7740k might even hit 6ghz on water since it has no integrated

https://www.nordichardware.se/nyheter/intel-core-i7-7740k-kaby-lake-x-syns-i-lacka-fran-sisoft.html

For reference, forgot it

File: single socket holocaust.jpg (28KB, 550x550px) Image search: [Google]

28KB, 550x550px

>>60084571
Oh yes THE KING IS BACK.

>>60084571
It'll max out at 5.2 for conventional cooling. Just like the 7700K.

>>60084623
People said the 7700k would max out at 4.7 like the 6700k

My response is that you're absolutely retarded.
Oh and enjoy your 2.2V and 600 watts socket melting CPU and tub of free LN2

>>60084650
Sorry raja you're falling even further behind in single core performance

>>60084571
What's the igpu have to do with anything?

>>60084643
Curr. running on a 5Ghz frequency (OC 6700k)

>>60084666
>raja
We're talking about CPUs you braindead /v/edditor.

>>60084672
It takes up space and it does process information so it generates heat

>>60084675
Revelation 28:8

>>60084698
?

>>60084687
What 15 watts at full load? You won't be using that anyway. Negligible benefit, if any.

>>60084713
It's a poojet that responded to you

>>60084571
The integrated GPU is disabled, not removed, and it's still going to have jizz for TIM, so you're still going to have to DELID DIS. It'll go 100 MHz over the 7700K, guaranteed.

AMD's response is to release more highly efficient price/performance champions that normal human beings who aren't retarded gaymurr fanboys obsessed with MUH CLOX will buy instead.

>>60084713
Heathen

>>60084733
>AMD's response is to release more highly efficient price/performance champions that normal human beings who aren't retarded gaymurr fanboys obsessed with MUH CLOX will buy instead.
Just like the good old days.

>>60084716
More space for cores and no igpu should allow even higher overclocks. If it does over 5.5 I will replace my 7700k

>>60084571
are you pretending to be retarded or getting paid to be retarded?
disabling the iGPU does not increase the overclocking potential one bit. It only increases stock clocks.

>>60084733
Don't act like AMD wouldn't want high clocks also. Ryzen can't clock for shit, it has no igpu and no TIM

>>60084780
No they don't. They can just increase IPC in future generations of Zen.

>>60084769
It's one less thing generating heat

>indian intelectuals

>>60084751
The GPU draws zero power when a dedicated GPU is plugged in, and it hasn't since 2010.
Any improvement(if any) in overclocking come from a better process, but since both still use 14nm+, there will be no change, Kabylake isn't even like 5 months old.

>>60084751
more cores = less clock
no iGPU = same as any 7700k with iGPU disabled, i.e. 95% of them
you = tech illiterate mongrel

>>60084801
>The GPU draws zero power when a dedicated GPU is plugged in

Is you retarded?

>>60084780
Zen is a new architecture. There is a lot of room for improvement. The one advantage Intel still has is single core, and that will be almost completely erased when the more efficient Zen 2 comes out. Tick-tock has now become a countdown to Intel's complete irrelevancy.
>>60084794
iGPU uses like 0.2 watts when not in use. Wow it's fucking nothing, cocksucking shill.

Stop respond to the fucking idiot.

>>60084818
No, Pajeet. You need to learn proper English first.

>>60084821
Igpu uses 15watts. Prepare for 6ghz poo man

>>60084794
less heat = higher stock clocks
nothing else

No iGPU does nothing for overclockers because they have the iGPU disabled anyway.

And the award for most retarded OP of the day goes to..

LOL! He trolled you AMDfags good, such fucking easily baited children lmao

>>60084849
Less heat means higher thermal headroom for overclocking you dumb indian

File: Screenshot from 26-04-2017 20:41:26.png (204KB, 503x488px) Image search: [Google]

204KB, 503x488px

>>60084870

>>60084571
with 170w-tdp maybe

>>60084985
B-but muh ghz and gaymin performance, goyim. BBBBBBBBBBBBBBBUT it's time for a new socket.

>>60084985
You mean half of amd's rx580? Lmfao

File: The lake, the lake, the lake is on fire.png (1MB, 1110x5110px) Image search: [Google]

1MB, 1110x5110px

>>60084985

Given voltage scaling is non-linear it'll probably drink more juice than a 9590.

>>60084614
>intel
>holocaust joke

it fits

>>60084571
AMDs response is known

MOAR CORES

>>60085097
I miss IBM, imagine how the kikes at Intel felt when they had to license practically everything((and still do a lot of stuff) from the folks who helped gas their 6 trillion
Of course a kikes quickly forget their dignity for money

>>60084571

This isnt possible. The reason I dont believe this is because the enterprise chips cannot deliver anywhere near this. If Intel has been sitting on this shit why are they not adding to their 6k chips to get people to buy them?

>intel sits on the usual i3/i5/i7 with IGP
>AMD releases 6/8 core Ryzen without IGP

>intel releases CPUs without IGP
>AMD releases 4 core APU with integrated Vega IGP

can't wait for the next months, they will be very nice.

>>60084571
With a free tube of Ln2? Lol, it's still 14nm+, you'd be lucky to get 50MHz more than the 7700k.
You can pray it's not TIM though.

>>60085146
Intel are not even proper kikes, they are a bunch of greedy corporate monkeys.

File: Intel designed final solution.png (297KB, 550x550px) Image search: [Google]

297KB, 550x550px

>AMD releases an extremely efficient new architecture
>Intel's response is to crank up the wattage, TDP, and clocks and keep prices high
We're just missing the OEM bribing and other under-the-table bumfuckery. I missed you, Netburst Intel.

>>60085244
We need ICC gimping ayymd again.

I can't wait for Agner Fog's analysis, the enthusiast/power user part of the market will literally fucking detonate with the force of a thousand suns, the guys words are practically gospel like David Kanter's arch dives are.

>>60084571
While running at 6 000 000 C

>>60085326
Just watercool it.

>>60085290
Who is Agner Fog? What does he do? I've been hearing his name lately, but he doesn't have a wiki page or anything. I've mostly just gathered that he's a computer engineer, but what separates him from any other guy?

