[ 3 / biz / cgl / ck / diy / fa / g / ic / jp / lit / sci / tg / vr / vt ] [ index / top / reports / report a bug ] [ 4plebs / archived.moe / rbt ]

Due to resource constraints, /g/ and /tg/ will no longer be archived or available. Other archivers continue to archive these boards.Become a Patron!

/g/ - Technology

View post   

[ Toggle deleted replies ]
File: 1.01 MB, 2160x1080, Screenshot_2018-07-27-07-55-12-050_com.google.android.youtube.png [View same] [iqdb] [saucenao] [google] [report]
66895902 No.66895902 [Reply] [Original] [archived.moe] [rbt]

>cannibalising your own product line this badly to stay competitive

>> No.66895934

>>cannibalising your own product line this badly to stay competitive

It's branding inflation.
i7 used to be top "prestigious" series, now it's a midrange CPU and all the enthusiasts will feel bad about having a shitty i7 and not a cool i9/

>> No.66895958

Given that I acquired an 8core/16 thread AMD processor for $170 and this will be nowhere near that price if not double this is a hard laugh for me.
5.00Ghz boost my ass @ 95W TDP

>> No.66896030

95w is only 8 cores at the base clock.
The 5.0ghz is only 1-2 cores at a time.

The power usage argument has always been used to favor people's bias as it suits their needs.
I don't care if it sucks 400w, if it's fast and can be cooled, it's all good.

>> No.66896039

What exactly are you laughing at? They look fine to me.

>> No.66896046

fucking lmao

>> No.66896053

I bet that the 5ghz boost is not continuous, but in burst, and limited to one or two cores.
No consumer cooler could handle 5ghz all cores.

Ps:$999,99+specific top of the line mobo only

>> No.66896060

I'll see what prices / performance these have when the benchmarks come out.

t. Non-fanboy with no plans to upgrade. Also owner of a 1950X.

>> No.66896061

You're also gonna pay more for that i7 than you're paying for the current i7, goy.

>> No.66896080

Of course it boosts up 5ghz only on 1-2 cores. So then it's possible the other leaks are true about 4.7ghz on all cores.

>> No.66896088

Don't forget that i9 can only work with the expensive Chipset, goy.

>> No.66896096

They can afford to loose money for some years just to make amd go bankrupt

>> No.66896127

4.7ghz all cores? In what scenario? 25c ambient+240 AIO only+fans full throttle

>> No.66896148

looks like ryzen 9 3900 with 16 cores will be a thing

>> No.66896191

Until it starts throttling. Procs are soldered now so it could be ok without absurd cooling solutions. Of course that assumes you shell out for the top tier Z390 motherboards to support that shiny new i9.

>> No.66896293

The total cost of a i9+mobo combo will make the epyc combo even more reasonable for the people that really need cores. But the i9 Will sell a lot for the "enthusiat" gaymen/streamer market, just for the status symbol alone. Intel now have the same sheep kind of following that Apple have.

>> No.66896439

>16 PCIe lanes
>Dual channel memory
>"""95W""" TDP
I don't think it's cannibalising the server market, perhaps the highend desktop line in power, but again not in IO.

>> No.66896504


>> No.66896517

>16 PCIe


>> No.66896530

>95w is only 8 cores at the base clock.
Intel actually matching their TDP numbers in real world power usage... kek
Yeah and my 8-core Ryzen has a 65W TDP and I can OC it reliably [ALL core] to a base of 3.7 at 95W TDP
> The 5.0ghz is only 1-2 cores at a time.
Not interested. Sound like a memory stall nightmare
> The power usage argument has always been used to favor people's bias as it suits their needs.
Given that I have a kikeripper that burns 250Watts at its current clock, I am not one such person.
> I don't care if it sucks 400w, if it's fast and can be cooled, it's all good.
Fucking neet that lives with his mum.

>> No.66896548

aka a fucking gimmick.
> Also owner of a 1950X.
My dude. Fellow owner myself
I have been an intel builder for a while. However, when Ryzen came out, they got dropped. Far better value and in a number of cases better performance and features

>> No.66896552

is the i3 getting dropped or relegated to Pentium duty?

>> No.66896565

> 16 PCIe lanes
These niggers will never learn.. hahahaha

>> No.66896568

Does an extra 4mb of L3 cache really make any difference?

>> No.66896609

>16 lanes
O I am laffin

>> No.66896639

16 lanes? What do you need 8 lanes for? Help police! This goy is trying to kill me over 4 lanes!

>> No.66896675

>gives more cores
>still only 16 lanes
>cpu to chipset is still 1 pcie lane
>a massive bottleneck fest this will be

intel 2018

>> No.66896696

16pcie lanes for my gpu, I didn’t need that msata ssd anyway. t-Thanks Intel

>> No.66896707

Is it backwards compatible with the 1151 socket?

>> No.66896710

>a socket change a year keeps the goyim in fear!

>> No.66896720

No, why would you want that?

>> No.66896742

>TDP 95w
Seems like Intel has misplaced a digit in the hundreds column

>> No.66896753

still gonna rape poozen in everything;)

>> No.66896763

you seem to misunderstand how CPU's work

>> No.66896773

When it says 16 pcie lanes does that mean you can attach 16 different things to it?

>> No.66896818


good boys change socket every year anon :^)

>> No.66896926

I don't. I most definitely don't for Intel CPUs :

1) Misconception 1: TDP represents max power rating
From the Intel Datasheets:
“The TDP numbers are not indicative of the maximum power the processor can dissipate under worst case conditions.”
Intel is listing TDP numbers that are significantly lower than the actual maximum power draw of their CPUs. They are then relying on the fact that most applications barely use the CPU, assuming that it will remain idle most of the time.

2) Misconception 2: Turbo-boost over-clocking is captured within the heat profile of TDP rating.
From Techreport on SB: The Turbo algorithm does something that may seem a little counterintuitive at first, allowing the CPU to ramp beyond its maximum TDP... taking advantage of the lag between when a relatively cool idle chip begins to warm up its environment.... the chip opportunistically pushes beyond its rated thermal peak by running at higher-than-usual frequencies within its Turbo Boost range. Once the surrounding system has warmed up or enough time has passed (the algorithm is complex, and Intel hasn't shared all of the details with us), the chip will drop back to operating within its TDP max... Turbo Boost algorithm incorporates not just the CPU cores but the IGP, as well; it can raise the operating frequency of the graphics processor when the CPU cores aren't at full utilization.

My Ryzen processors stick to their TDPs

>> No.66897020

do you just learn everything from news articles or are you just retarded? oh wait

>> No.66897093

Given that i've personally been building and OC'ing intel and AMD machines for well over a decade, I'd have to say its through personal experience and knowledge. Of course, when speaking to a brainlet like you, it helps to cite Intel's own data sheets. I dumped intel processors from consideration when Ryzen came out because its a better platform. That hasn't changed and its idiotic launches like this, gimmicks, PCIe lane gimping, security flaws, and outright lies that continue to solidify my choice...

