679x758, 54K)

DOA garbage trash

>little Timmy need more dopamine thread #48732

>the lack of 1 component send Intelavivs into a deep depression
how does she do it?

Attached: 1547697586525.jpg (218x250, 7K)

Attached: AMDISCOUNT BTFO.png (2782x1250, 1.69M)

next time can you please include "amd" or "intel" in your title? stop going around my filter list, thanks.

>mobile processors

cringtel back at it agane

Its the only area where Intel still holds the advantage and thus - suspiciously - becomes the most important thing to Jow Forums.

Reminder that a Ryzen 5 was the CPU shown at CES.

oh no no nononono

Lisa is just THAT powerful.

A 65w non-x Ryzen 5, at that. And based on the single core to multi core score ratios of the 2700 and 2700x, the ES shown at CES would have had to score around 200 in single core to score 2050 in multicore. An 8700k scores ~198 in single core.

With the 65w non-x part having 8700k tier single core performance, the 95x x version would only need to be 10% faster than the 65w ES to beat the 9900k in single core, while dumpstering it in multicore. And that's the mid-range chip.

The 16-core version is going to posterize Intel's entire HEDT line up, and do it on a mainstream socket. That terror is why Intel created that joke auction-only overclocked 255w thermonuclear house fire off-road map 9990xe, by the way.

Attached: Cinebench-single (1).png (1327x1222, 64K)

mesh bus based shit

inferior technology

>65w non-x Ryzen 5
Why does this matter? I thought amd & intel haven't been following their tdp rating for years now?

Intel's TDP ratings are bullshit now, but that's not relevant to what I was referring to. AMD makes lower power and higher power (non-x and x) variants of most of their desktop range, and the x versions are always clocked higher because they are allowed to use more power.

The 2700 had a 65w TDP, while the 2700x has a 95w TDP. So my point is that the 65w ES shown at CES is clocked lower than the 95w version will be, and STILL roughly tied an 8700k in single core and a 9900k in multicore.

>while the 2700x has a 95w TDP
it was 105.

Intel's TDP is set based on the chips base clocks and completely ignores their boost means yes, if you lock a 9900k to the 3.6ghz or whatever the fuck its base clocks are it really does have a 95W TDP.

Also: power draw is not TDP (even though they are correlated).

so did I fuck up by not buying a 2700x when they were on sale a month ago or not?

t. 2500k and r9 290 poorfag

>So my point is that the 65w ES shown at CES is clocked lower than the 95w version will be, and STILL roughly tied an 8700k in single core and a 9900k in multicore.

Damn, that's mighty impressive if true. Thanks for clearing that up.

Is Shlomo bootyblasted that AyyyyMD is selling 6C/12T chips for fucking i3 prices? Eat a bag of dicks.

You're a winner for not buying the 2700x or 9900k. Once ryzen 2 is out, it'll offer higher than 9900k performance at 2700x prices.

I had 2400 (non-K) and a 750Ti until two months ago. Went to a 2700X and a 1070Ti. Pic related.

Attached: feels-goodman.jpg (475x360, 34K)

yes this is an assblasted intelkike thread.

It is now.

Attached: 1525551728082.png (937x581, 161K)

how will IPC be?

CPUs are pretty good right now as far as options go, but GPUs are the thing holding me back. I wouldn't want last gen, but current gen nvidia is fucking expensive for small improvements

Attached: thats-not-just-savage-thats-vicious.jpg (1280x720, 423K)

Yup, I may have to upgrade my r7 1700 to a 16 core this year... These first world problems...

Screen capping all the intelifags because it will make Q2 - Q3 2019 that much sweeter.

that's the best part

youtube.com/watch?v=OxZPfMXtuF4

Attached: zen2.jpg (1280x720, 276K)

the burn, it is sick.

Just checked and you're right, I was mixing up the 2600x's TDP with the 2700x. Doesn't change get the underlying point though.

I'm aware of all that, I just think it's fair to describe a TDP set based on conditions that will never actually exist in use as "bullshit".

Well if Intel put is foot down and demanded mobo manufacturers stick to their TDP guidelines and not have the likes of MCE enabled all the fucking time it seems a lot less bullshit. Its why in the likes of Dell's optiplex units and other such deployments chips won't boost basically at all because the board adheres to TDP guidelines.

>The 16-core version is going to posterize Intel's entire HEDT line up
cucked into oblivion by chiplets, kek

Attached: 1519946649691.gif (200x200, 1.32M)

Didn't intel do this with their Bentium D and Core 2 Quads?

nah core 2 was answered pretty solidly by phenom 2

the real unrecoverable btfo was nehalem aka gen1 i7. answered by fucking bulldozer, which only bulldozed itself into a hole in the ground by which time the 2600 was already out.

Not entirely - despite the raw processing pwoer on offer you are still limited to dual channel memory which brings its own limitations (both in effective bandwidth and density). Thats why AMD can afford to bring monstrous core coutns to the desktop and not eat into their threadripper sales as - much like now - core count isn't the only thing that defines a cpu's performance and TR offers a lot of features not found on ryzen.

Threadripper zen2 chips will, however, annihilate Intel's HEDT lineup (remember user, double double).

You're truly delusional if you think this will actually be the case.
Even at 10nm AMD's TDP figures are shockingly high.