>>60085358
He found some code in starcraft that kill performance in amd CPUs

>>60085358
He's the guy with the most advanced SIMD library out(C++ vector class library), which is no small amount of times recommended over even the official AMD/Intel ones.
He's also the one who brought to light the whole ICC gimping bullshit which led to a 2 billion fine for Intel.

tl;dr he's the guy who's not employed by either company who knows the ins and outs of AMD's and Intel's chips the best, a real fucking wizard.

>>60085397
Oh. Well, shit. I guess if he says that AMD's IPC is generally better than Intel's, there's gotta be some truth to it.

File: Intel-Xeon-E3-1280-V6-Sysbench-Single-Thread-Benchmark.jpg (85KB, 859x593px) Image search: [Google]

85KB, 859x593px

>>60085423
You didn't actually need Agner fog to figure that out, there are a few benchmarks that actually take advantage of AMD's 4x complex decoders and it IPC it really shows, Intel's 1 complex + 3 simple can't really compare when programs are put to use said decoders.

Here's an example in a pure ST bench, when used it can feed the AMD cores far faster, even the 7700k is helpless, clockspeed matters shit when hardware has far more execution units.
If AMD gains marketshare, you can bet these things can be utilized more.

>>60085485
Oh wow what a shitshow

File: 1492899694971.jpg (20KB, 600x448px) Image search: [Google]

20KB, 600x448px

>>60085485

>>60085527

For these sorts of workloads Intel needs a hefty core advantage (nas naturally more cores = lower clocks) to make up the difference against a 1800x. Its why a lot of people are excited for what Naples may bring as it is going to be an absolute monster for a lot of stuff.

>>60085485

That review is quite interesting - the 1800x wins every single program they test.

File: Screenshot from 15-04-2017 01:35:14.png (72KB, 939x891px) Image search: [Google]

72KB, 939x891px

>>60085590
They're all Linux tests, under Linux the 1800X gets so close to a 10 core 6950X it's almost scary.

Windows being a piece of shit with new non-standard hardware isn't exactly a new thing, Linux kernel had full enablement patches months before Zen launched.

>>60085622
If or when do you think Windows 10 will get any patches?

People always talk about IPC. But I wonder:

what's the actual IPC *value* under normal use cases? 0.8? 2? 0.3?

>inb4 it depends

then make an example you moron

>>60084571
>7740k might even hit 6ghz on water since it has no integrated gpu
just like haswell-e unlocked super overclocking powers because it ditched the igpu? retard logic.

>>60085649
When MS stops sucking Intels cock, so I can't exactly say.

File: IMG_2801.jpg (89KB, 500x485px) Image search: [Google]

89KB, 500x485px

>>60085649
Never. Windows needs less but fast cores rather than a lot of slow cores like amd. Games are the best example. The gpu needs 1 or 2 but real fast cores instead of 8 slow ones

>>60085622

Makes Naples even more interesting - nobody cares about windows in the server space so this sort of performance is nothing to scoff at.

>>60085660

IPC is 1) not a static thing and 2) meaningless without a comparison. I suppose though the normal result will always be 1.0 as a chip will run any given task as fast as it can, thus IPC can never be greater (or below) 1.0 without external factors.

>>60085675
Nobody asked you, shill that shows up in every one of these fucking threads with the same exact day one benchmarks on games that have already been patched. Fuck, even the image name is the same every time. Fuck off already.

>>60085660
IPC isn't static, in one thing it can be .6 of the competition while in the other it can be 1.5.

It's hard to quantify this stuff, there's far too many variables.

File: Dawn of War III cpu.png (82KB, 523x440px) Image search: [Google]

82KB, 523x440px

>>60085675

>meanwhile, in the 41st millenium...

>>60085688
>>60085714

you fukin wut m8s
step 1: run a program
step 2: measure how long it took
step 3: multiply that by the clock
step 4: divide the amount of instructions executed by the number you just got

and that's your IPC. Also, assuming applications run similarly fast on two similarly fast processors from intel and amd, IPC will also be similar.

You just don't know the actual number (or what IPC means).

File: IMG_3347.jpg (85KB, 1277x683px) Image search: [Google]

85KB, 1277x683px

>>60085721
Where's the 7700k in that graph? Funny because you compare poozen to 2 generation old intels and your indian cpu still loses. Pathetic

>>60085753
Just fucking stop it. It's not funny anymore.

File: Screenshot from 17-04-2017 00:18:16.png (212KB, 977x688px) Image search: [Google]

212KB, 977x688px

>>60085745
IPC
IS
NOT
STATIC
YOU
FUCKING
APE

You have to look more than 2 benchmark to figure it out, if the IPC was same always, then you wouldn't see so much variance in benchmarks!

Some architectures have pipeline PITFALLS, meaning they fall FLAT ON THEIR FACE in certain workloads, like the one I'm posting right now.
Explain how the Broadwell-E is doing worse than Bulldozer? Explain how a chip with SMT cores that has over 55% IPC lose to fucking Bulldozer?

Stop talking about shit you have no fucking clue about, you shitty /v/ermin

File: IMG_2740.jpg (443KB, 2133x1242px) Image search: [Google]

443KB, 2133x1242px

>>60085706
Check out this DigitalFoundry benchmark posted in April. No amount of time will make your poozens boost enough to catch up with intel.

File: IMG_3353.png (42KB, 653x726px) Image search: [Google]

42KB, 653x726px

>>60085781
S-S-S-SHUTUP

>>60085800
Posting the same wojak edits are not funny anymore. At least make some new ones.

>>60085788
>Some architectures have pipeline PITFALLS, meaning they fall FLAT ON THEIR FACE in certain workloads, like the one I'm posting right now.
you mean it depends on the use cases? like exactly what I was saying in my first post?

>>inb4 it depends
>then make an example you moron

The fact that it isn't static doesn't mean that it doesn't have values at specific point in times. Stop making a fool of yourself and just stop posting if you don't know the answer.