I still own a number of intel systems because at the time I bought/built them, they were the best processors

>> No.66897094

Impressive if they get good performance. If intel wasn't absolute trash I'd buy this. But I'll stick with AMD.

>> No.66897140
File: 12 KB, 385x382, 4059.jpg [View same] [iqdb] [saucenao] [google] [report]

so you openly admit to being a know-nothing consumerist riding the latest trend, as you always have
seems I was correct all along

>> No.66897318


Inlel are finished

>> No.66897426
File: 15 KB, 250x250, 1445873750669.jpg [View same] [iqdb] [saucenao] [google] [report]

>((((((((((((((95 watts))))))))))))

>> No.66897439


>> No.66897442

I bet that the next line up, after this series, won't even have AVX2 due to the exploits that are starting to use it to leak data, even over a fucking network.


>The second measurement is novel and doesn't use the cache at all. Instead, it relies on the behavior of the AVX2 vector instruction set on Intel processors. [...] For this AVX2 side channel, the leak gadget is a fragment of code that speculatively uses an AVX2 instruction. The transmit gadget is a fragment of code that always uses an AVX2 instruction. If the processor speculates that AVX2 is needed then it'll start powering up the AVX2 units; this will make the subsequent transmit gadget run quickly. If, however, the processor speculates that the AVX2 code won't be used, the transmit gadget will take longer. These small performance differences are large enough to be measured over a network.

Where were you when Intel was literally kill in all fronts?

>> No.66897456

delid dis

>> No.66897463

>cool i9
But that's an oxymoron

>> No.66897467
File: 370 KB, 206x176, 1461226259654.gif [View same] [iqdb] [saucenao] [google] [report]

>They look fine to me

>> No.66897493

>if not double
Triple maybe

>> No.66897505

> tfw the i7 is priced higher than Ryzen 7
> 8c/8t
> tfw the i9 is priced in Threadripper territory
> 8c/16t


>> No.66897534

the i3 is fine, 4c/4t, will probably get a clock speed bump.

>> No.66897580

Everything 9xxx and HEDT is DoA since Threadripper 2 is a thing

>> No.66897612
File: 2.57 MB, 491x300, unnamed.gif [View same] [iqdb] [saucenao] [google] [report]


>> No.66897658

16 PCIe lanes
in 2019


>> No.66897676

shut up goy

>> No.66897697
File: 15 KB, 200x300, 1494420017221.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.66897843

These chips are pointless and a waste.

>> No.66897896

>AMDrones working double shill duty because they're going to get destroyed again and Ryzen 2 is nothing more but a wasted launch

>> No.66897941
File: 54 KB, 400x304, 1521676938108.jpg [View same] [iqdb] [saucenao] [google] [report]

>lga1151 (v2)


>> No.66897976

>to stay competitive
Intel isn't in any danger in the market that matters.

>> No.66897994


9900K is going to be memory bandwidth starved. Dual-channel DDR4 doesn't cut it for workloads that actually need eight cores. I suspect it is part of the reasony why 9700K doesn't have HT. I also suspect that 9700K is going to be 95% of the 9900K in overwhelming majority of workloads again due to memory bandwidth issues.

9600K = rebranded 8600K

>> No.66897995

The i7 and i5 can be used with the same LGA 1151 board, assuming your mobo manufacturer does a bios update

The i9 is different and requires the new Z390 board

>> No.66898021

They'll probably have to sell at a loss to match AMD. Intel has to develop a good interconnect and fast or they're done for

>> No.66898069

Intel will never price themselves low

At best they'll keep the i9 below 400 dollarydoos and that's a stretch

>> No.66898084


It is what Socket 1151 for 3xx boards should have called from the start.

>> No.66898086
File: 49 KB, 285x285, leff pez.png [View same] [iqdb] [saucenao] [google] [report]


>> No.66898098
File: 41 KB, 512x433, 1527000909230.png [View same] [iqdb] [saucenao] [google] [report]

7nm when already? CHADMD needs to deliver the finishing blow

>> No.66898115

8/16 CPU that overclocks to 5.5Ghz with soldered IHS and actual AVX512 support, the 9900K is going to be a beast. AMDrones seething.

>> No.66898146


> Only in Suicide Run with LN2/Phase-Cooling + Water pump

>> No.66898198
File: 112 KB, 691x771, 1499963016439.jpg [View same] [iqdb] [saucenao] [google] [report]

delid dis thread

>> No.66898236
File: 132 KB, 800x554, 1530708699933.png [View same] [iqdb] [saucenao] [google] [report]

>16 PCIe lanes

>> No.66898299

>so you openly admit to being a know-nothing consumerist riding the latest trend, as you always have
How is this an insult?
A common term in tech is "latest and greatest". Nothing wrong with following the latest trend in tech.

>> No.66898310


>> No.66898323
File: 72 KB, 1749x339, delet dis.png [View same] [iqdb] [saucenao] [google] [report]

>i9 9900K
>16 PCIe Lanes
>price: probably $3000
intel on suicide watch

>> No.66898353


>> No.66898384
File: 29 KB, 300x169, 2.jpg [View same] [iqdb] [saucenao] [google] [report]

What? No new socket? This must surely be fake news.

>> No.66898403




>> No.66898428
File: 58 KB, 831x1200, tb,1200x1200,small.2.jpg [View same] [iqdb] [saucenao] [google] [report]

>Memory channels: dual channel
are they ever trying anymore?

>> No.66898438

Anyone who buys this instead of waiting for a Zen 2 is a fool and supporting a greedy corporation with a horrible anti-consumer track record.

I just wish AMD or anybody would do the same to nVidia.

>> No.66898583

I openly admit I have a CE degree and am more than qualified both in formal education, my formal career, and in my ten years of building desktop computers in which I go towards a company with the best offerings and best value. Meanwhile you're a jackass making comments that defy logic, facts, and intel's own data sheet publications.

You are correct. However, I'm not the dumbass in the conversation. (YOU) are. Why do brainlets always resort to personal attacks w/ no basis when they're defeat? wait.. nevermind because they're brainlets who can't form arguments.

>> No.66898600

Intel is a complete write-off until they get their shit together and lower prices.
8 core priced in threadripper territory. The absolute state of Intel
> 16 pcie lanes (shintel)
> 64 pcie lane kikeripper

>> No.66898614

Agreed. This is how big tech giants fail and shrink down to reality.
Welcome to provide sound arguments and facts against the tide of substantial negatives

>> No.66898623
File: 627 KB, 782x782, oy vey the goyim know.png [View same] [iqdb] [saucenao] [google] [report]

mah nigga

>> No.66898649


>> No.66898659
File: 246 KB, 882x758, 1489100922662.png [View same] [iqdb] [saucenao] [google] [report]

intel is finished

>> No.66898676

Famous last words of a tech giant
>> Dual-channel DDR4 doesn't cut it for workloads that actually need eight cores
Top fucking kek. Sorry jackass, you're wrong

>> No.66898689

>16PCI-E lanes
>95 watt

I don't think so tim.jpg

>> No.66898694

People are really talking like AMD got back on track.

