AMD BTFO: 10 core Skylake-X i9 7900X 4.8GHz OC without delid

File: pHrnT.jpg (184KB, 980x602px) Image search: [Google]

184KB, 980x602px

>>60838297
LMAO

File: Screen_Shot_20170605_at_13.41.33fullpage.png (2MB, 1302x1942px) Image search: [Google]

2MB, 1302x1942px

>>60838297
>>60838409

File: TB2f4vIvrJmpuFjSZFwXXaE4VXa_2341477237_1.jpg (470KB, 1060x1569px) Image search: [Google]

470KB, 1060x1569px

>>60838297
>>60838430

File: 1430236296498.gif (112KB, 255x231px) Image search: [Google]

112KB, 255x231px

>>60838409
>>60838430
>308W
Netburst 2.0

>>60838297
A weapon to defeat 9590.

Stop circlejerking buy the best processor suited for you

>>60838297
Gee, I wonder why he didn't run any benchmarks?

>>60838409
>>60838438
Oh. That's why.

>>60838297
>inb4 hundreds of people try to replicate this but lose the silicon lottery and end up with a delidded disappointment

>>60838469
I wanna watch the world burn and dance in its ashes.
What CPU should I buy?

>>60838409
>>60838430
>>60838438
NOOOOOOOOOOOO

>>60838533
7980XE and overclock it to 4ghz.

i9 7900X

> requires massive overclock to be relevant.
> DDR4-2666
> 140W

Kek Intel are in panic mode and still offering fuck all to top Ryzen, Threadripper, or Epyc.

File: 6a00d83452a98069e200e54f76d93d8833-640wi.jpg (22KB, 470x312px) Image search: [Google]

22KB, 470x312px

>>60838560

$2000

95°C. On AIO.
Totally 24/7 reliable overclock.

>>60838297
lol https://www.hardocp.com/news/2017/06/05/intel_core_i9_skylake_e_clocks_no_higher_than_broadwell/

>>60838297
>Intel bulldozer

>>60838597
On a golden sample cherry picked to give to reviewers. And then delidded.

File: 1364076496407.gif (2MB, 320x200px) Image search: [Google]

2MB, 320x200px

>88°-90°C
>after delidding
>on a 280mm rad
>admits that it's a pretested golden sample

>>60838533
Ryzen seems pretty nice.
But if you're under a tight budget Sandy Bridge i5/i7s seem good. But I can't speak for much as I'm on Kentsfield

>>60838597
>>60838628
Literally worse than 9590.

I wonder how well 18c part will OC.

inb4 2000W PSUs.

File: 1476929491973.jpg (63KB, 250x323px) Image search: [Google]

63KB, 250x323px

>mfw Intel inherit AMD old housefire meme

>>60838662
But it's Intel (and then novideo) that produced the original housefires™.

>>60838679
LONGER BIPELINE
>LONGER BIPELINE
LONGER BIPELINE
>LONGER BIPELINE

>>60838689
Absolutely EBIN my friend.
MORE BINGBUS
>MORE BINGBUS
MORE BINGBUS
>MORE BINGBUS

>>60838652
About as well as an orange does.

File: 1492127495787.png (871KB, 801x1500px) Image search: [Google]

871KB, 801x1500px

>>60838297
>>60838409
>>60838430
>>60838438
MOAR COARZ
MOAR GIGAHURTZ

File: fx5800.jpg (9KB, 250x250px) Image search: [Google]

9KB, 250x250px

>>60838679
Yeah, I remember the FX5800.
https://www.youtube.com/watch?v=QvVr1El_q78

>>60838430
DELID THIS

>>60838724
I wonder if nVidia will produce the third set of housefires™ to fight off Navi.

>>60838297
lol comments are still full of talking shit about Intel. I don't think this "OMG LOOK AT THE UNREALISTIC OC" shit is working anymore.

>>60838773
Even kids that've never seen NetBurst don't want NetBurst 2.0, anon.

