Intel Core i7-7700K Kaby Lake

they priced it cheaper though
Intel is fully aware of what they're doing

>>58345808
It's a meme for desktops. Wait for Skylake-E for that.

For laptops it will have battery life and onboard video benefits.

Also it's more believable that silicon is hitting a plateau and going smaller is impractical, and that Intel may be focusing on non-silicon, like CPU manufacturers should've seen coming a decade away

>>58345808
Normies will think 7700k is newer and a bigger number than 6600k and therefore must be better, so it will still sell.

>>58345875
This. I don't understand why they're stretching silicon to it's absolute limit. Wouldn't they have better yields and easier development with fewer setbacks switching to Germanium now?

>>58345908
Also, may as well have taken the time to get the infrastructure to produce Germanium chips in place while AMD was stagnating. AMD would be over if Intel were shooting through 10, 7, 5, etc. nm on schedule instead of wringing out Silicon.

>assuming the typical consumer is smart

a chunk of the market will buy it because its new with no care as to its actual performance.

>>58345875
>>58345908
As someone in the industry I can tell you that they're doing just that but finding a replacement for silicone is much more difficult than you might think.

The industry's actually pretty fucked

>>58345907
im sure newegg reviews full of people claiming huge performance gains already

>>58345958
I'm also in the industry
Intel as a company are being pretty arrogant about PCIe 4.0 collaboration
Sandisk is still finding and choosing suppliers
Hyundai is keeping the same course, they're good customers
But of course people think PCIe 3.0 and 4.0 are completely different technologies and should be dubugged and optimized by whole new companies for some reason

>>58345808
In my eyes the biggest selling point for Kaby Lake is that it natively supports everything required for 4K HDR content all by itself. This doesn't really mean anything for people using discrete graphics, but it does mean that the laptops and things that most people will be buying will now have devices that can playback 4K HDR content, so we'll likely see 4K HDR adoption rates rapidly increase this year.

>>58345808
Kaby Lake is just Devil's Canyon 2.0.
They managed to tweak their buggy 14nm process a bit, they fixed some hardware bugs, and introduced a new media decoder to the benefit of their mobile SKUs. The only thing they did wrong was present it as a whole new generation of arch instead of the refresh that it is.

>>58345908
>>58345875
>when people who read popsci articles try to articulate ideas they can't fully grasp
Hoo boy.
Foundries are sticking with bulk silicon because each node is increasingly more expensive to bring online, sub 28nm was $3 billion+. and this is with guaranteed customers. Every new node shares a majority tooling with existing lines to reduce cost and time to market. Switching to a radically new material an entirely new substrate, would be an investment of tens of billions of dollars, and it would have uncertain results. They're not magicians and there is no rabbit to pull out of a hat.
SiGe is already widely used, and is introduced as a channel material in IBM/GloFo's 7nm SOI FinFET process. New high perf SOI is major progress, not that anyone on this board understands the difference or history of the two substrates.

It is logical to continue down this path, trending towards GAAs before considering gutting the entire basis of photolithography.

>>58345958
>Silicone
Ah yes. That must be why there are so many boobs on the internet.

>4k Netflix streaming
>Pentiums with Hyper Threading
>Unlocked i3
>DDR4-2400 support on the B250 and H270 mainstream chipsets
>Bigger OC headroom

Kaby Lake is a massive improvement over Skymeme.

>>58347282
>bragging about drm

I'm mulling this or the 6800k

Thing is I'm not sure in investing with a CPU with diminishing returns or an expensive platform that is almost obsolete

>>58347532
>X99
>Almost obselete
It's more capable than skylake and Kaby lake, both quad core maximum platforms with dual channel ram and limited PCIe lanes
Also the point of getting X99 is to get a balling CPU, most of which happen to destroy the 1151 CPUs in multithread and match them in single thread. A 6800k is just a 6700k/7700k with extra cores, but that *can* be a big deal

>>58347532
LGA 1151 is the only thing that's almost obselete here, you can expect the X99 market to stick around much longer than Z170/Z270, it's much more capable
At least until X299 or whatever the next upgrade is for that platform
X99 and Z170 are made for different CPUs and different demands, X99 is objectively better

>>58345868
Skylake has zero performance gain over Boardwell. Just buy a Boardwell-E and don't wait until August. If you are gonna wait at least wait until Zen comes out in March and (hopefully) AMD won't fall flat on their ass.

>>58347591
X99 only has PCIe 2.0. That's outdated shit.

>>58347574
>>58347591
>Intel chipset obsolete
But Intel releases a new chipset every generation

>>58347644
>Just buy a Boardwell-E and don't wait until August
Why?

>>58347700
>>Skylake has zero performance gain over Boardwell

>>58347282
My Skylake i3 easily reaches 4.8GHz, for $99. Why would I buy a Kaby Lake i3 with the same performance per clock, at 4.9GHz overclocked, for $177?

so if you need to o/c a 6700k to the same results, what happens if you o/c a 7700k?

>>58347646
Any guesstimates when the X99 successor will come out?

Hexacores should be the norm now

>>58347756
It will be X299
They've already said it will be only 4 cores for Kaby Lake X
6-10 cores for Skylake X, though

>>58347756
Late 2017

>>58345860
>they priced it cheaper though
Not seeing this on Newegg. Skylake is cheaper for now.

Looks like a 10% improvement across the board.

And given that I'm still packing my i7-920 it's going to be a huge upgrade.

File: bait.jpg (217KB, 3000x3000px) Image search: [Google]

217KB, 3000x3000px

>>58347282

Red pill me on the new intel jewery I am on 2500k. Should I Zen

Since intel don't discount thier old chips it will just replace the 6700k. Also where is the hate for AMD putting out slightly faster APUs for the past 5 years?

>>58347176
elitist fuck

I'll be getting one since I'm putting together my first build this year, one that is more than ready for 4K content. I feel like I'm getting meme'd but I also want that 5GHz OC for emulators and some older games.

>>58348301
Found a boob.

>>58347644
DDR4 though, that's gotta count for something.

>>58347748

c'mon guys. what are the results?

i'd never o/c a chip, too lazy

but i'd sure as fuck install the best that 4 cores could offer

>>58348233
I think there would be more hate if said APUs were the performance leaders for all of amd's processors.

>>58347028
HDR looks like trash

they hit 5ghz. that's a landmark and bodes well for 6 cores with soldered IHS's

>>58347282
>Unlocked i3
top kek. you might as well buy an ayymd FX

>>58349119
>i'd never o/c a chip, too lazy

presets are literally built into the bios. takes like 3 clicks of the mouse and hit save.

File: e57698f7d84986974df3fbfebdcd7a58c85141cafb9b4ae38af7db5f1838a16d.jpg (8KB, 250x234px) Image search: [Google]

8KB, 250x234px

>>58349103
Not a whole lot when you have dual socket 2011 and have dual quad channels that can push 100GB/s.

Why the hell did intel drop quad memory: Who the fuck knows.

File: dsaukfsfoisajfois.jpg (9KB, 184x184px) Image search: [Google]

9KB, 184x184px

>>58345808
>mfw kaby lake is shit

Who the fuck cares about consumer low-end shit? There is no life under 8 (intel) cores.

I'm buying one. but I got a z270 mobo for 130 bucks cdn so that kinda balances out the cost. and I'm upgrading from the ole 2500k so pretty significant for me

>>58349103
broadwell e already has ddr4

>>58345808
I just want it to come out so I can maybe get a cheaper 6th gen already.