But it feels like the mirror is actually true: Intel lost so many chances to remain ahead. It's the Pentium 4 era all over again. AMD just had to not fuck up completely a single generation.

So it's not like AMD is doing great, they just started doing their homework. Maybe in two generations they can make a good leap like Intel did from the Core 2 to the first gen of Core iX...

>> No.66898743

Nah, intel is launching the very same CPUs they were succesful with in the past, using the same socket change a year move for years, and only increasing core count after Ryzen showed up. Intel isn't fucking up, AMD unlocked infinity fabric and are making bank with that while Intel stagnates.

>> No.66898780

I wonder if Intel will be able to recover once 10nm will come out or it will be AMD staying ahead with new releases.

>> No.66898781
File: 158 KB, 1280x530, btfo.jpg [View same] [iqdb] [saucenao] [google] [report]

>> No.66898799
File: 31 KB, 300x295, 1352402985280.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.66898810


Well, I talked more about the pure "computational progress" aspect rather than the business. As in, their choices are quite suboptimal regarding increasing the computational power of their offerings.

It's quite a given that Intel didn't "need" to advance after making AMD eat dust with the Core iX while AMD derped hard after the Phenom II, as they are capitalists.

They fucked up progressively in a scientific sense.

>> No.66898847


>> No.66898906

>I can overclock to 3.7
Wow nice one, I remember I had the first i7 core generation I used to clock 4ghz.
Stay slow amd

>> No.66898915

Well, I've waited this long... Looks like I'm not missing out on anything. Ryzen2 will be my next build by the looks.
>t. 2600k owner

>> No.66898927

>i9 below 400

Lmao.Not even in dreamland pal.

>> No.66898969


>> No.66898983

When will we finally have an i11?

>his branding doesn't go up to 11

>> No.66898998
File: 67 KB, 181x201, 1500544851892.png [View same] [iqdb] [saucenao] [google] [report]


>> No.66899017

>make amd go bankrupt
I've used Intel chips for 31 years but I'm building a Ryzen computer once their 32 core chips come out later this year. Less expensive by a lot and won't be surprised if it blows out the i9 in performance on several CPU tests. Long term those 32 cores are going to be so much better to have than 8 with 2 that get "boosted" using some gimmick that isn't even proven to be beneficial.

>> No.66899027
File: 21 KB, 563x503, 1530903799953.jpg [View same] [iqdb] [saucenao] [google] [report]

burned just like what the i9 is going to do to the inside of it's owners computer case

>> No.66899038

Just a single graphics card already needs 16 lanes for top performance.

It will mean starving your graphics card to 8 lanes, so you can attach one SSD (4 lanes) and maybe 1 or 2 more cards.

>> No.66899040

AMD got back on track and intel arrogant stayed off track and kept trying to nigger the market in the face of it. Ryzen is a decade+ architecture plus they're leading using the industry's best fabs for die shrink. There's literally no competition for a decade since Ryzen arrived. The industry goes through these changes all the time and there's always a group of deniers who eventually jump on the bandwagon once its in its late stages.

I wouldn't go near a fucking intel processor or platform either in a server or desktop environment. They've lost me for 10 years. And yes I have a history of intel rigs. The only AMD I bought was the athlon XP back in the day (when it was better value at the time than intel offerings) Then I was back on the Intel train when AMD fucked up. There's no reason to be loyal to either. The best performance/value gets my $$$'s its as simple as that and intel has jackshit in the pipe to win me over and it seems they are going to arrogantly keep their prices in the stratosphere throughout this epic decline.

Even when AMD threw them a boon and let them license Vega for their APU, they still fucked it up w/ nigger tier pricing. No one's buying that barebone skull nuc for fucking $900.They're literally nonredeemable. I don't even think that shit has linux support yet.

They literally need to cut their prices in half. Since they will never do that, they're fucking DONE

>> No.66899047

it'll be faster in ST and MT for 7 months or so. then AMD will be back on top of MT. I like this

>> No.66899069

>My Ryzen processors stick to their TDPs
Have you actually measured it with anything than a shitty software?

>> No.66899076

I have an older Intel rig that I game on that has a higher stock clock and boost rate. It's 85W TDP (STOCK) However and is a 4 core / 8 thread processor. My ryzen kicks the dogshit out of it. It stopped being about clock rate some time ago. I'm guessing you never heard of memory stalls either in that you think your CPU can have such an insane clock rate w/ no matching DRAM access rate. Also, Im not a poorfag fanboy like you so I actually have multiple rigs. Some are intel some are AMD. The best tool for the job my dude.
> mfw kikeripper box is sitting right next to a housefire shintel box
> mfw I shit on intel but still have intel rigs
> mfw I have AMD rigs
Poorfags will never understand this.

>> No.66899089

>I'm a meme engineer
Appeal to authority here means shit
Give us some real measurements from hardware.

Anyway I know amd draw less power, it shits in pants anyway to reach near 4ghz

>> No.66899102

How much lanes Ryzen have? 9 series aren't hedt

>> No.66899106

>100mhz more coming right up.

>> No.66899111

BASED future kikeripper owner.
Welcome to the club my dude.
I'm going to forgo the 32 core kikeripper. From a technical stand point 2 dies or all 4 will be memory starved and/or 2 dies are potentially going to have zero wiring to RAM or PCI-E. They're going to bill this as intel's older 2 socket config where one socket has all of the I/O but I'm not buying such a shit tier NUMA config in this day and age especially when EPYC is properly wired. The 16 core kikeripper makes sense and I'm sure ryzen will get a core bump at 7nm but then they need to bump the DDR channels IMO. I'm comfy for now. I see the upcoming gen as a skip phase. I'm waiting on Ryzen 2. Comfy with Ryzen 1 until then.

>> No.66899125

> never heard of memory stalls eith
I know this an issue in amd shitty products that ties infinity glue™ to dram mediocre frequencies, while I can easily run ddr4 4000 on my 8700k.

>> No.66899127

Yes. I have a monitored PSU, UPS and clamp meters... Jackass
Your lazy ass is welcome to view the copious amounts of youtube videos that use even further probing for power util. What did you just crawl out from underneath a rock?

>> No.66899156

It's not an appeal to authority. It's a correct for the dumb faggot who claimed I had no qualifications on the matter (not that I need to in order to link to a dam intel data sheet that says exactly what I said). The real measurements have been done over years and are documented all over the interweb. The fact that you have no clue what I'm talking about regarding Intel's common real-world power utilization figures being far above their published TDP figures shows you're a fucking brainlet who hasn't been in the computer building scene for long.

> Anyway I know amd draw less power
> And actually doesn't kike its TDP figures and its processors match them
Then why are you talking faggot?

> it shits in pants anyway to reach near 4ghz
Clock rate jackasses.. Keep being shintel heating up your mum's house eating up the power bill.

>> No.66899163

Sad part is Intel is still going to dominate for gaymen and AMD will once again be relegated to counter-culture nerds on /g/ who do nothing but shitpost with 50 tabs open

>> No.66899170

I really believe you, with such shitty frequencies of zen, uless you try to force 4ghz it stay low on power.