File: Screenshot from 25-02-2016 14:05:08.png (266KB, 606x507px) Image search: [Google]

266KB, 606x507px

>>60085745
Stop posting, your post is literal diarrhea and it's obvious your whole knowledge about CPUs stems from some marketing grunt from youtube.

>>60085819
DESIGNATED

>>60084571
>might
:^)

>>60085790

>Nvidia gpu
>paired with ryzen
>in RoTTR

You're not even trying.

>>60085831
I'm actually making some test case in assembly to measure this, since I'm curious. And implementing a toy CPU in a FPGA and programming a toy cross compiler for it had nothing to do with my knowledge of CPUs.

Something tells me that YOUR knowledge about CPUs stems from /g/ though :)

>It's a /v/ tries to wrap its head around concepts it's incapable of computing but puts in cute effort before moving to shitposting

>>60085907
>"this benchmark is too old, not ryzen fault"
>"this gpu too good for ryzen, not ryzen fault"
>"this API no good, not ryzen fault"

I hope you see how stupid you look

>>60085884
It's good considers it will take amd another 8 years before reaching 4.5 ghz lmao

>>60085956
>clocks are everything >>60084614

>>60085971
>says faildozer

>>60086007
>muh faildozer
What was shitburst?

File: Dawn of War III 1400p gpu.png (152KB, 548x629px) Image search: [Google]

152KB, 548x629px

>>60085942

Its been proven that that specific combo of hardware and software gives lower performance than having ryzen and an AMD gpu. Why do you think /g/ suddenly loves RoTTR?

>>60086016
>both 3.5 meme and keklers are lower than non-x 290
My sides.

>>60086016
Ah, what happens when those 4096 ALUs can actually be put to use.

File: IMG_2558.jpg (504KB, 2094x1242px) Image search: [Google]

504KB, 2094x1242px

>>60086016
Or your indian cpu is just shit and can't communicate information to the gpu fast enough?

>>60086040
>indian
Since when Keller, Clark or Papermaster are indiands? Shoo shoo Rajeesh.

>>60086038

The order in that benchmark is almost exactly in order of a gpu's compute performance, so whatever DoW III is doing it isn't pushing the front end of those gpus.

>>60086053
Amd is a literal indian company. The guys that put ryzen together wipe their poo with their bare hands after they shit in the street you stupid pajeet

File: Panorama.jpg (600KB, 2550x1446px) Image search: [Google]

600KB, 2550x1446px

>>60085753
>Cherry picking this hard

>>60086068
You have to go back, Rajeesh.

>>60086058
AMD's GPUs prefer a continuous stream of data to the command processor while Nvidia's ones want data in batches, these batches probably overwhelm AMD's frontend and leads to stalls.

File: Best_Processors_April_2017_20170426_14.13.40.png (478KB, 1184x3530px) Image search: [Google]

478KB, 1184x3530px

>>60085753

>>60086058
Throwing fuckton of models and effects on screen at once.

>>60086074
>no source or watermark

You can do better

File: Panorama2.jpg (538KB, 687x5910px) Image search: [Google]

538KB, 687x5910px

>>60086040
Kill yourself, shekelchaser.

File: 1492587863785.jpg (395KB, 2400x2000px) Image search: [Google]

395KB, 2400x2000px

>>60086088
Commit sudoko

>>60085942
>I don't know that new architectures benefit greatly from optimizations
Wew

>>60086079

Perhaps, but geometry throughput is way, way higher on Nvidia's architectures compared to even polaris. Clearly the game is not hammering the areas where Nvidia holds the advantage.

>>60086088

Legit reviews.

http://www.legitreviews.com/intel-core-i7-7700k-versus-amd-ryzen-1700x-14-game-cpu-showdown_192508

File: IMG_3359.jpg (92KB, 687x683px) Image search: [Google]

92KB, 687x683px

>>60086093
You just BTFO'd yourself dumb indian lmfao

>>60084801
You must also consider that the new intel socket is huge and so you'll have the heat of a 4 core processor with no igpu being dissipated over a larger surface. The chips are massive. If they solder it too that would be even better. The power usage is probably going to be monstrous under it's max overclock though.
It should be interesting at the very least. I'm still hoping for the trend towards parallelism in games to continue or for zen 2 to catch kaby in single thread be it through ipc or frequency.

>>60086121
It's normal to expect a game not to use a tessellation level of billion so you're lead to thinking your GPU is shit.

>>60086134
>The chips are massive.
It's still that small fucking die just slapped on large PCB.

>>60085753
>Funny because you compare poozen to 2 generation old intels and your indian cpu still loses. Pathetic

185% increase in price for 25% increase in performance on a framerate 99% of people won't even be able to utilise

>>60086121
>tested with a gtx 970

Stopped reading right there

>>60086155

>most popular card in existence
>"it doesn't count!"

I'll add it to the list.

>>60086152
Not everyone is a poor indian. Some people don't care about $20 and would rather have the performance.

File: 1492842903517.jpg (2MB, 3840x1032px) Image search: [Google]

2MB, 3840x1032px

>>60086155
>Cherry picking: The post
E-en-jo-y yo-u-r stttt-utering -ho-house-fir-ez.

>>60086173
>would rather have the performance.
And heat.

File: ryzen7.png (2MB, 769x8779px) Image search: [Google]

2MB, 769x8779px

>>60086173

>>60086170
>test gpu performance by a gpu that bottlenecks them
>indian testing methodology

Come back with at least a 1080 poojet

File: RYZEN-1600X-1500X-34.jpg (128KB, 614x517px) Image search: [Google]

128KB, 614x517px

You know it's almost disgusting how at times Intel's CPUs fall flat on their face.

File: RYZEN-1600X-1500X-35.jpg (168KB, 614x570px) Image search: [Google]

168KB, 614x570px

>>60086190
Zen's FPU seems to be insanely powerful, I guess Agner wasn't kidding when he said it only loses in pure 256bit SIMD

>MOAR MHZ MORE HEAT
it's pentium 4 again
when does intel do secret shady OEM deals and get sued and lose to AMD again?

File: IMG_3360.gif (2MB, 177x150px) Image search: [Google]

2MB, 177x150px

>>60086176
>testing cpus
>with a 480

>>60086215
Wait a year or so. They are already desperate enough to produce another round of bentium4ee-tier hosuefires.