>>58349911
It came out today..... 8% gain I don't think any discounts are likely but maybe.

>>58349928
Oh it did? Nice. Either way I need it to hit a store near me or the prices will have 0% chance of shifting.

>>58349911
it's already on newegg

>>58349957
Some of us don't live in the US.

>>58345808
I've read compelling arguments that Intel planned to ditch 10nm/silcon acknowledging that Moore's Law is no longer sustainable and were switching to OQC but due to ongoing ECD bankruptcy proceedings they haven't been able to make the jump. This caused them to have to play catch-up on the 10nm process.

If this is true, INTC and MU will dominate all and both NVDA and AMD will get crushed (though many are speculating some form of buyout/merger of AMD and MU, though I don't particularly like it).

So, should I upgrade my haswell to skylake?

>>58345808
But it's still better than the 6700K right?

>>58350047

it is an upgrade, yes.

>>58350082
Unless there's a downside I'm not really seeing a problem then. It's not like it's super expensive.

>>58347176
How are they using SiGe? I can't imagine it's in patterned wafers so is it some deposition thing? To what end?

No alternative optoelectronic setup is made these days without patterning of some form.

>>58349957
Why the fuck is it the same price as the 6700K

>>58350090

just like every other advancement in cpu history

except now it's a bunch of guys who look forward to this stuff like it's the olympics.

small steps, not huge leaps, keep the tech industry moving forward

>>58345808
This reminds me of the move from Northwood to Prescott.

>>58345808
The box isn't even different.

>>58350152
Pentium 4 Damage Control edition was so hilarious.

Anyone else think it's odd that intel is treading water right before AMD drops their biggest CPU upgrade since the Athlon 64?

File: pnY4eLq22k4KcHqv3lI6xFqzsPGBsRmJEEcX4H2tDN4.png (892KB, 1023x768px) Image search: [Google]

892KB, 1023x768px

>>58350127

Are there any power usage benchmarks? I would like to see if they at least made it more power efficient, since at least that makes sense.

>>58350202
lmao

>>58350163
>>58350152
regale the millenialfags with stories bitte

>>58348304
At least wait for Zen.
Worst case is that it's awful and you buy kaby lake anyway.
Best case is that it's great and sends the prices of intel chips down.

Note that best case is not that Zen is better than kaby lake, just that it's good enough that it makes the intel prices look fucking stupid

I don't understand why these threads are necessary. It was known that kabylake was another 4790k in action. Intel has the luxury of selling better clocked versions of their current consumer line since their main competition is currently non existent. The moment you saw that they abandoned their tick tock bullshit is the moment when you knew it was just a shitty refresh. At least it has some HW decoding, what a piece of shit. Zen better deliver. I'll take their 8 cores option if it can hit 4GHz and it really has close if not better than Broadwell performance.

>>58350253
It may very well be better than babby lake.

>>58350246
Back in the old days intel was getting wrecked when AMDs Athlon was fucking the Pentium III, Intel was desperate so they had their R&D department work night and day.

AMD was nearing the 1 ghz barrier, and intel was scared that they were gonna get beaten to 1 ghz. So after working endless hours intels R&D team finally was able to develop the technology to to beat AMD to 1ghz, the paper launch.

(nothing was paperlaunched before the 1ghz pentium 3, which never really became available)

hooray

>>58350253
Why won't it be better than kaby lake?

>>58350317
Because AMD always managed to fuck up at least since Phenom days.

>>58350334
Times that AMD fucked intel out of nowhere:

Athlon
Athlon XP
Athlon 64
Athlon x2

>>58350356
And then they started fucking up. Repeatedly.

>>58350361
Phenom had a bug that didn't present itself until super late into the product cycle, phenom II was good but late, FX was a huge mistake

And they brought back the guy that fucked intel to design this processor, so...

>>58350382
He, praise Keller the certified shitwrecker. Also fuck Intel. Bunch of lazy cunts.

>>58350246
Once upon a time intel released a 1.4 ghz processor that was slower than their 933mhz processor and it cost significantly more and required ram that cost insane amounts of money, and required that you install ram in pairs.

They then switched sockets after about 6 months.

>>58345868
Tell me when they finally stop making mobile CPUs and create desktop CPUs that can fit in Laptops. If Nvidia can do it so can Intel.

>>58350317
Basic pattern recognition. Keller can only do so much with a broke-ass AMD.
Intel has more R&D funding than like 5 AMDs put together.

>>58350426
Intel R&D developed Optane. That is absolutely useless. Bravo Intel.

>>58350446
Intel also developed USB, which was also useless.
3d X-point has huge applications- as it's storage that's almost as fast as RAM itself.

>>58350426
The pattern you fail to recognize is Keller fucking up intel with a smaller team

>>58345907
Not if we inform people that at this point they can stick with a year old I7 and still achieve great results.

>>58350474
I'm not even going to waste time arguing with you.
When Zen comes out and it's exactly 10% worse than kaby lake you'll realise that I was right.

Note that it might not be exactly 10% worse, but it's going to be in that rough area.

>>58350496
And what if it's better?

>>58350496
>exactly 10% worse than kaby lake
>might not be exactly 10% worse

You mean you have to go to bed since you have have school in the morning. Goodnight kid, learn to not contradict yourself within 2 sentences next time

>>58350511
kek

If by better you mean 'more cost effective and overall a much better purchase', sure.

But if by better you mean 'better performance per core' then kek

>>58350536
Well it might be 9.9% worse and then you'd have grounds to say I was wrong.

>>58350547
Nigger you used the word exactly.

If you can't even understand the meaning of simple english words, why would you think that you can be an authority in the performance of CPUs?

>>58350537
They just demonstrated that with the 3.4ghz ryzen vs the 6900k

>>58350537
No, I am not talking about cost efficency. What if it's simply better? In every possible way?

>>58350566
And you're arguing with a complete lack of understanding of metaphor and hyperbole to avoid talking about Zen.

>>58350584
Pigs will be much harder to catch

>>58350578
Neither AMD, nvidia nor intel have ever exaggerated, cherry picked, or in any other way positively misrepresented their products in their marketing materials designed to promote their products before release.
Remember the 480? Better than the 980Ti!

>7700K

If you really wanted to go Kaby Lake you just get a 7600K and overclock it

>>58350280
Because AMD is there to disappoint.

>>58350861
They did fine last year. And yes, Keller usually delivers.

Ever since Haswell we only had shit

Igpwell
Skymeme
Baby Lake

>>58351221
>/g/ actually thinks diminishing returns = shit

Enjoy never being happy with a processor again.

tfw running OCed 6700 non-K to 4.5 Ghz

>>58351256
I'm very happy with my 14 core Xeon.

>>58350223
It draws the exact same. Maybe more because z200 boards default at 1.3+v.

post the 1% ipc gains diagram plox

>>58350043
Depends

Locked->unlocked sure
Unlocked->unlocked nah
i5k->i7k only if you know for sure you'll utilize the hyperthreading

Overclocks higher tho

File: 6174587.jpg (143KB, 855x1017px) Image search: [Google]

143KB, 855x1017px

>>58350417

File: -------.jpg (61KB, 640x480px) Image search: [Google]

61KB, 640x480px

2500k is still good enough for any task unless you need MOAR COARS in which case this is not the type of cpu you should be looking at.