>> No.66899181

i have a skylake motherboard with the 110m chipset, do i need a new mobo for 9600k?

>> No.66899184

>Remove HT from i7
>6C/6T to 8C/16T 14nm are all same 95W TDP

>> No.66899201


>> No.66899202

>Clock rate jackasses
Really sad when your CPU can't even break 4.0ghz without 1.6v

>> No.66899213


>> No.66899215

Truth > memes

>> No.66899238

> I can easily run ddr4 4000 on my 8700k
At what timings? 19-21-21-41
LOL grasping at straws this hard.
I guess you didn't get the memo pleb, intel's shitty architecture doesn't expose the internal bus to the RAM timing so you literally get no performance boost above a certain threshold :

In some cases, you get worse because of the nigger tier timings you have to run at to get those RAM clocks

I run high clocks on all my ryzen rigs w/ tight timings 14-14-14 because I don't buy shitty ram after being bankrupted by shintel pricing. The internal memory controller runs at the RAM clocks and scales excellently w/ higher clocked ram.

> mfw some intel dweeb tries to argue about RAM clocks and realizes his shitty processor architecture doesn't even take advantage of it compared to a ryzen proc.

What other poorfag intel fanboi has a comment?

>> No.66899282
File: 11 KB, 299x168, images (17).jpg [View same] [iqdb] [saucenao] [google] [report]

Lol ring/cache is running @ 4.5 GHz here pleb.
Anyway, you already said you can't even break 3.7, that for sure won't stall memory unless you are using zenfag certified memories

Just admit you are shopping in the bargain bin

>> No.66899295

> muh magical comeback story that isn't reflected in roadmaps or their increasingly reduced market share in gaymen
Ryzen has been dominating gaming rigs for some time. Steam and others publish figures regularly. Time to wake up from your dream.
> Frequency vs performance
> housefire power utilization and temps that increase exponentially with diminishing returns
Stop letting everyone know you live w/ your parents. Let me give you an inside tip Gen-Z fag...
> Clock rates are limited by transistor size and physics.
> Power utilization goes up exponentially towards the physical limit
> 7nm is where the clock rates have room
> Most modern CPUs do a better job at Overclocking via intelligent smart boosted baked into the hardware
The hobby of overclocking was big when none of these features/limitations impacted processors. You're literally a nostalgia fag.

Why are intel brainlets so easy to shit on?

>> No.66899326

Really sad that you know nothing about computer architecture or real world performance.
> Gaymen faggots and their house fire clock rates
> Meanwhile 32 core server CPUs run at 2.0-2.6Ghz and cost 10x more.

Not everyone spends their lives glued to a gaymen screens fapping off to FPS numbers that their physical brain can't even perceive.
> Muh 4k gaymen
> Muh 180 FPS
> Eyes can barely process 90FPS in high action scenes

Dumbfags will never learn

>> No.66899349

i get because you need 24/7 16 cores running who knows what parallel load (at lower speed)

>> No.66899352

>muh magical comeback story that isn't reflected in roadmaps or their increasingly reduced market share in gaymen

>> No.66899384

course not. Some coworker asked me what cpu he should buy for gaming. I said once you are getting 60fps the logical thing is to use zen, the overall cost is much smaller. Then he said "what about intel"?

Im in the process of finding a new 4 cores laptop, since i dont need ultra frequencies but a nice frequency + 4 cores, i was prone to get one ryzen, especially considering how much meltdown toil has on performance, the thing is, i cant find, barely anything zen laptops. Even in US the options are severly restricted. Same as with retail servers on dell. For the majority of the market, its like zen doesnt exist.

>> No.66899412
File: 256 KB, 802x816, FUCKING_BRAINLET.png [View same] [iqdb] [saucenao] [google] [report]

> Doesn't own a ryzen rig
> Doesn't know you can run DDR4 @ 3200 14-14-14-34
> Doesn't know you can run it @ 3600/3466 at CL 14/16
which is the peak operating point of RAM clock/timings

> Lol ring/cache is running @ 4.5 GHz here pleb.
Doesn't know fuck all about his intel's internal microarchitecture or how memory clocks work. Openly admitting this like a pleb because you don't realize what you're saying. I don't even think you understand that DRAM clocks are half of what is stated.

Ending this shit tier exchange :
> you don't know what memory stalls are
> you're oblivious to how frequently they occur
> You don't know that the advertised clock speed of RAM is double what it actually runs at. It's called DDR because it operates on both the rising and falling edge of th eclock
> Doesn't know anything about his intel's micro-architecture in that the internal clock for the memory controller is not tied to his ram

Go to sleep.
It's your bed time.

>> No.66899449

ryzen will never hit 5ghz

>> No.66899460

Oh, right, because you need 16/32 C/T running at 4.0ghz to play fucking call of duty.

>> No.66899464

You get nothing. Every comment you've made has been corrected and shat on.

Revisiting what I said :
> Most modern CPUs do a better job at Overclocking via intelligent smart boost baked into the hardware
The 16 core processor smart boosts cores when others are dormant. When the other cores are not in use, they hardly receive any power (modern day CPUs). So, there's absolutely no loss to a high core count HEDT processor. I game on a shitty ass intel 4 core processor w/ a dated ass nvidia card and it plays most gaming titles fine.

I recall what it was like to be young and feel like I am a master computer Guru because I can set some numbers in bios and overclock a processor. Those days are gone. Both Intel/AMD slam their CPUs to the wall straight out of the factory using intelligent-boost hardware/firmware. You'll never be able to more efficiently boost clocks anymore since these features rolled out. The processors are also binned to shit especially shintel. So, you're doing nothing special w/ clocks.

It's literally nostalgic faggotry that people engage in to feel connected to their hardware. I grew out of this stupid shit when I realized that the processors come out of the factory slammed.

>> No.66899473

AMD has 0 laptop presence because they have no ultra low power components with any amount of performance in them.
Meltdown and Spectre are not even a remote concern for you, or me, or anybody else that isn't running VM cluster farms. Just get an ultrabook and be done with it, if you're not compiling giant C++ applications or doing actual computational work on it you won't notice the difference.

>> No.66899481

I'm on an intel based laptop right now and have several intel rigs + amd. For low core count (4 or less) non-custom rigs, intel is fine. For 6+ I go w/ ryzen. Laptops are not that serious to be worrying about what processor is in them.

>> No.66899496

And nothing of value is lost in an age where system memory continues to be the bottleneck
You dumb faggots will never learn.