>>60086218
That's the best amd has. Just wait for vega.

File: RYZEN-1600X-1500X-33.jpg (145KB, 614x517px) Image search: [Google]

145KB, 614x517px

>>60086190
Zen, the CPU for terrorist encryption and porn stashes

>>60086218
>oy vey goy please test with Nvidia we'll gimp your Ryzen properly

>>60086250
yet amd tested poozen with titans
back to the designated street poojeet

File: 7700K%20OC.png (84KB, 837x903px) Image search: [Google]

84KB, 837x903px

>>60086241
>Wait a year or so.
More like wait 2 months, the 7740k will be close to pulling 170-180W at stock.

Btw here's a "91W TDP" chip using 150W at stock.
While TDP isn't power usage, exceeding it this far is only allowed on mobile parts.

>>60086243
Which is shit. But then again, you're a brainlet indian and can't factor that gpu bottleckecks shouldn't be present in cpu testing lmao

>>60086243
JUST WAIT

I think naples is going to be the most interesting thing coming our way. Zen is an integer monster and performs really well with the linux kernel. Hopefully they make enough money to go insane. It would be so nice to see RTG hit a home run with chips after vega. Even if you have owned nvidia cards in the past if you have any interest in adaptive sync gsync adds 300 dollars to the cost of the nvidia card in comaprison to the radeon.
Future APUs on modern AMD archs with innovative memory solutions could be really interesting as well. Once an AMD can do 1080p 60fps we'll see some crazy form factor computers and laptops.

>>60086338
JUST WAIT

>>60086338
After Vega comes Navi and it's all about advanced packaging methods applied to GPUs. Boohoo, first real innovation in god knows how many years.

>>60086281
>99% of 7700k's reach 5.0-5.1GHz*

[spoiler]
* After winning silicon lottery, deliding, sanding on a waterloop and with an -20 AVX offset[/spoiler]

>>60086366
JUST WAIT FOR NAVI

>>60086214

To be fair I severely doubt /g/ will be discussing his findings. Even The Stilt's rather simplistic analysis is ignored because it goes beyond what most anons understand. Actually as The Stilt focused a lot of effeciency it provides an interesting insight into what AMD focused on in the design and i'd wager the desktop chips are essentially (obviously it more involved) just the shit chips AMD has been stockpiling for Naples and possibly mobile chips.

>>60086368
[spoiler]ha[/spoiler]

>>60086393
I'd gladly wait for something actually NEW in GPU space. MCMed GPUs sounds fun as fuck.

>>60084571
It probably has an integrated gpu, but it's just disabled.

>>60086338
Actually it seems Zen has a beefier FPU >>60086214, but loses slightly in vector SIMD due to way shorter pipes.

>>60086366
The innovation was when you jump from 60 fps with a 980ti to 200 fos on a 1080ti. Raja has a thing or two to learn

>>60084666
>>60084722
>>60084794
>>60084830
>>60084876
>>60085753
>>60085790
>>60085865
>>60086040
>>60086053
>>60086068
>>60086077
>>60086126
>>60086173
>>60086187
>>60086276
>>60086293
>>60086427
Hey guys! Just wanted to let you know that racism against Indians isn't cool. We're all humans on this Earth, so please show some respect! Phrases like "Pajeet" and "DESIGNATED" are very demeaning, and I'm sure the actual smart people in this thread would appreciate it if you added some diversity to your vocabulary. Thanks for reading!

>>60086400
>desktop chips are essentially (obviously it more involved) just the shit chips AMD has been stockpiling for Naples and possibly mobile chips.
They are. Every golden Zeppelin goes for Naples for dat sweet sweet perf/watt.

>>60086368

As of 2/22/17, the top 78% of tested 7700Ks were able to hit 4.9GHz or greater.

As of 2/22/17, the top 59% of tested 7700Ks were able to hit 5.0GHz or greater.

As of 2/22/17, the top 28% of tested 7700Ks were able to hit 5.1GHz or greater.

As of 2/22/17, the top 7% of tested 7700Ks were able to hit 5.2GHz or greater.


Even if you do delid and repaste the chip you still have just over 50% chance you can actually hit 5ghz which I find out as /g/ tells me all 7700k's hit 5ghz.

>>60086400
/g/ is irrelevant to Agner's analysis, nobody gives a shit what /g/ or reddit think, but wanna know who cares about what Agner says? HPC folk.

>>60086427
That's not innovation, that's throwing more hardware due to dieshrink. NVIDIA fails at innovation, CUDA is their only notable invention.

>>60086440
Don't these fags test overclocks with a 1 hour prime95 test? I don't consider overclock a real overclock unless it can do IBT or Small FTT over 10 hours.

>>60086471
>overclocks with a 1 hour prime95 test?

Someone tell me this isn't real.

>>60086471

Basically - their testing os woefully poor for what they test (lel speed binning). Still it indicates the actual stable clocks will be lower than what they claim.

>>60086485

lolololo

Passed the ROG RealBench stress test for one hour with these settings:

50x CPU Multiplier
1.376V CPU VCORE (Or less)

>>60086471
Only 30 minutes if you don't want your amd pootherboard to catch fire

>>60086485
It's real, and it's retarded as fuck.
These CPUs are for people that use their CPU for a few hours and turn off the PC like some fucking console.

If can't do that without delidding it doesn't count.

>>60086517
You can delid any poozen and it still won't hit 1 mhz higher than 4.1ghz

>>60086538
Ryzen has nothing to do with new Intel housefires.

>>60086546
Besides the fact it's forcing Intel to pull out more hilarious housefires like the 7740k as a response.

>>60086561
Well, that's classical Intel.

>>60086546
Rich coming from 580 hoursefire

>>60086578
GPUs have nothing to do with CPUs.

>>60086593
FIREPOO

>>60086623
Are you insane?

>>60086593
AMD GPUs seem to have an advantage with Ryzen against equally priced Nvidias
Most probably driver compatibility

I love shill always bragging about day one benchmark

>>60084571

s-six ghz? i guess i'll hold off on buying that 7700k. hopefully ram prices will be down a little bit by then as well.