>AMD
>10nm

Attached: hmmm.png (470x454, 11K)

I meant the chiplet thing, IIRC they had to slap two P4 chips on one substrate, and again with two C2D's

>(remember user, double double).
how can i doubledouble my boner even further beyond

Attached: 1541538409899.png (298x279, 30K)

Why does AMD almost always have to be the one with the innovations

head to head test proved it, either dwi or kys

I really want to see what the fuck that IO die is for

P4 was shit from the start, putting two piles of shit together won't make it better.

Because Intel has always sucked at innovating in processor design and has always relied on its node advantages, marketing, and shady tactics to carry it in the marketplace.

Not arguing that point, user.

>intel's response to chiplets is putting two dies ON TOP OF THE OTHER
Why, the heat dissipation will be fucking retarded

Are you laughing because AMD will most likely actually surpass these leaks within the year?

so far this prediction is on track, why are you laughing?

The R5 from CES was clocked at 4.5 and the 3600 from that list is 4.4 lul

OP is either trolling or more likely false flagging, but the fact of the matter is that Jim waaaay overpromised, even after he clarified to product "announcements" and not "launches". We're not getting an actual official product lineup until June.

Also, AMD never actually committed to 16c Ryzen 3 AM4 parts and is being weaselly about potentially stopping at 12c:
> youtube.com/watch?time_continue=488&v=M5S_qX2wYvY
> I think that you might expect that will have more than 8 cores. I didn't say how many more. I said more.
There is a chance 16c Ryzen 3 parts will come out, but they might want to hold back at 12c for now due to TDP caution, clearer segmentation vs. the Threadripper line, wanting to hold back something for the 7nm+ Ryzen 4 launch, etc.

do you think lisa would agree to be my mom?
youtube.com/watch?v=GcE-20XauqM

>There is a chance 16c Ryzen 3 parts will come out, but they might want to hold back at 12c for now
You clearly don't understand AMD's product strategy for x86. They're going to double core counts up and down the product stack, and as soon as they get on 3nm, they're going to double them again.

AMD can never stop being an 'also ran' in the x86 CPU market unless they disrupt the market, so that's what they're going to do. This is also why they are combining ticks and tocks every time they can. They want to push the pace of innovation to the max because they believe they can out innovate Intel and seize considerable market share under those circumstances.

Whoa.. calm down

yikesfu

Attached: jim.png (1531x902, 842K)

It blows my mind how much time people spend typing out long replies like this

>5 lines
>long reply

Where did you see the clock speeds?

It was an engineering sample using 50w less to equal the 9900K in multicore cinebench. That's all we know.

You do realize that the core-doubled parts are going to have considerably reduced clocks to stay within TDP, right? E.g., 64c Rome is anticipated to have base/turbo clocks in the neighborhood of 1.4/2.2 GHz, which will still be enough to handily beat CL-AP in cores*Hz and c*Hz/W.

Workstation chips have a higher Watt/core budgets, but you're not gonna see 16c parts with great looking all-core turbo speeds in a ~105W TDP. If AMD works with board vendors to recertify an AM4 derivative as a ~130W TDP platform, 16c Ryzen 3 would be more reasonable, but companies generally avoid goofy arbitrary platform splits to not confuse and annoy customers.

AMD is not beyond 125w parts

AMD has a storied legacy of making HOUSEFIRE tier parts from time to time, but AM4 is not currently certified for that. In contrast, TR4 was always a 250W platform that initially had only 180W parts (i.e., basically a designated overclocking platform) before moving to the 24c/32c parts this Fall with base 250W TDPs.

To make a decent 16c Ryzen 3, AMD would have to have effectively split the platform into low- and high-power segments after the fact, with separate AGESA firmware for different mobos depending on whether they could get recertified, etc. It's clearly not impossible, but it's not hard to see why AMD might be hesitant to over-promise here.

> 16c Ryzen 3
Imagine how much Ryzen 7 will be able to offer. Crazy!

meant 3rd gen/7nm Ryzen, not R3 gimped core series.

fpbp

>The Jew always wins!
Intel now sell crippled CPU's with no iGPU for the same price. Intel shills will defend this.

> I meant the chiplet thing
No, they've never put an I/O part to a separate chip. MCM chips by themselves are nothing new.

Attached: bluey.jpg (499x374, 30K)

just buy it

Attached: 1492855909643.png (1228x1502, 944K)

Multicore pentium Ds, C2D, C2Q and the first gen of icores(i am not 100% sure about this one) were multiple single cores into the same package.
They didn't have cache coherency, each core worked as a single core, it was not aware that it was on a dual or quad core sku. The OS was responsible of distributing the load on each single core.
This is not a mcm or a chiplet design, it's just glueing together several single cores and call it a multicore.
Athlon X2s had a proper multicore design. There was cache coherency between the 2 cores.
What difference does it make? Well, performance.
Also all intel's cpus with cpu+igpu have the same design, the cpu cannot share data or utilize parts of the gpu.
Amd apus can do that. This one is called HSA. The previous design is called UMA.
Often intel's skus had e.g. 14nm cpu part and 22nm igou part. Intel did that in order to drive cost down.
Chiplet is part of the operations of the chip in a different die. Intel has never had any of these.