>> No.66899504

cpu is the bottleneck for my daily applications. sorry but you can rub your dick over ryzen/gnu/linux all you want, but in corporate environments, intel and windows is king

>> No.66899580

Honestly, 5GHz is just too hot in the current technology available. I'd rather have more cores at 4GHz

>> No.66899596

> cpu is the bottleneck for my daily applications.
Memory is fundamentally the bottle-neck in present day architectures (period) as is PCIE 3.0. If CPU is your bottleneck, it's time to stop running your tasks on poorfag rigs.
> sorry but you can rub your dick over ryzen/gnu/linux all you want,
I'm just stating facts you don't seem to be able to stomach
> but in corporate environments, intel and windows is king
Different set of priorities.
Number one : The IT/SysAdmin doesn't want to get fired in the event something occurs and its blamed on hardware that the shit brain CTO is unfamiliar with
Number two : Support contracts (which have nothing to do w/ the performance/value of a processor)
Number three : Relationships beyond your paygrade
Number four : Developer support beyond your paygrade

You can grasp at straws all you like. It just makes you seem even more pathetic. You admit you got blown out for the desktop now you're trying to go even further out of your knowledge base into the enterprise eco-system. You don't have a leg to stand on and that's clear.

Do what you want w/ your money but don't come up w/ bullshit arguments as to why you do what you do.

>> No.66899611

5Ghz fags are nothing but shills. Running that at 12nm is beyond retarded as is water cooling to achieve it. It's literal brainlet hobby tier bullshit that companies milk stupid fags for their highest margins

>> No.66899621

lmaoing at your life

>> No.66899632

I wonder if Ryzen 2 will hit 5GHz at 7nm tho.

>> No.66899720

i'd be fine with base clocks of 4Ghz.
It's been about core count for some time.
That and I'm looking more forward to PCIE 4.0 and a possible innovation in system memory. The key IMO was to lockin and enjoy the bump from 14nm to 12nm DDR4/PCIe 3.0. Nothing is entertain-able until 7nm/PCIE 4.0 and a major change in system memory and storage. 8 core is a clear standard at this point w/ higher core counts for more serious tasks. Whatever the clock rates are .. they're set by physics and temperature dissipation. 5Ghz and other silly ass figures are a thing of the past when core counts were low or there were gimmicks.
> PCIE 4.0
> HBM 2.0/3.0/4.0
is where things get interesting.

The majority of what's going on right now are iterative milking steps. Even the new Nvidia cards this year look to be milk. I'm not touching that shit until its 7nm. Comfy on 14nm GPUs until then.
> PCIE 4.0
> 7nm
for GPUs

I don't think a lot of anons have been building for a while but its always retarded to get milked on the small bumps. You build and ride it out for 5-8 years

>> No.66899746

>AMD up literally 10x from where they were like a decade ago
>go bankrupt

>> No.66899776
File: 90 KB, 1200x654, 1524454497419.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.66899884


>> No.66899902

I don't see anything wrong here

>> No.66899936

yes it is

>> No.66900008
File: 12 KB, 215x211, 0o.jpg [View same] [iqdb] [saucenao] [google] [report]



>> No.66900103

I will say this again:
> 8c/8t > 6c/12t

AMD is still ahead tho. Threadripper will rape that i9.

>> No.66900174

Man, intel didn't just get caught with their pants down, they got caught jacking off to gay porn with a massive dragon dildo up their asshole. They have nothing, and now it's clear that their performance lead was based on shoddy work and tradeoffs against security, they deserve to suffer.

>> No.66900447

that's the consumer i9
later on they're releasing that cascade lake X abomination

>> No.66901061

Ok you wanna talk about stalls, what about the shitty latencies on zen cache? No even talking about accessing l3 that is in another ccx with latencies all over the roof.
Skylakex is another shit using mesh for 10-12 cores, it is somewhat better considering the latency across a large number of cores. Epyc may implement 4 channels but it's still a pos if you need to address memory that is across infinity fabric latency goes over the roof.
Zen Arch is good if you need a lot of cores and can drive break your load in several numa domains.
Meanwhile, Intel ringbus is the fastest interface for 8 or lass cores, there isn't even comparison

>> No.66901950

I'll let you in on a secret faggot.
See that mesh layout of cores at Shintel. The latency literally depends on which core is attempting to communicate with another core. If its the top left with the bottom right, the latency is worst. Intel's numbers don't cover the jackass latency issues that coincide with this routing mechanism.

> Zen Arch is good if you need a lot of cores and can drive break your load in several numa domains.
NUMA is a new norm. Especially given how fast I/O is becoming.
Mesh is more suited for GPUs where you're doing highly parallel non branching compute
Intel made an epic mistake in their CPU architecture.
> Intel ringbus is the fastest interface for 8 or lass cores, there isn't even comparison
8 cores is the new MINIMUM. So, there goes that.

>> No.66902955

AMD won't release 64 core CPUs simply because Intel can't compete.

>> No.66903042

24. 16 for gpu, 8 for other things.

>> No.66903051
File: 120 KB, 299x299, 1529312192947.png [View same] [iqdb] [saucenao] [google] [report]


>> No.66903085

>cascade lake X abomination
Litreally DOA, what's the point?

>> No.66903091

>16 pcie lanes
Why is this still a thing?

>> No.66903092

They still have the mindshare. Skylake-X was a disaster and broadwell is pretty old so customers still need a replacement.

>> No.66903099
File: 30 KB, 202x249, 1436734391281.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.66903105

Because PCIe is upgraded on time and consumers normally don't even use full bandwidth. 8x PCIe 3.0 is enought to power modern GPUs and another 8x is enough for everything else. PCIe 4.0 arrives next year so next gen GPUs will be fine with 8 lanes too.

>> No.66903109

Yeah no >>66896696

>> No.66903160

Let me guess: Pretending?

>> No.66903209

ok ram-boy

>> No.66903225



How is this even plausible.

>> No.66903237

oh god, did they just disable hyperthreading and call that a new chip? oh nononono

>> No.66903253


>> No.66903257
File: 1.13 MB, 320x240, 1509997189004.webm [View same] [iqdb] [saucenao] [google] [report]


>> No.66903259

You can't bin wattage.

>> No.66903279

discount cerberus

>> No.66903344


>> No.66903354

If you want more lanes buy HEDT (broadwell-X or threadripper).
If you want more lanes it doesn't actually mean you need them.

>> No.66903368

Why not?

>> No.66903465

>1151 v2
So it seems there's absolutely no change in these cpus other than the name. Which means that they will be compatible to current mobos, but also means intel now tries to sell you the same cpu for more, hoping that it can boost itself to specs without melting.

>> No.66903522

good goy buy HEDT to unlock moar lanes

>> No.66903523
File: 359 KB, 1300x599, 1497918643725.png [View same] [iqdb] [saucenao] [google] [report]


>> No.66903545

>5GHz 16 thread
What's wrong with it?
Intel is fulfilling their promises.
28 core 56 thread 5GHz CPU
Amdrones on suicide watch!

>> No.66903555

Because only analog tech can go to 11

>> No.66903568

Same thing with Ryzen 2 then

All those talk of mobo compatibility and AMD still shills you to buy the 470 boards

>> No.66903579

No, if you don't need more lanes don't buy HEDT. Can you be any more dumb?

>> No.66903598

If you assemble a new machine there is no reason to buy an outdated mobo. Are you retarded?
Compatibility is there, if you have an older mobo you can use it. Honestly intel shills trigger me so hard with their stupidity.