>>60086682
It's probably even better, I bet 6.3GHz

>>60086682
Not gonna happen, RAM prices are still going to go up because the big DRAM makers are basically colluding with each other.

>>60086745
We know it's AMD's fault, if Ryzen wasn't released the priced would remain low

>>60086812
shit b8, you'd have to be a retard or live under a rock to believe this.

>>60086883
Prove it poojet

>>60084571

7740K-7760K are going to be beta-test-tier shit for X299 platform.

You can't even use the quad-channel DDR4 and 40+ PCIe 3.0 lanes on X299 platform.

It will have marginally better overclocking headroom then its socket 1151 counterparts.

>>60086743
Fuck that, they're gunning for 7 gigaheezys. The Tejas prophecy shall be fulfilled.

https://en.wikipedia.org/wiki/Tejas_and_Jayhawk

>>60086883
Amdrones pushed the meme that their cpus perform like shit because of slow ram and they would perform better with faster ram (they don't) so ram companies took advantage of this and cranked up the price

>>60086942
>51 pipeline stages
What in the ungodly fuck.

>>60086968
>No graphs
>No charts
>No tables
>No links
>No embedded videos
>No evidence at all
Kill yourself, shekelchaser.

>>60086986
Where did you think the BIPELINE :--DDD jokes came from. Intel was working on a weapon to surpass Metal Gear.

>>60084614
>it hits 6 million ghz

I can only imagine

>>60087001
careful not to pop your hemorrhoids poojeet

>>60087002
How motherfucking insane one has to be to design THAT? Prescott was fucking retarded, but what were they thinking with that.

>>60087038
Intel was about to drop a 7ghz cpu in 2004 meanwhile amshit can't do 4.2ghz in 2017 lmao

>>60087051
0/10 the worst bait I've ever seen.

>>60087051
MORE BIBELINES! MOAR GIGAHURTZ! MOAR TDP! MOAR HEAT! MOAR DELIDDING! MOAR LN2 COOLING!

Ah the joys of a totally dead end strategy :^)

>>60087067
You clearly haven' seen Tom's reviews of PressHOT if you think this is 0/10, this is no more than noeffort/10

>>60087067
It's not bait read the article and then try oc'ind your poozen past 4.1ghz

>>60087086
My more is slow as shit I know what we'll do poojets just add MOAR COARZ

Please don't indulge the wild retard, it only makes him stronger.

>>60084571
Intel will nerf it with shitty TIM to save some pennies in this times of need

>>60087175
At least when intel fucks up they don't hire 30 pajeets to shit up this place

It feels like there's a bunch of bots spamming random combinations of "pajeet(s)", "poo" and some other words.

>>60086432
fuck off pajeet.

File: Screenshot from 17-04-2017 23:15:20.png (143KB, 665x699px) Image search: [Google]

143KB, 665x699px

>>60087199
They do better.

File: IMG_2741.png (609KB, 854x640px) Image search: [Google]

609KB, 854x640px

>>60087234
Rich coming from the poo guy

>>60087234
>Whye
>sounds like Whey
>it's very white

I'M BEING OPPRESSED BY HER TOXICLY PROBLEMATIC NAME HELP ME

File: mpv-shot0028.jpg (137KB, 1280x720px) Image search: [Google]

137KB, 1280x720px

>>60087269
Intel's got you covered.

File: 374.jpg (76KB, 709x520px) Image search: [Google]

76KB, 709x520px

Meanwhile in the home of the Shekelchaser Grinch

File: Screenshot from 26-04-2017 23:45:32.png (135KB, 399x437px) Image search: [Google]

135KB, 399x437px

>>60087319

>>60087319
Oy vey delet it goy.

>>60087319
>>60087373
Do you think he stuck his gold plated shriveled old dick into one of the poo's poo holes?

>>60087105
Try speaking English, Shiester Shekelstein.

Today I will remind them
https://ark.intel.com/products/93791/Intel-Xeon-Processor-E7-8893-v4-60M-Cache-3_20-GHz

>>60084571
Intel is now the Just Wait(TM) company. What a time to be alive.

>>60086249
and that's a problem why? :)

File: 12345241332.png (438KB, 1064x1619px) Image search: [Google]

438KB, 1064x1619px

>>60086281

126w 4.2ghz at 1.150v since its base clock.

>>60088630
7700k turbos all cores to 4.4

>>60087223
Because it is, we're being invaded by Intel shills. Same buzzwords, same asinine arguments, same day one benchmarks with the same filename posted over and over again.

File: egg.png (111KB, 465x290px) Image search: [Google]

111KB, 465x290px

>>60085244
>the past decade
>Intel: the king of efficiency in their architecture, allowing for die shrinks, performance gains, lower power consumption
>AMD: MOAR COARS

>Launch of Ryzen
>AMD: finally has their shit together with an amazing new architecture that is proving to surpass Intel's improvements and then some
>Intel: MOAR GHZ

>>60088684
No, more like a bunch of retards shitposting like crazy.

>>60085649
>If or when do you think Windows 10 will get any patches?

never. It's no secret that Microsoft has been in bed with Intel for a while now. The best that W10 users can hope for is some bios updates that can fix the issues with the hardware and probably squeeze more raw power from the CPU.

>>60088677
set multiplier in bios
fixed

>>60084571
haha.
no.

>>60086432
I agree but you're on 4chan, so the best would be to ignore this as the rest.
Fuck off newfag !

>>60084571
pssst.
you can turn off igpu on the current delid lake.

>>60084571
>amd's response
amd's sempron

File: 5ghzSkylake.png (568KB, 1185x935px) Image search: [Google]

568KB, 1185x935px

>>60084675

File: Daily.png (226KB, 410x393px) Image search: [Google]

226KB, 410x393px

We should be at 1.21 jiggawatts by now
5ghz was nearly 10 years ago on a desktop chip

Pick related i have been running 4.9ghz on 6 cores for nearly 7 years now

>>60089052
with which cooling and what's your cpu-z bench?