And gaymuhrs think that skymeme and babby cake will improve their fps, it will only prevent slight FPS drops when your GPU is too slow or the game is poorly optimized, in which case you're likely fucked anyway. The whole point of a GPU is to prevent the CPU from going full retard on calculations because it will lead to poor performance.

Also, you're all cucks for buying into this shit.

>>58352268
>2500k is still good enough for any task
VT-d please.

>>58345808

My i7 920 still handles everything I throw at it, nine years after I bought it...

>>58352282
>VT-d
For what reason?

>>58352322
Hardware passthrough..? Specifically GPU.

>>58345868
>Wait for Skylake-E for that.
BWHAHAHA

INTEL IS NOW THE MORE COARS COMPANY

INTEL IS NOW THE WAITFAG COMPANY

HWHWHAHAHHAHAHAHA

>>58352331
>Hardware passthrough..? Specifically GPU.
Why not use Wine? Or do you want to play Cowadoody etc?
You may think i'm an asshole for suggesting this, but if you switch to lunix only to install a windows vm you've defeated the point of the exercise.
If it doesn't run in Wine, abandon it. If you have specific professional needs, get a workstation for, well you know work.

File: 1456591795524.png (289KB, 1200x600px) Image search: [Google]

289KB, 1200x600px

>>58345808
Its the result of a dead end arch reaching its conclusion. Anyone who thinks this is just Intel being lazy is delusional. They still have to get people who have older Intel products to buy new ones.

Increasing single core performance significantly at this point would take exponentially larger cores and it would make netburst look like a freezer. The future is more cores.

>>58352370
Wine is fucking shit. I've been full time Linux for a year and I think I've played games for less than a total of a week's hours because dual booting is a pain and Wine has way too many problems and way too little compatibility. Supposedly it has lots of compatibility but every time I go to do something there's problems. Games and software both.

I'd rather just clean everything up, drop Windows as a boot entirely and have a walled off high performance VM I can use _without shutting down my linux environment_ to play games or use windows only apps when I need to.

I mean sure I could just get a second PC altogether but in that case why not just put that second PC into one PC and do it this way, also it's more fun this way.

Saying not to do this because you desperately want people not to buy a new CPU (new workstation instead though lmao?) would be very dumb.

>>58352404
Don't expect Linux to ever be a competitive platform for games if you continue to give Microsoft your support.
I see why you would want a gaming VM for pure comfort, but i assume you left Windows for a reason. Why would you continue to use it, when you could support development and interest for a better alternative(regardless of the faults it may have in it's current state). Especially considering that now there is even a official Linux Steam client, and a vague hope that Vulkan isn't pure vapor. 2017 is the year of the Linux desktop(i promise).

I suggested getting a workstation because honestly if you have need for a professional solution you'd probably be better off having it separate from your private PC where you can dick around as much as you want. It's just good practice in my opinion.

Does kaby lake use a new socket?

>>58352475
>Don't expect Linux to ever be a competitive platform for games if you continue to give Microsoft your support.
I don't expect it to be period, the tiny amount of Linux users compared to Windows babbies for games will never shift. At one point I dreamt of Linux having it all but now I understand it's not to be and it's best to use the right tool for the right job. As it turns out the only job Windows is good for any more is playing games on.

I left Windows for a reason, but you ask why I'd continue using it like I'd actually be 'using' it at all. I'd be simply booting the VM and launching a game immediately, I wouldn't really interact with the OS any more than would be necessary to install the games I'd launch. Windows would become just a video game bubble, and I like that.

Problem with 2 machines also is unless you do some tricky KVM switch setup you'd have to move back and forth to use them. I like the idea of having a game (in windows) on one monitor but having my linux desktop right next to it, easily accessible in under 2 seconds if I want it. I could continue using my linux environment fully and completely neglect windows then.

>>58352494
No.

>>58352522
I think we've reached a point where i don't really disagree with you, and i don't really care to argue my own points enough to continue.

So i'll just call you a fag and go make a tuna sammich. Fag.

>>58352543
lmao, i love this place

>>58352574
Who doesn't.

>>58349173
>more bit depth is bad
Are you right in the head?

>>58347646
>PCIe 2.0
>95% as fast as 3.0
>outdated

PCIe 3.0 was always a meme.

>>58347829
Then why Even have enthusiast lineup if it wont be anymore special?

>>58352404
>dual booting is a pain
>goes on and on listing problems that would be solved by dual booting
>dual booting is a pain guys i swear

>>58353151
They want you to buy the older technology or watered down new technology so you have to upgrade again sooner, goy.

>>58348233
There is no hate on AMD senpai.
But amdrones are shilling everywhere and pooposting their biased benchmarks.
Thats the problem.

>>58350107
SiGe is used as insulation under the source and drain wells of ever 28nm mobile chip on the market. Its been used for a long time.

>>58353302
You're absolutely delusional

>>58345907
>tfw all local computer stores have sold out of 6600K and only have newer 6700K

I am going to be a fag enabler and buy one.

>No windows 7 support

Pass.

>>58352398
>increasing single core performance
>larger cores
Are you retarded

>>58345808
>So who is going to buy it?
People who want faster comptuers, with faster ram.
Dont give me any shit about arsbullshitnia...

>Can Intel actually afford to release a CPU they know damn well nobody with a brain is going to bother buying?
People with brains will be buying it,
people without brains will be buying it,
and people with shit for brains will not know the diffrence.

>Why not just release nothing?
Ram Speed increase.
Otherwise just go FUCK YOURSELF.
your arguments are invalid. >>58345808
>So who is going to buy it? Can Intel actually afford to release a CPU they know damn well nobody with a brain is going to bother buying? Why not just release nothing?

File: kek.png (74KB, 1070x491px) Image search: [Google]

74KB, 1070x491px

>>58345972

Vega 10 competes against Titan X Pascal.

File: gfgsdgsdgsdiphy.gif (356KB, 500x442px) Image search: [Google]

356KB, 500x442px

>>58352475
>>58352404
Artists are still fucked as well, not sure if anyone noticed but it's one of the first groups to go for if you want more games, which = more users.

And don't give me that gimp/blender is just as good crap, that's like telling you programmers that "you're wrong and should be using sublime text/atom instead of Emacs/Vim/Visual studio"
It's easy to have an opinion on replacements when you don't have experience.

>>58345808
>nobody with a brain is going to bother buying
That's the problem, brandcucks don't have a brain and Intel knows that.
Why would they waste money on retards who'll thankfully swallow their shit anyways?

>>58355138
Nothing will have win 7 support anymore.

>>58353061
Muted color paletes that the flim artist intended are more aesthetic

>>58345860

I think they meant price on release, I'm seeing skylake cheaper everywhere

>>58350246
Millennials are old enough to remember it. You also need to be at least 18 to post here.

>>58345808
>Intel brings desktop performance to a standstill.
To be fair, AMD did that years ago on their line. Without competition what is their motivation?

File: intel.png (234KB, 800x612px) Image search: [Google]

234KB, 800x612px

>>58345808

>>58353246
It is a pain, and there's a better solution than it so why the fuck would I do it.

>>58356118
>by 12%

>>58356118
Right. I forgot that everybody is stupid, everybody but me.

>>58347829
4 cores for X299? [Citation needed]!