>> No.66903607

i15 when

>> No.66903613

X370, and even B350, works perfectly fine with 2xxx, you dumb Inturd-gargling faggot.

>> No.66903658

You're talking to AMD enthusiasts. Idiocy is a given.

>> No.66903664

I call on your bullshit
H310 boards will run it without overclock.

>> No.66903676

You're talking to an AMD enthusiast too, senpai.

>> No.66903707

There is a Z390 board

That board is for the i9, everything i7 and below will run on 1151 boards with mobo manufacturer bios update

>> No.66903780

Every modern GPU runs fine with 8 lanes but I'd still rather have the full 16

>> No.66903839


>> No.66903891

Apparently there's no SSD that uses more than 2 lanes.

>> No.66903897

APUs have 16 though

>> No.66904166

All is good at Intel, all is fine. You know nothing. See

>> No.66904431

of course you can, the best parts will be 95W with 8C/16T while the worst parts can be 125W 8C/16T or whatever, I don't know the range for their process.

>> No.66904458


Keep telling yourself that kiddo. Not everyone specs a desktop/workstation computer for silly shit like gayming and facebook.

>> No.66904872

I thought we were talking about many core parts.

>> No.66905051
File: 39 KB, 625x415, 2f14e1f1c56ae4a49f8dee19b6a78022bae9c298ec9357fbc32e34c30d4c51bd.jpg [View same] [iqdb] [saucenao] [google] [report]

This actually makes me concerned. If intel terribly fails, what's stopping AMD from ripping us off?

>> No.66905053


>> No.66905065


>> No.66905079

Gaming requires some of the most powerful hardware out of any common task. Actually the most.

Why are you newfags on this board so retarded?

>> No.66905291

nGreedia will always be there, don't worry.

>> No.66905312

Intel's mindshare.

>> No.66905347

Right, more cores means more MHz, hence more speed in games.

>> No.66905377

nvidia doesn't make cpus though

>> No.66905394
File: 26 KB, 605x377, 2990x.jpg [View same] [iqdb] [saucenao] [google] [report]

1950x master race here, upgrading to the 2990x when it drops, lol'd heartily, thank you OP

>mfw 32 cores and 64 threads

>> No.66905401

>Network connection establishment
>"With all mitigations enabled, the mainline kernel is slowed down by approximately 15 per cent without and 21 per cent with the User-Based Firewall.
>The GRSecurity-enabled kernel is also slowed down by 15 per cent without,
>but 67 per cent with the User-Based Firewall."

Who needs hardened secure routing VMs anyway goys

>Disk access
>"With all mitigations enabled, the mainline kernel is slowed down
>by approximately 50 per cent on local disk
>and 33 per cent on Lustre.
>The GRSecurity-enabled kernel is slowed down
>by 90 per cent on local disk
>and 33 per cent on Lustre."
Who needs disk IO anyway goys?

>Computationally intensive code – Kernel changes had little effect, "as they make few requests for kernel services". That's the good news.
>The bad news? "A noticeable slowdown was seen with the microcode updated...
>These slowdowns were measured at 21 per cent for pMatlab
>and 16 per cent for TensorFlow for the baseline kernel with all mitigations
>and 19 per cent and 15 per cent respectively with just the microcode."

You know those $40mil contracts you had with AWS with 2% overhead factored in? You're now 20% short of raw compute power required.
Pay up goys.

>the absolute state of intel and their customers 2018

>> No.66905421

so intel will match 2700x 8C/16T with core i7 8C/8T? this is stepdown in multithreaded application.

>> No.66905500
File: 95 KB, 360x360, the-brain-is-the-most-important-organ-according-to-the-brain.jpg [View same] [iqdb] [saucenao] [google] [report]


By Intel's standards, yes, 95w. Only measured at stock non-turbo clocks, that is.

>> No.66905526

>H310 boards will run it without overclock
One ligttle NIKO MOSFET two ligttle NIKO MOSFET

>> No.66905592

Yeah, Intel might have finally realized forcing people to change mb's will bite them in the rear. However:
>got my pc for cheap, owner said it was nowhere near as powerful as he thought it would be
>odd, a 4790 and 16gb ddr3 at the time shouldn't be slow by any means, then I saw the el cheapo 30€ H81 mb it was mounted on
Needless to say, with a 4 core non-k cpu the VRM's were already throttling out the ass. I shudder at the thought of the same type of board handling an 8 core.

>> No.66905795

novidiots don't have brains,
they don't know that.

>> No.66905902

They have actual serious plans on start doing it soon. Been planning it since as far back as 2015.

>> No.66906144

Not getting the check my repeating i111

>> No.66906196

Their headquarters aren't in Israel.

>> No.66906252

this desu

>> No.66906277


>> No.66906373

The 9700K will beat the 2700X in every benchmark while consuming less power.

>> No.66906413

kek <- this one

>> No.66906423

No it's 95W at all core turbo, intel CPU's have a power limit set you can disable in the bios.

I have a 8700 and it shows 95W while under aida 64 stress test 4.3GHz on all cores

>> No.66906472
File: 67 KB, 572x600, 0a9843_6397243.jpg [View same] [iqdb] [saucenao] [google] [report]

>when Intel makes cpus so bad a switch-level fix will probably be created slowing down the entire internet instead of holding the kikes responsible

>> No.66906481

>Thermal Design Power (TDP) represents the average power, in watts, the processor dissipates when operating at Base Frequency with all cores active under an Intel-defined, high-complexity workload.

>> No.66906561

So you really concede you shitty zen CPU have much more memory stall at cache level

>> No.66906586
File: 86 KB, 840x864, Capture1.jpg [View same] [iqdb] [saucenao] [google] [report]

See Turbo Power Max
The CPU isn't allowed to draw more than 95W and still reaches all the boost clocks.
There are also test's with clamp meters showing not more than 105W drawn from 8Pin which translates to 95W due to VRM inefficiency.

The reason they say base frequency is for a worst case scenario, if someone get's a really bad binned chip.

>> No.66906606

This i9 is planned to compete against who? Threadripper?

>> No.66906685

Keep dreaming , The Product is been availible for the last 4 years . THE BENCHMARKS ARE OUT THERE.
Go look and see the power consumption on any recent 8 core intel CPU with HP.
Enjoy the 220watt to 270W for 5ghz everything.

>> No.66906697

You already know Pentium Silver and Pentium Gold, but now there's a new successor: introducing Pentium™®© Platinum©®™

>> No.66906740

You can't even read the Image you just Spent..
I can see short burst of 135W... RIGHT THERE

Go read some reviews

You will not sustain the all core turbo AT 95W unless you consume more POWER ...

Frequency = power consumptiong ...
Higher Frequency = more power consumption
UNLESS you have a new process aka new transistor.
Which it is not running on anything... therefore enjoy your 150W 6 core

>> No.66906776

It pretty much depends on a mobo and settings. Shitty mobos from hp prebuilt machines will stick to the base TDP level while high-end motherboards will output as much as their VRM can.

