HAHAHAHAHAAHAHAHA

>>57923703
>engineering sample @ z170

:^)

File: KEK124.png (16KB, 625x347px) Image search: [Google]

16KB, 625x347px

>>57923709

>>57923703
>>57923709
just wait for zen. amd will impress the whole world when they release cpus that are even worse than their FX line.

>>57923703
>Four percent lower power draw
IT'S OVER ARM IS FINISHED INTEL CONFIRMED FOR KING OF EMBEDDED

>>57923703
BUT MUH 4K HEVC

File: KEK123.png (15KB, 580x341px) Image search: [Google]

15KB, 580x341px

>>57923728

File: 1477747416594.jpg (24KB, 496x499px) Image search: [Google]

24KB, 496x499px

>>57923703
im glad you're getting it out of your system poorfag OP

it's alright, we're here watching

>>57923703
Not sure why this is a shock, same lithography and same clock speed with no real change to the chip outside of the encoding / decoding hardware and the IGP, which no one gives a shit about on a desktop.

That's great news to be honest, I was almost regretting buying a Skylake on sale during Black Friday with Kaby so close but now I'm glad I pulled the trigger.

>>57923703
Its a line refresh, is not supposed to be good.

>>57924127

Same here. I bought a 6700k on black friday, was worried I would regret it, I don't at all.

Goy the whole point of kaby is being able to run at higher clocks. It defeats the purpose to run it at the same clock as skylake unless your goal is to prove that there are no IPC improvements, which there aren't. Right?

>>57924127
>>57924155
I also bought a 6700k during microcenter's black friday sale

>>57923703
>implying these wattage readers are that accurate
>implying it couldn't easily be any other part of the system that's drawing this much power

Its literally shit. Still no reason to upgrade my 2600k. The server chipset in launching ~2Q2017 looks like it might be pretty good though. Gonna be a ripoff but the first real upgrade in a long time.

>>57923742
They're probably measuring with a killawatt which is +/- 5w iirc

>>57924178

Kaby Lake is also going to have a much better chipset.

No reason to upgrade my FX-8350 in 2016. Nothing to see here.

>>57923703
>skylake has better performance
>skylake uses more power
really makes you think

>>57923703
Probably just more efficient for laptops like before with skylake

>>57924155
>>57924184
did the same thing

god bless microcenter

they take back items no questions asked too

>>57924350
>Still no reason to upgrade my 2600k

Nooo! A socket change a year was supposed to keep the goyim in fear!

>>57923729
I agree with you my brother; Intel cock tastes better

>>57925749
Kek further, there was no reason to buy it in 2012 when it came out

It was already underperforming Intel single core performance by a factor of 2.

>>57923742
4% lower system wide.

File: 1480023622195.png (9KB, 493x317px) Image search: [Google]

9KB, 493x317px

>>57923703
That's not how you make charts.

>>57927226
Really makes you think

Intel can't defy the laws of physics so we're stuck with meager performance gains until graphene replaces silicon

>>57923703
>same socket
>same lithography
>consumes less power
>produces less heat
>overclocks higher
>new hardware encoding and decoding for media
>no one is forcing you to buy it if you don't want to
>you don't have to upgrade as often now

I legitimately fail to see any issues here.

>still waiting for a cpu that makes a big leap like the second gen "i"s did to upgrade
>Migth never upgrate my overclocked 2500k

>>57927707
Pretty sure Sandy bridge to skylake would be a pretty big jump.

File: 1481143742040.jpg (36KB, 400x460px) Image search: [Google]

36KB, 400x460px

>>57927696
>no one is forcing you to buy it if you don't want to
>you don't have to upgrade as often now

wow thanks intel

File: 1481241034893.png (798KB, 996x868px) Image search: [Google]

798KB, 996x868px

>>57923709
>that awkward moment when Intel stops improving their processor by 0.1% every two years and actually starts going backwards

By the time they reach cannonlake they'll be 10% slower than skylake

>>57923709
>>57927801
All that graph is showing is that Metro is not bottlenecked by the CPU in that benchmark

fuck you, poorfag. just bought my lappy with a sweet Kaby cpu (shit is cash) and it's freaking amazeballs.

File: 1479787347437.jpg (37KB, 934x697px) Image search: [Google]

37KB, 934x697px

>>57923703

File: 1310429820535.jpg (13KB, 300x300px) Image search: [Google]

13KB, 300x300px

>tfw 2600k at 4.4 ghz

File: intel inside of youuuuu.png (41KB, 493x317px) Image search: [Google]

41KB, 493x317px

>>57927226
what did they mean by this?