>>60084571
>>60084584
says nothing about 6 GHz kill yourself fucking insane bullshit tard

>>60085208
yeah right, they fund feminist frequency and shit like that, which is exactly what you would do if you want to push for the jewish agenda, and exactly what you wouldn't do if you wanted to appeal to your customer base of gamers and white/gook businessmen

File: 1486961261431.jpg (678KB, 1035x1360px) Image search: [Google]

678KB, 1035x1360px

>The single-thread instructions per clock rate of Ryzen is higher than for any Intel processor, except for 256-bit vector code.

THREE
BUS
RINGS :--DDD

>>60084584
Where do you pretend to have read 6 GHz in that article, retard?

File: housefire inside.png (126KB, 357x330px) Image search: [Google]

126KB, 357x330px

i7-7740K hitting 6 GHz on water right behind this door.

>>60085485
That's strange, though. Most single-threaded benchmarks I've looked at seem to show a result at least slightly slanted towards Intel.

File: JUST_krzanich.png (1MB, 1002x1480px) Image search: [Google]

1MB, 1002x1480px

>>60089276

>>60085660
It really depends extremely much on the program you're running. But for what it's worth, here are some examples when I'm running the same things on an i5 2400, a Xeon E3-1271 and an Athlon 64 that I have access to (in that order).

>GCC compiling one of my projects -- 1.17 / 1.42 / 0.70
>Stupid file hashing program that I wrote just for benchmarking -- 1.66 / 2.62 / 1.42
>mpg321 -- .2.49 / 2.88 / 1.70
>A Certain Game Server -- No test / 0.90 / No test
>Chromium loading CNN -- 1.04 / No test / No test
>Gnumeric starting up -- 1.14 / No test / No test
>git log through 11k commits -- 1.35 / 1.67 / 0.82

>>60084780
Ryzen doesn't clock high since the process is tuned for efficiency, Zeppelin and Zen were made squarely for the server market
>>60084840
iGPU hasn't used anything at all when disabled since 2010, and only draws 15W when ran at full load with turbo clocks dumb /v/edditor
>>60084985
A weapon to surpass the Intel Designed Final Solution
>>60085279
ICC never stopped gimping AMD and VIA, old versions even gimp newer Intel processors, and newer versions gimp the old versions
Agner Fog still has a thread on this where he has revisited the issue since 2003
>>60085622
>Linux kernel had full enablement patches months before Zen launched.
It didn't, it was way more patched than Windows (which had ZERO real support), but some stuff didn't made it to 4.10, and some stuff isn't even submitted anywhere yet
You're also missing that all his binaries are compiled with recent GCC versions optimized for the CPU's he's testing, so no ICC gimping
>>60085423
He also tests IPC using the performance counters on the CPU, and by measuring the clocks it takes to execute instructions
He does this for plenty of scenarios, and analyzes the arch in depth
>>60087038
It was already in tapeout
Management couldn't believe that they wasted 5 years on Netburst when the mobile division could make better CPU's in a few years based on P3's
They finally accepted reality when the antitrust started
>>60088047
>a-add moar bing busses :------DDD
>>60088692
Intel
>MOAR GHZ
>MOAR BING BUSS
>MOAR ITANIC AGAIN
>:-------DDD

>>60084571
6ghz is overkill anyway,amd stiill is better for cheaper and faster than average i3.i would not buy anyway even if i had money,there is no practical use for that speed

>>60086432
Says the Pajeet.
Welcome to 4Chan. Fuck off.

>>60089364
Read the post. Again.

>>60089749
Are you saying "Sysbench CPU" is a test specifically designed to test decoding throughput?

>>60089185
Gigabyte X58 SOC?

>>60089276
You mean to say that when operating out of the µop cache, Zen can deliver, rename and dispatch 6 ops/cycle? If that's true, that's quite cool.

>>60084840
http://www.anandtech.com/show/11232/intel-launches-xeon-e3-1200-v6-family
E3-1275 v6 73w with igpu
E3-1270 v6 72w

File: gutswat.jpg (617KB, 1600x1200px) Image search: [Google]

617KB, 1600x1200px

>>60089185
>>60090214
Its on water loop

Asus rampage III black edition

File: 1212692824726.jpg (199KB, 562x791px) Image search: [Google]

199KB, 562x791px

>>60084614
>that filename

File: 1354482626418.jpg (527KB, 1920x1200px) Image search: [Google]

527KB, 1920x1200px

>>60089675
Its not overkill just way WAY outside what silicon is meant to run at
Voltages and amp loads are going to be redonk

>>60085675
>Ryzen
>Ram speed not listed

>>60089633
>>Stupid file hashing program that I wrote just for benchmarking -- 1.66 / 2.62 / 1.42
Interestingly, we can also simulate this to some accuracy, since it's a very simple inner loop. Might be interesting to some Anons. Here's the C source for the hash function:

int hash(char *fn)
{
    int i, h, ret, fd;
    uint8_t buf[65536];
    
    fd = open(fn, O_RDONLY);
    h = 0;
    while((ret = read(fd, buf, sizeof(buf))) > 0) {
    for(i = 0; i < ret; i++) {
        h = (h * 31) + buf[i];
    }
    }
    close(fd);
    return(h);
}

And here's the assembly that GCC generated from that inner loop:

1:       89 da                   mov    %ebx,%edx
2:       43 0f b6 4c 05 00       movzbl 0x0(%r13,%r8,1),%ecx
3:       49 83 c0 01             add    $0x1,%r8
4:       c1 e2 05                shl    $0x5,%edx
5:       29 da                   sub    %ebx,%edx
6:       89 d3                   mov    %edx,%ebx
7:       01 cb                   add    %ecx,%ebx
8:       44 39 c6                cmp    %r8d,%esi
9:       7f e6                   jg     400880 <hash+0x50>

(Upon entry, %esi holds the value of ret.)

cont.