>>58347644
>March

Jesus fuck I thought it was going to be released in January

>>58356686
The absence or presence of HDR has nothing to do with how muted colors are.
So either you think HDR = displays as set up in Best Buy, have no idea what muted colors are, or you're just plain stupid.

File: kabylake-bench-bf1.png (9KB, 831x440px) Image search: [Google]

9KB, 831x440px

Intel quality

>>58361049
why the fuck is Ivy Bridge so fucking high
and why is sandy vagina even still relevant, holy shit intel put your poo in the loo

>>58348233
Because each generation of APUs was actually noticeably better than the last. The improvements weren't HUGE, but still better than Intel's.

I'm not understanding what's being said about this supporting 4k video or whatever.

You don't need particular hardware to run video at particular resolutions to begin with provided you just have the prefomrance to play the file, no?

>>58361887
>You don't need particular hardware to run video at particular resolutions to begin with provided you just have the prefomrance to play the file, no?
There is a difference, it's software vs. hardware. But the cheaper solution is an RX 480 or any GPU that can output 4k.

>>58361887
Dedicated hardware video decoders just save power mostly. There are some instances where your CPU could simply be too weak to decode a given video without stuttering, but its mostly about saving power.

Your laptop might pull 20w when playing a 1080p video decoded by the CPU threads.
That same video decoded by a dedicated block can decrease power consumption to just a couple watts.

So it's not worth getting a 7700k over a 6700k just for the sake of preformance, but is it worth getting a Z270 board and a 7700k by extension over a z170 and a 6700k just for the 7270 presumbly sticking around and being supported longer?

I got a Asus z170 board and a 6700k over blak friday and I'm wondering if it's worth trying to return them to get these instead.

>>58361559
desu I remember when sandy bridge came out and /g/ was insisting not to buy it because

>DRM that will disable your pirated files

It sounds fucking retarded thinking back, but that got me on the red team

>>58361983
Its entirely a possibility, that fear wasn't unfounded. Windows has the ability backed into it.
It was originally envisioned that HDMI could detect pirated media and disable devices that played it. ARM teamed up with the MPAA and created a hardware ASIC for preventing mobile devices from playing pirated movies.

Pretty much the only thing stopping companies from enforcing it is the bad PR they'd get from the innocent bystanders caught in the crossfire.

>>58362019
Oh shit it was real? But I imagine the screenshots were faked

>>58362054
It was very real.
There were a couple things in Sandy Bridge that people freaked over. One of them was a suite of business oriented features, including some anti theft stuff, that could be used to remotely track and disable a computer if reported stolen. Basically the same thing all cellphones have in them now. Report a computer stolen, and as soon as it connects to the internet it can be killed remotely. Its easy to see why people would be suspect of this functional and its potential for exploitation. This still exists and has been expanded upon by a lot.

The other big one was some internet streaming DRM hardware. Intel teamed up with some movie studios who were planning to introduce sort of a Netflix distribution method. They were thinking of trying day one digital releases, but they wanted to ensure it wouldn't be an open door for pirates. So Sandy Bridge has something on die thats capable of preventing you from re encoding certain streamed media.
To my knowledge it was never used, but it exists.

There are a few other things too, but they weren't as high profile back in 2011~

>>58362149
The Anti theft technology required a subscription and was discontinued a few years ago.

>>58345808
>Every review out is called this thing a meme and the result of AMD's embarrassment and Intel genuinely not giving a shit anymore.

Has anybody considered that perhaps maybe Intel is having trouble making the chip perform significantly faster? Is there another processor out there that can actually match or beat theirs in terms of performance? Of any architecture?

It looks like they're more interested in energy-efficiency these days because they want a slice of that low-power and embedded pie that ARM is currently the king of. That's not necessarily a bad thing if you think about it, it means that laptop batteries should be lasting longer at the very least.

File: certified shit wrecker.jpg (13KB, 381x286px) Image search: [Google]

13KB, 381x286px

>>58362339
>Has anybody considered that perhaps maybe Intel is having trouble making the chip perform significantly faster? Is there another processor out there that can actually match or beat theirs in terms of performance? Of any architecture?
Hmm.

File: Halo Cal-141.jpg (76KB, 840x540px) Image search: [Google]

76KB, 840x540px

>>58361977
bumping this

>>58362355

I wouldn't be surprised to see AMD's new CPU's approach or equal Intel's current offerings.

I would be shocked if they significantly surpassed them, though. Intel is already cutting it incredibly close to Moore's Law, and AMD is playing from behind. I don't think the AMD has some unheard of ace up its sleeve that Intel hadn't thought to do.

>>58362509
Intel went the lazy shitcunts way so who knows. 8-core already clocks higher than anything Intel has to offer. Yet. Maybe Skylake-E will have some damage control edition SKU to compete.

>>58350399
It's what happens when companies open themselves to gender-equalitarianism and other SJW bullshit. Spending your money on unqualified women and faggy feelings doesn't engineer products.

>>58362675
/pol/ pls.

>>58361887
>>58361915
>>58361934
There's also the matter of supporting HDCP 2.2 which is required by almost all commercial 4K sources such as Ultra HD Blu-Ray and Netflix streaming. Keep in mind that many people these days don't have towers they could slap a discrete graphics card in if they were even comfortable with removing the side panel.

>>58362695
/lgbt/ pls.

File: pol.png (73KB, 1056x869px) Image search: [Google]

73KB, 1056x869px

>>58362710
Fuck off.

>>58362734
no

File: 12208254_1567018506943471_217479177006874016_n.jpg (40KB, 696x637px) Image search: [Google]

40KB, 696x637px

>>58362734
kys your self

>>58361977
>I got a Asus z170 board and a 6700k over blak friday and I'm wondering if it's worth trying to return them

Don't. You won't be getting Kaby and/or 270 boards for good prices anytime soon, and the performance difference is pretty much zero. No IPC change. Slightly better OC on Kaby, but again, close to zero real world impact. Just wait for the Skykek and be happy you got it at a good price.

File: 1458507925176.png (15KB, 455x440px) Image search: [Google]

15KB, 455x440px

hey guys how do you say Kaby? Kay-bee or Kah-bee? I know alot of people pronounce it as kay-bee but i like to say it as kah-bee :3

>>58362695
He's right though.

Pandering to minorities and women doesn't work. Hiring them doesn't work. If you want competence and competitiveness you hire white/asian men. If you want barely passable for cheap, you hire indian men.

If you want to drive your business into the ground and go bankrupt, you hire women and SJW's.

>>58363121
Please do not leave the house, a van is on route to your location to take you to a reeducation facility.

>>58360761
still no release date confirmed besides a hard q1

>>58356118
honestly, neither of these reviews are bad.

one wanted ddr4 one went from ivy i5 to 7700k

>>58360904
Alot of hdr content and displays fuck with the colors currently, to the point I can tell a difference between displays on a T240hd, I should not be able to, It would just be extra in between colors, possibly a noteable difference in contrast that you can see between one panel and another because of the camera, not night and day one is oversaturated shit and looks worse then the on paper objectively worse in every regard panel.

the difference between 10 bit and 8 bit is 3 extra colors in between, and with 8 bit, you are VERY hard pressed to see the difference between 180,140,200 and 180,140,201

The big thing is contrast, and realistically that is all that matters and ever will matter in being able to see if a panel is better or not with the naked eye.

tn not being accurate is something you get use to, more colors have niche uses (banding which is almost un noticeable outside of pausing a movie and looking for it) and are largely un noticeable, but the first time you see an oled display, it shows you the future, and the future is real blacks.