>> No.66906818

That's very true but the common idiot thinks that's the TDP on the box is what the CPU consumps at turbo range.
And Intel only advertise single core turbo ...
So they assum the single core turbo applies for allcore which it does not ...

>> No.66906939

The turbo boost short power is like 5 seconds, and it's just an additional measure for stability.
>UNLESS you have a new process aka new transistor
There is a binning process involved, here, just because the 9900K will have 2 more cores of the same architecture doesn't mean it needs more power. It will just be better binned parts.

Look how the 8086K runs much cooler than the 8700K even though they run at the same all core turbo. Only because it's a better binned chip.

>> No.66906944

Imagine doing nothing but gaymen in your PC

>> No.66906978
File: 72 KB, 589x534, 2a3.png [View same] [iqdb] [saucenao] [google] [report]

How can you possibly be more Jewish than this?

>> No.66906999


>> No.66907096
File: 161 KB, 1262x826, 1523874112611.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.66907118

Let me add what happened with my 8700k CPU on z370 Asus hero.
First I overclocked the CPU to all core 5ghz. It was really setting 5ghz across all cores, however, it was strange it was keeping 50c across all cores. I started prime95 and it kept 50c, so something was up. The frequencies were all bouncing on cores between 4 and 5ghz even with raised max multiplier.
Then I found out, by default, bios set the current limit for CPU to tdp, so by default, it can only keep those raised frequencies for a short burst, when it dial back to hold tdp values.
Tldr, tdp is achieved with current limit. This means you end up with 8700k around 4.3 GHz across all cores for 95w. If you remove all current limits it tops around 130-140w on stability tests, and on prime it goes around 160w on my tests.

>> No.66907120


>> No.66907172
File: 1.44 MB, 2298x2000, 1a2243d5238d48094ae566297bfddcbe.png [View same] [iqdb] [saucenao] [google] [report]

>one let's you use the same socket for three generations
>other changes with each release even thought it's still Broadwell
>same thing

>> No.66907234
File: 260 KB, 659x800, da0593a6a0c68440ae06cc7d2582abc7.jpg [View same] [iqdb] [saucenao] [google] [report]

Why isn't the product line as follows:

Celeron 9000K
Pentium 9009K
i3 9090K
i5 9099K
i7 9990K
i9 9999K

>> No.66907291

It's arguably shit to change sockets so often, but Intel used until now if nothing changes, to have free design over generations without considering socket limitations
When you see 1151 v1 it was delivered in a time for skylake when 10nm CPUs was expected to be delivered soon.
What we see on 1151v2 is increase in power delivery pins, when clearly the 1151 plataform wasn't mean to delivery the power required for 6-8 cores CPU.

>> No.66907371

Is that kelvin?

>> No.66907402

Will probably do 5 GHz after delid at 300+ W consumption
While AMD brings a 12C/24T up to 4.6 GHz turbo
>TFW there will be Intelfags with "muh 400 MHz more"

>> No.66907425

No that's the number of kikes shilling Intel on /g/ daily.

>> No.66907643

No. They literally renamed i7 to i9 and i5 to i7.

>> No.66908199

And there will loads of people buying i7s just because of the name. Its all just marketing. I guess the i5 and i7s will be fine for most gaymers.

>> No.66908216


>> No.66908280

>a change in CPU architecture needs a new socket
get fucked, retarded shitter

>> No.66908328

hits too close home

>> No.66909353

9900k will surely beat the 2700x, so Intel will be at the top, but for what purpose? No way are these going to be competitively priced with how expensive the 8700k is. They also would need solder if they want to be close to those clockspeeds.
Will AMD respond with a 2800x or just wait until Zen2 to fuck Intel over?

>> No.66909364

>16 PCIe lanes
This can't be real

>> No.66909423
File: 141 KB, 439x290, 1455687281173.png [View same] [iqdb] [saucenao] [google] [report]


>> No.66910617

>but for what purpose
9900k will most likely be priced around $500

>> No.66910723

>but for what purpose
To maintain the halo of the fastest gayming CPU.

>> No.66910755

You know, this is the time that one would think intel finally play fair and puts out something good for an acceptable price to be competitive. But nope, they rev up the jew to 11.

>> No.66910792

nobody needs more than 4 pcie lanes

>> No.66910819

You had better enjoy that 16x slot wired up to be 8x and the rest spread across the other slots and M.2.

>> No.66911628

>imagine being this delusional

>> No.66911636

Watch out german bro, they are going to send you to thought crime prison for that.

>> No.66911682
File: 24 KB, 240x300, 1511303367638.jpg [View same] [iqdb] [saucenao] [google] [report]

No semites -> no antisemitism.

>> No.66911864


>> No.66912000

>28 core 56 thread 5GHz CPU
the incinerated state of intel

>> No.66912283

Lmao who cares? 7nm Zen 2 early next year.

>> No.66912394

explain. i dont get it.

>> No.66912498


>> No.66912506
File: 178 KB, 330x319, CP2077_Sun.png [View same] [iqdb] [saucenao] [google] [report]


>> No.66912509

>thinking that a 28 core 56 thread cpu is NEXT from JEWTEL
mate they wont release that for alteast another decade

>> No.66912524


Where are the prices

>> No.66912556


>> No.66912599

I have the cheapest h310 mobos you can buy and it runs my 8400 fine. sure it will probably throttle a 8700/k but who buys the cheapest board with the most expensive cpu?
benchmarks are geared towards threads for the most part.
Intel doesnt have a 100% IPC lead over ryzen.
you do the math.

>> No.66912759 [DELETED] 
File: 76 KB, 885x665, 8v4ml1wwmum01.jpg [View same] [iqdb] [saucenao] [google] [report]

Oy vey!

>> No.66912787

put me in the /r/4chan screencap XDD

>> No.66913073

How much zen non treadripper has?

>> No.66913116

But this is 14nm+++++???

>> No.66913823
File: 69 KB, 678x492, Blender OC_575px.png [View same] [iqdb] [saucenao] [google] [report]

Literally the FX 9590 all over again, but at least it'll run games good.

>> No.66913945

I don't know but they better release it before Ryzen 88

>> No.66913957

LGA 1151 V2 ,V3,V4,V5,V6,V7

>> No.66913966

>5Ghz boost
Yeah right

>> No.66913984

24, wired as 16+8 or 16+4+4 when you install an NVMe drive IIRC

>> No.66914013

For a while they did price-gouge the Athlon X2.

>> No.66914030


Gayming is baby-tier as far as computing requirements are concerned.

You can perfectly game on a low-end and mid-level customer-tier desktops. "High-end" desktops are only needed for certain niches or people who fell for the 4K gayming, 200FPS+ meme non-sense.

>> No.66914043

Its 16 (graphics) + 4 (M.2 slot) + 4 (Chipset link). The actual die supports 32 but the remaining 8 lanes are inactive.

>> No.66914056

I know this is bait but..
>16 PCIe Lanes
>i7 no longer has HT
that's bad

>> No.66914071

9900K will only be faster at four threads or less.