>>57923728
This is a terrible benchmark because the entire critical path is hand-coded SSE/AVX. The performance is basically only going to be dependent on the clock rate and nothing else.

(Maybe RAM speed and cache size to some degree)

>>57927805
>>57927849
Btw, you're not going to get much success out of benchmarking Kaby Lake vs Sky Lake because they have the exact same IPC, architecture and feature set. There's going to be zero difference.

Here's a list of things changed in Kaby Lake, copied from wikipedo:
>native USB 3.1 Generation 2 (10 Gbit/s) support
>Increased clock speeds across all CPUs models (increased by up to 300 MHz)
>Faster clock speed changes
>Improved graphics core
>200 series chipset (Union Point) on socket 1151
>Up to 16 PCI Express 3.0 lanes from the CPU, 24 PCI Express 3.0 lanes from PCH
>Support for Intel Optane Technology

tl;dr the CPU performance at the same clock is going to be same, you just gain support for new tech and slightly higher clock speeds

>>57923703
>>57923709
>>57923728
>>57923763
>comparing an overclocked skylake chip against a stock kaby lake chip
all of my why

I still use an i5-3750. There's no point upgrading is there?

>>57927882
Because like every Skylake can hit 5GHz like it's nothing.

>>57928160
And this should not be the case for kaby lake because...?

KL has higher base clocks than SL

File: images.duckduckgo.com (72).jpg (44KB, 404x405px) Image search: [Google]

44KB, 404x405px

>>57923703

HAHAHAHAHHA

>tfw waiting for AMD Zen, which will be more energy saving and performing better

meanwhile, Intel releases """"""new"""""""""" processor every 6 months that perform 2% better and consume 20% more power (and forces intelshitters to buy new motherboards and ram)

>inb4 Intel charges $50 for unlocking power savings mode on Kaby Lake (they did actually charge $50 USD to remove software block so that some intel pentium processors could become overclockable, look it up)

>>57928198
>meanwhile, Intel releases """"""new"""""""""" processor every 6 months that perform 2% better and consume 20% more power
>>57923703

>>57928182
>And this should not be the case for kaby lake because...?
Oh great. Thanks for releasing a processor with the same IPC, the same power consumption, and the same overclocking abilities. gr8 shit, will gladly pay $400 for it.

I seriously don't know how people on /g/ are fine with this. It's like they want to get assfucked by AMD.

>>57928220
>I seriously don't know how people on /g/ are fine with this.
I honestly don't see the problem

They release a new product with new features and improvements (>>57927875) and charge a price for it.

Your old CPU didn't stop working. This might be news to you, but you don't HAVE to upgrade every single generation, anon. Upgrade when your current setup is broken or too slow/old.

>>57925971
I dunno, theres roughly 30% improvement since your 2600k. Even if the year on year isn't impressive the gradual improvement is pretty large.

File: babbylake power.jpg (38KB, 1400x317px) Image search: [Google]

38KB, 1400x317px

>>57927226

>>57927801
>slowly decrease performance by 1-2% ever year
>in 10 years our old CPU's will be dead so we'll have "upgrade" to slower ones
>rerelease skylake under a new name with 20% performance increase

when you stand in front of a wall, to go forward you have to go backwards first.

>>57928307
>when you stand in front of a wall, to go forward you have to go backwards first.
yes, which means NEW TECH not sinking money into the sinking silicon lithography ship

>>57928328
Coming up with something else may cost even more than EUVL.

File: 172836whs1560j1aeeh0t1.png (42KB, 428x318px) Image search: [Google]

42KB, 428x318px

>>57928328
There is nowhere left to go but MOAR CORES. Making it wider would mean exponential core complexity.
Higher clocks would mean exponentially higher power consumption.
Not to mention x86 limitations.

>>57928343
>may
you mean *will*. That's sort of the point, lol

The more money you keep sinking into silicon, the bigger the gap will be when moore's law prevents you from pursuing that path further. At that point, you'll make a far, far greater loss than had you simply researched alternative technologies earlier when the difference was not as big.

Silicon is a dying industry that's being kept alive via IV drips and artificial lungs, and the companies just don't want to realize it yet until the R&D costs come crashing down upon them as customers stop buying new CPUs.