>>60090876
Analyzing it for for the i5 2400, a Sandy Bridge part, we can see that the critical path through this loop are the instructions #1 -> #4 -> #5 -> #6 -> #7 -> #1, since each of those depend on the result of the prior instruction. Each of the instructions are 1-µop instructions that all have a latency of 1, so each iteration of the loop must take at least 5 cycles. Looking at the remaining instructions, #2 and #3 are simple 1-µop instructions, while #8 and #9 macro-fuse into one µop. #3 forms another dependency loop with itself over loop iterations, but since it has a 1-cycle latency, it's hidden by the main dependency loop. #2 is a D-cache load, and so has a latency of 4 cycles, but only #7 depends on it and in turn only depends on #2, so it should be more than well capable of running ahead so that its value is ready when #7 needs it.

All in all, then, the core should find no trouble executing #2, #3 and #8&9 "in the background" while running the critical path, so that each iteration of the loop takes 5 cycles. Executing 9 instructions even 5 cycles should give an IPC of 1.8. I'm honestly not sure why it doesn't quite reach that. Perhaps there's some issue conflict between #4 and #8&9 (since those are restricted in what ports they can be dispatched on), but seeing how much headroom the processor should have in executing #8&9 while the rest of the critical path is running, that seems like stretching it. If anyone has some insight, please share it.

>>60090876
Also, fuck 4chin for not handling tabs properly in the code tag.

>>60086357
COMING SOON

>no integrated gpu
muh gamers

File: total nuclear annihilation.jpg (7KB, 240x180px) Image search: [Google]

7KB, 240x180px

>>60090365
how about this one

How many watt would 7740 6hz pull
300?

>>60090884
Analyzing it next for the Xeon, a Haswell part, the main dependency loop is of course the same, but Haswell's front-end is smart enough to be able to eliminate #1 and #6 completely in the renamer, so the real dependency loop is #4 -> #5 -> #7 -> #4, only three instructions long. That leaves much less headroom for running #2, #3 and #8&9 in the background, but not so much that it should block anything, especially since Haswell has so many ports available for ALU operations.

In theory, then, Haswell should be able to execute a loop in three cycles, leading to an IPC of 9 / 3 = 3.0. Again I'm not sure why it doesn't reach that and stays at 2.62. In this case, a dispatch conflict between #4 and #8&9 is perhaps a bit more likely, however, since the critical path is so much shorter. #4 is constrained to ports 0 or 6, and #8&9 to port 6, so conceivably there could be some cycles where a #3 occupies port 0, forcing #4 and #8&9 to conflict for port 6. #3 can also dispatch to ports 1 or 5, though, so that shouldn't be a problem, but perhaps the scheduler can't cope with constrained µops generally enough.

Interestingly, on Sandy Bridge, the average number of cycles per loop iteration is 5.42, while on Haswell it is 3.44, so in both cases it is almost exactly as many cycles (~0.43) above the theoretical minimum, so there could be some common factor at play. No idea what that would be, though.

>>60091019
>Implying AMD's housefire series can hold a candle to Intel's housefires
Kek. AMD BTFO one again.

>>60091019
6/10 - underage tier
>>60091111
Nice digits

>>60090884
>Interestingly, on Sandy Bridge, the average number of cycles per loop iteration is 5.42, while on Haswell it is 3.44, so in both cases it is almost exactly as many cycles (~0.43) above the theoretical minimum, so there could be some common factor at play. No idea what that would be, though.
Managed to figure it out, and as expected it was much simpler than that. It turned out that the remaining ~half clock cycle per byte is spent in the kernel, in the read() call. Apparently Linux needs that much time to copy data to userspace.

File: 2017-04-26.png (615KB, 877x953px) Image search: [Google]

615KB, 877x953px

prasanjeet my son

Their response is not having 72% profit margins and fucking the consumer.

>>60091471
>X skews
>not fucking the consumer
>not having a large market share is a good thing
ehehehehehe

>>60091375
Pajeet working for AMD.

>>60091375
who else going to see this kino? Loved part 1

File: 1492864332826.jpg (690KB, 1680x1050px) Image search: [Google]

690KB, 1680x1050px

>>60091375
Prasanjeet looks like an alpha, damn, are all AMD users like this?

>>60091618
>buff indians who shit in the streets
probably I dunno I've never seen one in public

>>60086432
gr8 b8 m8

>>60091181
>I-It doesn't mater if it's relatively recent
You amdpajeets are hilarious. Also bartons running at 100C was daily occurance.

>>60084571

6Ghz?
You mean 6 million kilohertz?
Oy vey truly the Shoah of our time.

No fucking way it's hitting 6.0ghz.
I say it will hit 5.4ghz in average. Not like that extra GHz is gonna save it from the 6th core competition

File: words can not describe my anger.png (416KB, 703x764px) Image search: [Google]

416KB, 703x764px

>>60084571
>might
>might
>might

File: dumb fags.png (21KB, 900x350px) Image search: [Google]

21KB, 900x350px

>>60084571
>6GHz on water
>when ONLY the top 7% of 7700Ks were able to hit 5.2GHz or greater.
Is Intel about to become the space heater meme?

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

495KB, 1070x601px

>>60084571
>What's amd's response to this?
Better IPC instead of MOAR GIGAHURTZ

>>60086485
The most I will go on p95 is 20 minutes, I hear some retards do 24 hours or 16 hours.

>>60092115
Amshit is far behind intel in IPC regardless

>>60093347
You mean ahead? Because they are.

File: IMG_3144.png (3MB, 2208x1242px) Image search: [Google]

3MB, 2208x1242px

>>60093362
Nope

>>60093373
Yes, according to Agner.

>>60093385
Real world test > what some random pajeet says

File: 1352197404494.png (89KB, 399x382px) Image search: [Google]

89KB, 399x382px

>>60093385

Let's see what Agner has to say on the subject.

http://www.agner.org/optimize/blog/read.php?i=829
>The single-thread instructions per clock rate of Ryzen is higher than for any Intel processor, except for 256-bit vector code.

>>60093399
Agner > random pajeettube test

>>60093404
Real test > words

Or is it the other way where you live in india?

>>60093385
>>60093400
who is this Agner guy? seems interesting.

>>60093424
Wow, one cherrypicked benchmark in a bideo game.

Really tickles my neurons.