>>58362339
yea, everyone considers it for a bit, maybe they cant push the architecture better, maybe there isn't something better, but then we realize that a new architecture from scratch costs a few billion and come to the conclusion that intel doesn't want to make a better cpu, they want to milk what they have as long and hard as they can.

At least until they are challenged, and because no one here can tell us why they CAN'T get better performance, we should assume that intel is being cheap lazy cunts.

the reason why they went for power over performance is the one area they are challenged in is the server/enterprise market, where many times power comes before performance if the performance meets a requirement.

amd is potentially doing 8 channel memory, which will hit intel as the best they will do is 6 channel and that will bottleneck the people who need the throughput,
amd has an architecture that looks more power efficient, so intel will get hit again.
all intel has left is absolute peak performance where they will be the only choice for 20% of the server market.
And because of amds better gpu and seemingly better power efficient cpu, another spot intel will get hit on is mobile.
then to kick them one more time while intel has been sleeping at the wheel, the entirety of the desktop market is up for grabs because intel refused to give people more cores, cheaper cpus with a choice of no gpu, a more expensive platform, and year over year, very minimal improvements, fuck, even amd's cpus gave compelling improvements architecture to architecture, now that amd will give people an 8 core that could be i7 prices, it will be the kick to the balls intel needs as if they do make 8 core 16 thread 350~$ that makes the entire i3 i5 and i7 lineup pointless with their 6 and 8 core in do i need it, if amd put out a 4/8 for 150, that kills intel's low low end too, because amd gets the cheap as dirt motherboards.
All i can say for sure, things are going to get interesting.

File: ahmedinejad.jpg (22KB, 280x412px) Image search: [Google]

22KB, 280x412px

>>58350417
>They then switched sockets after about 6 months

>>58362734
Triggering: The Post

Is Kaby Lake the new Ivy Bridge with TIM issues?

>>58364757
http://www.pcgamesn.com/intel/intel-kaby-lake-tim
Sure seems like it.

>>58364757
It's worse. It's even funnier considering Zen is going to be soldered. In a good old fashioned way. I'm sure Intel can't solder consumer-grade CPUs because, I guess, reasons.

>>58365046
>I'm sure Intel can't solder consumer-grade CPUs because, I guess, reasons.
Im sure it involves money grubbing.

>>58365070
That's the point. Even when the nearest competitor is about to bitchslap them they are still trying to go full jew. I guess Krzanich is either too greedy or he's a really bad CEO with no common sense whatsoever.

>>58365104
I was gonna upgrade to a new Kaby Lake from my shitty A10 5800K, but the thermal issues have stopped me and now I have turned into an AMD waitfag.

kek has forsaken me

>>58365137
Well Ryzen is q1 2017. And mobos will be cheaper.

>>58365137
>implying Kek wouldn't bless AMD for the lulz
Meme magic is always with which ever victor would be funnier.

>>58365161
>Keller sticky
Fuck I want to see that.

I've heard that an AMD CEO has said that they have been focusing on APUs in the past

Which sounds retarded as fuck

>>58365186
What's wrong with APUs?

>>58365197
They are like obese people

>>58365250
But you are probably using an APU because Intel is too greedy for moar cores on desktop CPUs.

>>58365289
Aren't nearly all CPU's these days technically APU's?

>>58365289
Also because the majority of consumer computer use doesn't need a dedicated video card.

What's an APU? Please be gentle :(

>>58365306
Yes. Intel's z170 offerings are all APUs. They just don't call them that because their marketing department told them not to.

>>58365326
Integrated GPU and CPU. Simple as that.

>>58345808
There are some changes in the iGPU, I think it has better/more complete HEVC decode support and also supports whatever DRM faggotry Netflix needs in order to allow you to stream their 4K shit. So I guess if you're building a HTPC or something you might as well get a Kaby-something CPU.

>>58365336
Sounds shitty, is the performance any good?

>>58365353
Stick a Kaby Lake in your computer and tell me.

Keep in mind that gaymen comprises only a fraction of total consumer computing time. Office work doesn't require a dedicated GPU. Every second you spend on 4chan is a second spent doing something that can be done perfectly well on an APU. Of course you need the graphics card when you need it, but the point stands that the majority of even your home computer use doesn't consist of gaming. Laptops are mostly Facebook machines. APUs work perfectly well for office work, internet, and light gaming.

>>58365393
But iGPUs are absolutely pointless in k chips.

>>58352370
>WINE
even when it does work you're limited to Win XP era DX9 graphics.

>>58365306
My Xeon E5s aren't

>>58361983
kek, you poor soul. I almost bought a Phenom back then, but went with i5 sandy vag. It's still all good.
Can't wait to jump ship to AMD.

File: edgeville.png (74KB, 587x488px) Image search: [Google]

74KB, 587x488px

>>58365169
/g/ would explode.
I want that to happen.
I want to see this place burn.

>nobody with a brain will bother buying
Kaby Lake is very useful because of hwdec support for H265. High end desktops might be able to get away with software decoding but laptops will benefit immensely.

Is there any reason to get a 7700K over a 6700K? The price difference is negligible where I live and it seems like it's literally the same shit.

I'm not 100% sure I'll actually go for an upgrade though, my i5 4690 is still doing fine, will depend on how many CPU-whoring games I decide to dedicate time to.

>>58366905
fixes some bugs

>>58345808
just ran 7zip since its the quickest to reproduce and my 2600k with a pleb tier overclock (every setting in the bios on auto except turbo multiplier set to 45) averages the same or above the [email protected] in the benchmarks in OPs link

intel laughing all the way to the bank

So is Broadwell the only processor available right now that isnt a meme?

>>58365421
Not necessarily, for instance the vast majority of desktop-oriented distros won't even boot on my GTX 1080, because nouveau shits itself and I get no video output at all, not even a terminal. Integrated graphics are useful because I can actually install normally using that, add proprietary NVIDIA drivers and then have a working system. I'd need to take out the 1080s and replace them with some older card to do this otherwise, it would be much more of a PITA that way.

>>58367154
kek, Broadwell was the biggest meme. Intel was on full damage control all that time to finally release 14nm. Fastforward to now and they're going to be using that process for years to come where even GloFo is expecting to release 7nm before the next Intel shrink to 10nm.

>>58367272
or just own a old GPU, do all that shit, and stop crying like a bitch. i have a gtx 275 in a box and i dont even use linux

>>58367291
Congratulations

I'm getting the 7600K. Last week I was about to buy a 6700K, over the 6600K, cause of the clock speed. But yay 4.2Ghz boost out the box.

>>58367272
>buys 1080 for Linux
Are you retarded?
The only use for iGPUs are facilitating PCI passthrough.

>>58367372
>mfw wanting to upgrade Zen
>mfw no iGPU
Shit, I'm going to have two discreet GPUs. I think I'll just dual boot and disable SATA ports on the BIOS before booting.

>>58367372
>The only use for iGPUs are facilitating PCI passthrough.
You dont need iGPUs for PCI passthrough. Stop being a fag and run ESXi.