We're pretty close to the limit of moore's law - and I don't mean this technologically, I mean it economically. Look at the evolution over time of the number of companies that can afford to compete with the state of the art in semiconductor manufacturing. That number is dwindling quickly over time, because the smaller companies can no longer pay the insane R&D costs required, shrivel up and die (or enter alterantive markets).

Currently that number is at 3. (It's basically just Intel, TSMC and Samsung). Pretty soon it will be 2. Then 1. Then 0.

Exponential investment cost is a bitch, and due to the finite resources on planet earth we can't afford to keep inflating it.

Silicon *will* die, and we will be left in a dark age of computing where no progress can be made until alternative technologies catch up, which will be extremely slow due to the lack of competitive market drive on products that will initially be significantly less powerful than our current CPUs. Brace yourselves.

>>57923703
In case you didn't notice, progress is dead. All you get now is more botnet and removed features.

>>57928416
You do realize that the main problem with "moar cores" is synchronization? You do realize that parallel buses died exaclty because of this, because increasing speed on parallel data lines beyond a certain point stopped making sense as performance gains were nullified by synchronization delays on the other side?

>>57928441
The way forward is specialization and parallelism.

>>57928470
This. Multicore utilization and parallel processing are still a problem after well over a decade. If it were the way to go, we would have CPUs with thousands of cores by now.

>>57928472
>parallelism

Yea, that's why disks are all on PATA and why SATA died, kek.

>>57928472
No, it really isn't.

Specialization just buys you a few orders of magnitude. CPU vs GPU is about 2 orders of magnitude, and GPU vs ASIC is about 2 orders of magnitude as well.

It will buy you time, and the companies desperate for revenue will pursue it because of that. But it's just digging their grave deeper by delaying the inevitable even further.

Parallelism has been a meme for decades. Amdahl's law is a bitch, as are hard data dependencies and communication latencies. Not every problem is a monoid.

Now what might be interesting is the death of the von neumann model. 3D-printed biological computers, wetware, etc. could also take us significantly beyond the current capability of computation, even well before we begin to hit thermodynamic limits

>>57923703
>>57923709
>>57923728
>>57923763
Did anyone seriously expect Kaby Lake to be anything more than a new Skylake stepping?

>>57928497
No, it's just an AMD psy-op tactic to get people “disappointed” about kaby lake and “hyped” for zen

>>57928472
>parallelism
bulldozer was too early to this party, at least amd have this tech refined with excavator

>>57928497
people expected 5% more performance

File: no-country-for-old-men-tommy-lee-jones.jpg (99KB, 500x332px) Image search: [Google]

99KB, 500x332px

>>57928198
>tfw waiting for AMD Zen, which will be more energy saving and performing better

Just like you waited for the 480 to be better performing than the 1070?

AMD Zen will not beat skylake, let alone kaby lake.

It will make up for a big chunk of what they were behind on, but they will still be behind.

AMD claims 40% IPC gains, they would need closer to 60/70 % to match Skylake.

And that's not even counting the fact that they probably won't be able to match the clock speeds, as most leaks so far have shown.

Zen will be good, and it will bring much needed competition, but if you think it has any chance of beating kabylake you're delusional.

>>57928569
>480 to be better performing than the 1070?

Nobody except currytech said that the 480 was going to be a 1080 killer or something. Even AMD themselves openly stated that the 480 was aimed at a different segment of the market.

>zen and skylake

Zen is competing with broadwell-e, so it's very likely that zen will outperform skylake in multi core performance. I really can't say much about single core but considering how has/broadwell-e chips match their non-e counterparts in single core with moderate overclocking, single core perf might not be a problem if zen turns out to be a good overclocker

>>57928813
>Nobody except currytech said that the 480 was going to be a 1080 killer or something
Except you're wrong, because I was here and I saw the daily threads up until the launch, the expectation by /g/ fanboys was 1070/980ti performance.

>Zen is competing with broadwell-e, so it's very likely that zen will outperform skylake in multi core performance

That's based on one cherry picked/best case scenario benchmark by AMD, with a highly underclocked Broadwell-E chip. If you think that's going to be representative of actualy performance in typical scenarios you are mistaken

>>57928482

>CPUs with thousands of cores by now

They are called GPUs. Failing that the xeon phi.

>>57928910
>the expectation by /g/ fanboys was 1070/980ti performance.
whats the actual performance of 480 referance and xfx 480 compared to those

>>57925952
fuck yeah they do, I stopped using sites like newegg entirely because there's a microcenter near me

>>57929120
There's a significant gap between a 1070 and an RX 480.