>meanwhile, in the real world
>>60085485
>>60085622
>>60085721
>>60085788

>>60093435

Read >>60085397

File: IMG_2802.png (538KB, 1279x720px) Image search: [Google]

538KB, 1279x720px

>>60093441
>linux kernel compiler

Try a real benchmark which would help most people not just forever NEET virgin masturbators

>>60093465
>four benchmarks in different use cases where AMD blows the Intel processor out of the water
>mentions the one linux kernel compilation benchmark

You break my heart. I even included another video game test, just for you Anon.

>>60093451
bwhahahahah
dohohohohoh
https://www.pcper.com/news/General-Tech/Intel-still-hasnt-paid-AMD-12-billion-USD-anti-trust-fine

>>60093490
>game where amd loses
>linux compilers
>literally who benchmark
>literally who benchmark

>>60089203
He never said anything about the article saying that holy shit you imbecile fuck.
It just makes sense that it has no iGPU so has more thermal headroom.

Holy shit.
> inb4 fanboy
Fuck off, reading comprehension and common sense has nothing to do with brands.

File: 1424180323693.jpg (148KB, 500x500px) Image search: [Google]

148KB, 500x500px

>>60093529
>1080p gaming benchmarks
>in the year of our lord 2017

You have to go back.
>>>/v/
>>>/hm/

>>60086152
LOL it's a benchmark.

>>60093550
>what are 240hz refresh monitors

>>60093565
>"muh input latency at 240hz"

Didn't know people were cosplaying as CS:GO pro players now. What's your rank, Gold Nova II?

>>60093582
Open a elementary anatomy book, human eye can see up to 245hz

File: SMUGGED.png (169KB, 720x470px) Image search: [Google]

169KB, 720x470px

>>60093582
>>60093565


>>60093591
>he said with no source
Lies, someone sitting in a chair and actively guessing at how high a framerate is can, on average, interpet up to about 150 frames per second.

>>60093582
> being such a pleb
kek

>>60093614
Dumb anime poster

File: fucking really.jpg (611KB, 1680x1050px) Image search: [Google]

611KB, 1680x1050px

>>60093642
>he thinks it's anime
You're missing out on some top notch smug.

>>60085048
I'm sorry since when u put rx580 in cpu socket ?

>>60094116
Poo in the loo Pajeet.

File: intel_j.png (148KB, 763x930px) Image search: [Google]

148KB, 763x930px

>>60084571
let me guess:
minor improvement, charge 3x as much as AMD

>>60094170
>minor
It's clocked 100mhz higher. That's it. There's no improvement.

>>60086080
>those prices

jesus christ, intel must be on fucking suicide watch

>>60094170
Don't forget the socket change! Truly one of our greatest allies.

>>60094191
fkn lol, typical Intel

File: Screenshot_2017-04-27-11-47-25-642_org.mozilla.firefox.png (127KB, 1080x1920px) Image search: [Google]

127KB, 1080x1920px

>>60084571
Wow, it's nothing.

>>60094324
Why would they release a CPU that's similar to the previous one, but on a different socket?

>>60092052
What cooler and thermal paste is this?

>>60094356
They are panicking.

What's Intel's next name for their cpus after Coffeelake? Tealake?

Or after Vega and Navi what's AMDs next name for their gpu, vladimir?

File: the year of intel netburst 2.0 spurdo.png (871KB, 801x1500px) Image search: [Google]

871KB, 801x1500px

>>60094324
>has no iGPU
>112W compared to 91W i7 7700k that has an iGPU
mein sides

>>60094419
>what's AMDs next name for their gpu, vladimir?
Shiva
the creator and destroyer of worlds

>>60093373
> 7600k > 7700k
what? How is this possible?

>>60094452
Is this confirmed? After Zen plus? Nirvana? These fuycking names are hilarious,

>>60086080
Whoa 1700 is selling like a hotcakes.

>>60086040
Was gonna get Ryzen so I could play CS:GO on my new 1080p 500hz monitor but gonna go Intel now. Thanks Intel!

>>60086080
I love it when Intel is in trouble. The kikes have to innovate again.

>>60086432
https://www.youtube.com/watch?v=_peUxE_BKcU

>>60094517
>Intel
>innovating
Things that never happened.

>>60094556
intel was the first the release quad core.

>>60094586
It was two dual-core dies slapped together in one package. That's no innovation.

>>60094609
Like amd's 2 4core ccx?

>>60094695
That's a single die with different topology.

>>60087373
thats the most evil looking jewish person i have ever seen

>>60091184
what the fuck are you doing on /g/?
you're too smart and knowledgeable for this shithole.

>>60086040
>brb, buying a 400Hz monitor so I can AWP 12 year olds

File: LRM_EXPORT_20170427_072254.jpg (2MB, 766x8714px) Image search: [Google]

2MB, 766x8714px

>>60093373
>Measuring IPC in FPS
There really is no limit to how low shekelchasers will sink.

File: LRM_EXPORT_20170426_154238.jpg (2MB, 2539x1446px) Image search: [Google]

2MB, 2539x1446px

>>60093465
Try cherry picking your benchmarks less, shekelchaser.

>>60095070
desu for me 7700k is just the same shit ryzen 7 1700x is.
Only difference is that R7 might have better performances with the time since programs might be more and more multithreaded + optimisation.

>>60093373
According to your garbage benchmark 7600k is better than the 7700k in gaymen

>>60086074
>>60086106
>>60086176
>>60092052
>>60095070
>>60095065
Holy shit Intel is complete garbage at literally everything.

>>60093400
We don't know exactly what he meant yet, though. It's quite possible that he means that the theoretical throughput is higher, rather than that it exhibits higher IPC "in practical applications".

That interpretation is not at all unreasonable, seeing exactly as how Zen has more capable decoders (4 vs. 1+3), a wider fetch from the µop cache (6 vs. 4), much wider retire (8 vs. 4), and more issue ports (10 vs. 8). So that, on specifically tuned code, it may very well be able to blow Intel out of the water. But that doesn't automatically mean that it beats Core in code that hasn't been specifically tuned, or that doesn't exhibit inherently high ILP.