>>58350260
>I don't understand why these threads are necessary. It was known that kabylake was another 4790k in action.
i think you know why these threads are here

/g/ has more money than they know what to do with

>>58364234
you're an idiot

>>58362988
Hey-Peess

>>58362988
kay-bee

File: ryzenslueth.png (846KB, 1023x768px) Image search: [Google]

846KB, 1023x768px

>>58350202
unedited version

>>58350127
things are usually more expensive at launch. you'd rather have a 7700K than a 6700K, so you'd be willing to pay at least the same right now (or not buy either one of them right now). just wait a bit for the supply/demand to balance out

>>58367943
>you'd rather have a 7700K than a 6700K
no.

>>58368066
if you're building a new pc you'd probably prefer it assuming it's not more expensive

>>58367943
Why is Broadwell still more expensive?

>>58368066
>would rather buy the buggy new architecture than the refinement for the same price

Do you enjoy being a cuckold beta tester for intel that much?

>>58368099
afaik broadwell never really took off except for in some laptops. maybe they stopped production or haswell is cheaper to make or something

Rysen for win then ?

>>58368187
q1 2019

>not waiting for Coffee Lake
You're missing out mainstream six-cores.

>>58368218
Coffee Lake is going to be the meme of all memes.
Its going to be the 4th 14nm incarnation, and is actually a scaled up 10nm.

>>58368201
Saw a couple of rigs running games 4k at 60fps in ces yesterday and a lot of maufacturers announsed am4 motherboards too for rysen. Maybe q2 2017.

>>58345808
>So who is going to buy it?
People like me who haven't done an upgrade in 4+ years.

>>58368155
their fabs fucked up, they wanted to do same thing as kaby doing now but couldn't yields were too low

>>58368323
I haven't done upgrade since june 2011, I don't want any mainstream intel CPU even for free.

File: 1475158491018.jpg (68KB, 448x529px) Image search: [Google]

68KB, 448x529px

>tfw you can completely skip this technological generation for an arbitrary reason like "I don't like the name Kaby Lake" because your current processor is just as good

>>58368260
AMD has said they plan to have it out before the end of Q1 2017

3770k @ 4.5ghz

Still no reason to upgrade.
Those newer asus mobos are enticing though

>>58352582
Me. It's degenerate

>>58368996
Coffee Lake is even shittier name and Cannonlake will be released in late 2018.

>>58349234
Actually the highest end i3 is a 7700k just with two cores. For gaymers that's literally the perfect CPU.

>>58370003
For gamers the Pentium G4560 is more than enough. It's better to invest more into the graphics card.

>>58370032
but what if i have money to burn

>>58370032
>For gamers the Pentium G4560 is more than enough
are you intentionally spreading false information?

>>58353246
>dual boot
It's a pain in the ass if you want to hibernate both OSs.
Mounting the filesystem of a hibernated OS from the currently running OS will kill the data on the filesystem.

It's not their fault you sperg. The transistor is much harder to make smaller as you reach 10 or 5ns. Sure, they would make enormous chips of more cores and whatnot but that's not progress, that's just more multiprocessing, and it doesn't usually increase single-thread performance.

>>58352522
>Problem with 2 machines also is unless you do some tricky KVM switch setup you'd have to move back and forth to use them. I like the idea of having a game (in windows) on one monitor but having my linux desktop right next to it, easily accessible in under 2 seconds if I want it. I could continue using my linux environment fully and completely neglect windows then.

VNC if shit happens and steam streaming for gaymes.

File: 1.jpg (44KB, 600x400px) Image search: [Google]

44KB, 600x400px

> mfw my i7-4790k will be good for the next 20 years

> asus claiming it's "reasonable" to overclock to 5ghz with 1.35volts on that transistor process

asus prove once more they are crooks.

they will kill your chip within a year.

>>58370032
Not this meme again.

>>58370204

It was a good meme, but it had of course great caveats, the game had to be largely made of spergs that can only make it single-threaded, e.g. WoWarcraft.

>Having hope for deepwater horyzen

That shit is competing with what is effectively a 5ghz intel quadcore that is guaranteed to dominate this year. They haven't had a worthwhile CPU since Athlon64. They're what you buy if you are impoverished. They don't have any aspirational or top-of-the-range feeling products. Same story with the vega shit, nvidia will just drop 1080Ti and then that will be the one everyone wants.
That said I have not bought a new PC since 2010 since this is a stagnant garbage industry and it would not improve my life one bit.

File: IMG_0089.jpg (17KB, 469x480px) Image search: [Google]

17KB, 469x480px

>>58370303
True. Sad.

>>58370303
Bait

>>58370396
>>58370303
Samefag phoneposter

>>58370160

I do wonder why Intel's design can't take the abuse the 32nm one piledriver is built upon given the huge advantage in just about every way Intel has for the physical silicon. If there is one design that can be tortured is piledriver. 5ghz 1.5v across 8 cores is no joke by any metric and igorning the required power draw piledriver will run at that 24/7 without flinching.

>>58347176
>so many boobs on the internet

I c wut u did thar

>ayymd razen barely overclocks to 3.9 on water
>7700k does 4.5 stock

HAHAHAHAHAHAHAHAHA

>OCing a 6700K produces the same result

... but can it take on a OC'ed 7700K? No?
And that's why it will sell.

>>58370663
Yes. They are identical. And TIM is even worse.

>>58370663
6700k already destroys zen. It's not even worth comparing it to 7700k

>>58370685
7700k did 7ghz on ln2. 6700k only could do 6.3

File: good_old_goyim.png (445KB, 617x515px) Image search: [Google]

445KB, 617x515px

>>58347738
>compairing prices from different time frames
Dumbass.

>>58370711
Who cares about ln2 OCing. Faildozer did even more.

File: IMG_1090.jpg (19KB, 250x241px) Image search: [Google]

19KB, 250x241px

>>58370723
With all cores disabled

Krzanich please at least solder your shills.

>>58364234
>the difference between 10 bit and 8 bit is 3 extra colors in between
I cant even react to this. Just stop please

>>58370303

Only spergs equate Intel to NVIDIA. NVIDIA is the same beggar as AMD. They don't have their own transistor process.

The sad truth is that Intel is the only choice on CPUs, but on GPUs both choices are mediocre.

NVIDIA gives the delusion they are good is because they overdesign and therefore overprice.

>>58370504

It's nothing secret. We simply reach the limits of Quantum Mechanics. Electrons leak out of the transistors.

It's the main reason why for the last ten years the advances in CPUs are minuscule.

In the nineties or before one single year gave almost +100% speed, now it's ~4%.

>>58371014
What's next then? 3D stacking?

>>58371014

Also it's why the only way they can impress people is by multicoring. The single thread performance (which still matters the most) remains almost stagnant and just add cores. It's the main reason GPUs give the delusion of progress because they use almost the same transistor technology but they give more transistors (because those work in parallel by definition) but the catch is they become more and more expensive, and we reach ludicrous levels of NVIDIA asking an arm and a leg for a single gaming card when they have nothing better than AMD technologically, just more transistors making it more expensive.

>>58371052

There is no much left on the silicon transistor. Whatever they do it leaks because quantum mechanics is real (it will be even worse when they approach 10nm and smaller) so their usual scam is to create multicore processors or overdesing the GPUs(they are lucky on those because they are by definition multiprocessing units, but the catch is they become more expensive the larger they are).

The real advances may come from completely different technologies, e.g. biological computers or graphene transistors, but those have a chance to be worse than anticipated.