Anything above that, the dual-channel DDR4, ring-topology start becoming a problem and bottleneck. Thermal wall is also an issue. Solder is not a band-aid to problem either. Solder just makes IHS fittings at the factory more consistent. It doesn't magically reduce temperatures in itself.

I suspect 2700X will edge out a small victory at hilariously parallel tasks but otherwise contest the 9900K at six-threads or more.

>> No.66914161


>Brand inflation when your major competitor is starting to take back large pieces of market share.

Yes Intel, this is a brilliant idea.

>> No.66914804

Threadrippoff is still 1GHz slower than the i9.

>> No.66914816

>Hurr clockspeed everything
You enjoying that housefire there?

>> No.66914827
File: 1.29 MB, 418x402, SRSLY 080.png [View same] [iqdb] [saucenao] [google] [report]

>cool i9

>> No.66914843


>> No.66914861

>ancient memes
Is it 2007 again?

>> No.66914870

Oh shit that is pretty cool. Is MSRP unchanged? kinda regret jumping on Ryzen day 1 now. Also this is no different than previous iterations of i5 being 4/4 and i7 being 4/8.
Yeah but Intel still has better IPC and this will probably OC about 20% more. really depends on your usage DESU. As has been the case since Ryzen launched really.

>> No.66914919

>he thinks it's a meme and not genuine reaction

>> No.66914923

Yeah and the i9 has better IPC.
To me the part makes no sense though.
In that a the price has to be more than the i7 which is $400. And a 16 core threadripper is like $650 now.
So while single threaded performance is better. I just can not see a scenario where you want an i9 for hyper-threading but your workload wont scale perfectly to 16/32 and run much better on thread-ripper.
Like if your workloead is poorly multi threaded get an i7 or even an i5. but if your workload is scaling to 16 threads Threadripper is better and similarly priced. It is still only on the mainstream platoform to so you can say you need single threaded performance with Enthuist features like more PCIE lanes or ECC.

>> No.66914960

What a joke, I'll never buy an Intel again.

>> No.66915994

7nm doesn't matter

>> No.66916012


Can it officially be added to the list?

>> No.66916016

because burning the entire neighborhood down is a good thing

>> No.66916069

All CPU's should be benchmarked at stock settings. Manual overclocking is cheating.

>> No.66916160

>***NEW*** 7nm doesn't matter!

>> No.66916275

Why? Who the fuck uses stock clocks?

>> No.66916296

The majority of people.

>> No.66916300

>8 cores at 5ghz turbo
What's the problem exactly?

>> No.66916324

It is the exact same part as the i7 only with Hyperhtreading, making it cost more and presumably pitting it against a part with twice the core/thread count that exists on an enthusiast platform with much more features such as more PCIE lanes. Basically see

>> No.66916370

The majority of people are not looking up benchmarks or CPU specs.

>> No.66916413

So what is your point? That 150 billion dollar corporations should cater to your special snowflake ass because?

>> No.66916429

Intel is not the one's benchmarking it. Nor is their $150billion dollar valuation based on enthuist level unlocked socketed CPUs, which quite literally are made to cater to my special snowflake ass. So I am not sure your point.
If you are buying a k part to put into your computer, you should be manually overclocking it to get the best possible performance. Most do this, especially the ones reading PCper or Anandtech or whatever. if you are too retarded to do this. Dont buy a 9700k or whatever.

>> No.66916448

Remember to not overclock your k cpus, though, as intel stated.

>> No.66916455

kek, I forgot about that statement.

>> No.66916556

>eight cores
>dual channel memory


>> No.66916670


>> No.66916743

It's not. Intels TPD is for stock clocks aka 3.6GHz and not 5GHz.
OC that nuclear reactor to 5GHz and watch the TDP double, at least.

>> No.66917160

>What's the problem?
It OverClocks only first two cores and heavily downclocks others. AMD currently has slightly lower clocks by comparison, yes, but it ALWAYS OverClocks ALL cores & threads with no issues. Basically, you get those "5GHz" only on first two cores with Intel while also having thermonuclear temps & power draw and everything else past first two cores goes to utter shit as you do it. With Zen, you get stable "4~4.2GHz" on ALL 8/16 cores & threads AND better thermals/power envelope at the same time.

It's for you to decide in the end what appeals you more, but I personally prefer to have stable and same high clocks on all cores while sustaining decent/tolerable power draw & thermals (I have top tier cooling solutions anyway) than to have much higher clocks on just first couple cores and very bad power envelope with hery high temps.

>> No.66917177

When will there be more than 16 PCIE lanes on an enthusiast grade Intel CPU? Is threadripper literally the only option?

>> No.66917191


>> No.66917222

you forget intel was getting more from both cpu and meme chipset each year

now they had no option but to have a backwards compartibility meaning a huge portion of income wont be there

>> No.66917307

Intel 6core parts can hit 5ghz stable on all 6 cores with any half decent cooling and not terrible luck at the silicon lottery. Also worth noting that the IPC is still marginally better. Even then the 2700x so it is not apples to apples. Regardless AMD winning this one would really only be based on SMT, better pricing a less retarded chipset situation and the fact they still might have a part or two waiting to counter.
In any 8thread or less workload intel will pull ahead even with stock/turbo.

>> No.66917367

Tfw 10 intel cores destroy 16 iampoo corelets

>> No.66917408


AMD confirmed dead an buried FOR REAL THIS TIME GUYS huh?

>> No.66917619

>Multi-core doesn't matter!
>Productivity doesn't matter!
>Price/performance doesn't matter!
>Performance per watt doesn't matter!
>Power usage doesn't matter!
>Temperatures don't matter!
>Soldered dies don't matter!
>Stutters don't matter!
>Streaming doesn't matter!
>Data centers don't matter!
>Locked CPUs don't matter!
>OEMs don't matter!
>Hyperscalers don't matter!
>Upgradeability doesn't matter!
>Anti-competitive business practices don't matter!
>Locked platform features don't matter!
>Synthetic loads don't matter!
>PCI-e lanes don't matter!
>Burnt pins don't matter!
>ECC support doesn't matter!
>*NEW* 7nm doesn't matter!
>*NEW* HEDT doesn't matter!
>*NEW* StoreMI doesn't matter!
>*NEW* Security doesn't matter!
>*NEW* Stock coolers don't matter!
>*NEW* Games don't always matter!
>*NEW* Hyperthreading doesn't matter!

>> No.66917974



>> No.66918054

Feels good to still be on LGA 2011-3 with my 8 core 16 thread @ 4.6.

>> No.66918128

>not soldered

>> No.66918411

There is no point in overclocking all cores when you're only using 2.
Intel's approach is pretty clever here if you ask me, they can sell you an 8 core that outperforms a quad core in single threaded benchmarks.
But even with all 8 cores it still runs at 4.7Ghz apparantly, which is a lot better than AMD does right now.

>> No.66918590

>But even with all 8 cores it still runs at 4.7Ghz apparantly
Not without a delid + watercooler + expensive motherboard

Name (leave empty)
Comment (leave empty)
Password [?]Password used for file deletion.