A realistic future would be to remain on this technological state, more foundries to get 14nm and lower, and therefore the prices to drop, but that would still rely on multicores.

File: BASICS_figure.gif (68KB, 570x493px) Image search: [Google]

68KB, 570x493px

>>58370504
GloFo's 32nm process that Zambezi and Vishera are built on is IBM's 32nm PD-SOI. IBM designed it for their POWER7+ line which hit nominal frequencies over 4ghz.
SOI based devices aren't like bulk devices because they're inherently much more insulated. Their voltage scaling characteristics are vastly different.

The process was also designed to handle high levels of leakage, so the 5ghz Piledriver chips aren't doing anything out of the ordinary for the node.

>>58371014
Stop regurgitating things you don't understand.

Process is not limiting architecture.
Short channel effect is not part of quantum mechanics any more than water going through a sieve is.

>>58371084
Wouldn't Intel bring new, noteworthy performance improvements with their new x86 uArch in development?

>>58370977

this

it was ogre before it even started, intel immediately patented the transistor processes and everything so AMD and Nvidia either have to make something completely and utterly new or stay second for a long time

>>58371102
>Stop regurgitating things you don't understand.

If you are a sperg, don't show it. Quantum tunneling is a main problem of modern transistors.

>>58371102
Rude.

>>58371124

They just don't have their own process. They are clients of foundries like TSMC or Global Foundries or whoever they pay nowadays, even Intel. Intel has their own foundries and the best in the world, on a mass scale.

Maybe Samsung I heard has good things lately, but they mainly make memory I think.

A lot of stagnation or slow down comes from their aggressive copyright barriers.

File: 28-14.jpg (87KB, 1064x714px) Image search: [Google]

87KB, 1064x714px

>>58371134
> Quantum tunneling is a main problem of modern transistors.
Again, stop regurgitating things you don't understand. Normal leakage current is not some aberrant poorly understood principle of quantum mechanics. Current FinFET devices have less current leakage than any FETs we've ever used in silicon up to this point. You're flat out talking out of your ass, saying things you do not even understand in the slightest.

Leakage is not a limiting factor of FinFET devices today.
Magical electron tunneling isn't an issue.
If you understood anything about transistors at all, you'd understand that a FinFET inherently is a QWFET, capable of electrical insulation on all sides of the channel region.

Stay on reddit if you want to meme about popsci buzzwords with other 80 IQ subhumans.


>>58371171
Samsung has a handful of fab facilities. They produce NAND, DRAM, various micro controllers, they do advanced packaging, and they make their own ARM SoCs of course. Outside of serving their own needs they produce parts for other vendors. Various chink ARM parts, some FPGAs, Qualcomm's latest line of SoCs is being fabbed by Samsung.

>>58371221
>Stay on reddit if you want to meme about popsci buzzwords with other 80 IQ subhumans.

This sperg things that 4chan will give him a nobel prize instead.

>>58361049
BF1 horribly optimized game

>>58371221
>Leakage is not a limiting factor of FinFET devices today.

>today

>Today

>TODAY

For anyone that didn't get what this sperg is doing, is that his whole argument is that "we found a way to slow it down for now". Yes sperg, of course Intel and others are into slowing it down, but it can't go forever.

Next thing you'll tell us that quantum mechanics is a buzzword.

>>58371221
Impolite smartass.

>>58371321
No. You're still talking out of your ass, incapable of doing anything but regurgitating buzzwords.
Leakage does not break a planar device. There was no aberrant electron tunnel that was shutting down devices, stopping a transistor from functioning. Sub 28nm short channel effect was so severe that leakage current was approaching equal levels with actual current used to power the device. Still, this doesn't interrupt nominal switching performance.

Leakage is not harmful electron tunneling. You do not know what these words mean. You're an 80 IQ moron mindlessly regurgitating popsci buzzwords.

FinFETs, ALL FinFETs, intel 22nm Trigate, their 14nm Trigate, TSMC 16nm FF, Samsung 14nm LPP, these have radically less leakage at higher power than any prior planar processes.


Just stop regurgitating things you don't understand. Stop pretending like you know what you're talking about.

>>58371387
>Stop pretending like you know what you're talking about.

But sperg, I know what I don't know and I look at the big picture, it is you that have the delusion you need a nobel prize.

>>58371410
>still just talking out of his ass
Stop fallaciously conflating short channel effect with electron tunneling. Stop regurgitating "quantum mechanics" when you have no idea what it even means. Stop presenting yourself as an expert to other people when you're blindly talking out of your ass.

Spreading ignorance and misinformation is the act of pure scum. Either you're an autist desperate for attention, or you're a pathological liar.

>>58371437

> I'm an expert

You are a delusional sperg that learned a couple of things anyone with access to wikipedia can get. If you had any perspective of the big picture here you would know the transistor is much harder to shrink for the last 10 to 20 years and it's the main reason chips are incredibly harder to be made faster per single thread per year.

So you can pretend we owe you a nobel prize all you want. The fact of the matter is that I was perfectly right. The transistor is harder to be made faster and the main scam that makes people thing there are big advances is multicores and multithreading.

>>58371437
Could you tell me where Zen is worse compared to Kabylake from the information we know so far? I want every detail.

>>58371474
Now you're pathetically shifting goal posts because you know you were in the wrong, but your nature as a lying piece of shit won't allow you to stop lying.
Area scaling becoming more difficult, more expensive, is not because of electron tunneling. Like all other inherent electrostatic hurdles, transistor topology addresses it. The issue with smaller processes is the limited wavelength of light being incapable of readily forming structures through lith masks. This requires double, triple, and even quad patterning. That increases complexity exponentially which drives down yields. Additionally in bulk devices structures themselves become more complex which requires more critical mask layers, but only a few companies have an amassed wealth of SOI based IP.

>>58371490
There is a chart of instruction throughput and latency floating around from some guys on twitter who benched their test chip.
http://www.overclock.net/t/1619110/cpc-first-unofficial-ryzen-benchmarks#post_25730086

Zen doesn't have the physical FPU resources that the latest iterations of intel's arch does. intel hand process native 256bit ops, the Zen core can do 128bit ops. There are tons of nuances and implications of this, but those larger FPU ops are actually still uncommon. The caveat is that in some instances a larger FPU may let you execute numerous smaller ops concurrently, so you end up with the throughput advantage with say a 64bit op. This type of thing tends to show in some synthetic benches.

Over all though Zen is incredibly close to Broadwell, and Skylake on average is less than 3% faster than Broadwell per clock.

>>58371267
No.

>>58371575
Would you say a 6 core Zen would be better or just as good as an Intel Core i5 7500K?

>>58361559
>why is sandy vagina

Because every bridge is iteration on Sandy, their acrh did improve but marginally.

>>58371602
Ultimate performance will come down to what clocks you can hit on Ryzen when overclocking.
The i5 7600k has a base clock of 3.8ghz, and peak turbo is 4.2ghz.
Kaby Lake has a slight IPC advantage, whatever practical benefits come from the beefier FPU, and it has a clock speed advantage over the fastest stock Ryzen we've seen so far.

6c/12 Ryzen should still win out in most heavily threaded benches though, providing its clocked in somewhat the same range.

>>58371636
Thank you!

>>58345808
I wouldn't be surprised to learn Kaby Lake is a literal rebadge with support for Optane and 4k streaming DRM enabled. If 'Optimization' doesn't mean better IPC, better thermals, better power profile or cheaper to make then what the fuck does it mean?

>>58345931
Intel isn't feeling any pressure to advance their processors at all, silicon or otherwise. AMD is so far behind there's no need for them to work hard.

Perhaps if Ryzen turns out to be actually something, Intel will start trying again. But I doubt Ryzen can live up to the hype.

>>58371680
It means buy new CPU, goyim.

>>58371575
>http://www.overclock.net/t/1619110/cpc-first-unofficial-ryzen-benchmarks#post_25730086
>dat lack of pipelining on the complex units

AMD has always kind of done that, even back in the A64/P4 era they were strong on legacy x87 math but not so hot on SSE2.

Intel's a bit more forward-thinking, but AVX is poorly designed enough that I don't think it will matter much since it's so awkward to fully use it and nobody bothers (the fact that Pentiums and Celerons still don't support it doesn't help)

>>58371680
What are your doubts on Zen?

>>58371712
It's an AMD arch. And you know how faildozer ended.

>>58371490
>>58371602
>>58371662

Let me guess, you want to buy a new processor and you are doubting between burning the money now for kaby or wait for zen.

What happens to the people who bought Broadwell 6800k?

>>58371712
Name one thing ever that lived up to the hype.

>>58371770
Athlon, Phenom.

>>58371770
Donald Trump

>>58361049
>Keskimääräinen ruudunpäivitysnopeus
My new password

>>58371706
High level arch never shows the details and intricacies at the low level, but I do the feeling that AMD tried for elegant simplicity. It makes sense considering their financial situation, wanting to make the most with the least available resources.
They've got 4 generations of upgrades to tweak things and over complicate it.
Zen is only the starting point after all.

>>58371662
Any time.

>>58371878
Thanks
t:NSA

>>58367372
I'm not retarded, but you might be. I didn't buy a 1080 "for Linux" but I dual-boot.

>>58371878
Mine is أَشْهََدُ أَنَّ مُحَمَّدًا رُسُولُ ٱلله!

I have a neighbour whose WiFi username is Keskimääräinen.

Are we perhaps neighbours?

>>58371575
>you know you were in the wrong

Any delusional narcissist sperg would claim that. I explicitly told you I look at the big picture which obviously implies I do not everything. Only a narcissistic child like yourself could ever pretend that he knows everything on the subject.

You have to be literally autistic to not see that the transistor is harder to shrink and the chips harder to make faster for the last 10 years especially.

Whatever buzzwords you heard in wikipedia are not going to change the fact that multiple conglomerates competing on it still find it hard.

>>58369176
>3770k @ 4.5ghz
>Still no reason to upgrade.
>Those newer asus mobos are enticing though

What voltage? I'm at 4.2 on 1.19v, I did 4.4 on 1.21

>>58372531
Repeating the same fallacious nonsense and continuing to shift goal posts is just highlighting that you're a pathological liar. You were talking out of your ass. I called you out. Your autism won't let you stop shitposting.
Stay on reddit.

>>58372633

Sperg, I repeatedly called out your narcissistic tendency to pretend that we owe you a nobel prize. Completely out of nowhere you started spewing all those wikipedia-sourced buzzwords to pretend everyone else around you wrong.

Guess what moron, nobody cares for the exact reason the processor is harder to be made faster. But that's my entire argument. That it was harder to be made faster.

So kindly fuck off because I never questioned the fact I do not know everything. I clearly claimed that it's just harder to make faster single-threaded chips.

>>58372724
>hey guys I totally know what I'm talking about, its this quantum mechanic buzzword I read about on reddit
>STOP MAKING ME LOOK STUPID
>I-I-I'M STILL RIGHT

Just stop it.

>>58372747

What's most ironic is that quantum tunneling problems in modern transistor making brings up hundreds of thousands of google results.

So even if the sperg is right that photolithography is mainly to blame, it doesn't negate the fact quantum tunneling is big factor.

And in any case nobody gives a shit in /g/ about the exact inner working but that simply all foundries find it hard to shrink.

There will be ANOTHER generation of 14nm after Kaby Lake, too, called Coffee Lake.

The worst thing is that right now I need a very powerful CPU more than ever. 240hz monitors are out and the CPU is universally the bottleneck at that FPS.

I will have to buy a high-bin siliconlottery 7700k to have any chance of running many games at 240 FPS.

>>58372822
>I need a very powerful CPU more than ever.

That's a hard search nowadays. Almost any CPU after the Sandy Bridge era is only a minuscule advantage over the last. e.g. if you have 4970k today, it would be practically like throwing money out the window.

It's not for lack of trying, or because AMD doesn't give competition, that's total nonsense. It's just that the transistor is much harder to be made faster. e.g. because of harder photolithography.

The most cost effective thing you can do is to get any high end i7 of the basic desktop variety. If the software is single threaded little will change with the usual modest overclocks anyway.

>>58372822
What's the point of that refresh rate when LCDs have garbage persistence that will never be overcome. I have a 120hz samsung TN, a 90hz korean meme IPS, and a flat AOC CRT and even at only 75hz the CRT looks faultless. It's like looking into another world by comparison to the others.

>>58372822

That's because those games use largely only a single core. Single cores aren't that fast nowadays compared to 10 years ago. I wouldn't blame the software developers that much either since there are inherent bottlenecks in multithreaded programming, especially when the software is interactive, e.g. in gaming.

Multicores seem impressive mainly on encoding, decoding and non-interactive benchmarks that nobody gives a shit about in practical computing.

File: 1452717560680.jpg (35KB, 593x452px) Image search: [Google]

35KB, 593x452px

>>58373008

>>58373016

Have you ever tried to program a multithreaded application? If you can split it in parts like in offline encoding you can almost double the performance with two cores. When you have to deal with sharing of variables on an interactive application like a game, not only not double, you may even make it slower if you are not careful. The main reason computing in general is much harder to be made faster is that single threaded performance advances are simply shit.

>>58347282
>4k Netflix streaming

It is drm it will get cracked like every other drm

>Pentiums with Hyper Threading

Nobody likes pentiums

>Unlocked i3

Yeah let me overclock the facebook machines,

>>DDR4-2400 support on the B250 and H270 mainstream chipsets

Good

>>Bigger OC headroom

Good

>>58347282
>DDR4-2400 support on the B250 and H270 mainstream chipsets

doesn't every skylake i5 and i7 already support well above 2400mhz ram

>>58345808
getting locked into windows 10 when you can buy a cheaper product that can perform the same and run 7

>>58373016
>muh cores

https://en.wikipedia.org/wiki/Amdahl's_law

>So who is going to buy it?

People who want to upgrade from lesser CPUs like the 2500K or shitty CPUs like the FX8350?

It's not like the 7700k is more expensive than the 6700k or the 4790K

>>58373008
Meh.

Games are simulations, and those are usually pretty easy to thread. The bottleneck these days are shit tier APIs (DX11 & OpenGL). If we had Vulkan/Dx12 api as a dominant one 4 years ago, then half of the world would run 8 core amds because MOAR COARS would make sense.

However if 70% of your time is spent in molesting a serial API, then threading the rest makes no sense. Maybe large simulations, like Supreme Commader would scale nicely with threads, but your every day TPP shooter with velcro walls does not.