File: 1496424178719.jpg (84KB, 640x429px) Image search: [Google]

84KB, 640x429px

>>60838785

>>60838785
This is the trouble.
Millenials who were barely able to comprehend how to turn a computer on are now old enough to be learning about this site.
None of them know of the dark ages.
>Though how they could be that dark with housefires burning into the night thanks to Fermi and Netburst is a mystery

>>60838757
I wouldn't put it past AMD to fuck up nvidia with Vega yet, though.
Even if you picture it as an overclocked Fury, It should still score between 1080 and 1080Ti.
I think them, not showing FPS on Prey demo may have been deliberate.
We'll see. At the end of the day, price is the decider.

>>60838785
>tfw still on presshot
My PC's side panel is off because of this volcano. Summer is suffering.

>>60838836
nVidia can fight off Vega with bigger dies. If Navi will bring MCMed GPUs nVidia will not be able to EVER slap more silicon than AMD.

>>60838297
>/g/ is finally redpilled about intel

what a time to be alive

>>60838836
Screentearing in 4k means one of two possibilities;
>150fps and no vsync

>underdeveloped drivers
AMD have a well established history of somehow being able to fuck up every product launch they do.
If they get the drivers sorted for release, and give us something solid and stable, it'll be a very good launch. Assuming Lisa doesn't pick up the T-shirt cannon again.

>>60838856
Without knowing the limitations of Infinity Fabric it's hard to say what AMD can do with it.
They're doing well to keep the hatches closed and stop anything getting out about it beyond that it can fuck shit up for days.
A high latency interconnect may not play too well in a GPU.
But, if AMD are to implement IF with their SSG idea, then it should provide a very tasty workstation solution.

>tfw really want a 256GB SSG for solid works
>AMD only make it in 1TB

Why do I see this republic of gamers bullshit everywhere. Fucking asus chinks spreading their stupid gaymer bullshit everywhere. I hope china nukes taiwan, fucking chinks.

>>60838856
That's the sad truth, really.
Even back when 4870 and following 5870 were kicking nvidia's ass at half the transistor count, they didn't sell.
They were fucking cheaper too.
AMD has no reach to normies.

>>60838866
That presentation was beyond retarded and Raja should probably gas whoever was responsible for it.
>showing crossfire RX Vega on 60hz projector with no freesync
The fuck.
>>60838886
Are SSGs good? I have no idea about prosumer cards.

>>60838905
>Fermi outsold Evergreen
I still wonder why the fuck would anyone buy these housefires instead of fucking 5850.

>>60838906
>Are SSGs good? I have no idea about prosumer cards.
The idea behind it, yes. Having a huge, high speed storage media right there on the GPU so it doesn't have to pass via CPU and PCIe before it's available makes it super easy to work with large file sizes. At their presentation on it, they ran 8k with realtime cleanup at 96fps. Whereas the nVidia quadro they were comparing it to was barely pulling 15fps.

https://youtu.be/vyKgT5QUU2Q?t=4003

Being able to have those files open, and so easily accessible by the GPU makes a massive difference to your workflow.

>>60838945
Maybe this time, the good publicity from Ryzen will help their brand.
They'd better not fuck it up though.

>>60838951
Neat.

>>60838974
If AMD can recapture a good number of OEMs who are held in high public opinion, like ASUS, then normies start to get a grasp of the situation.
What needs to happen though, is AMD need to be able to deliver comparable battery life to low power, high performance intel mobile solutions and comparable performance. The difference being, AMD have such massive potential to put a hugely powerful iGPU in there and make all the difference.

>>60838974
Considering how many people are crying about Vega (they'll never buy it anyway), I don't think Ryzen would help. Nvidia mindshare too stronk.
>>60838992
RR will have more perf with the same battery life. And AMD iGPU is ages ahead of Intel.

>amd btfo
>10 core overclocked without AVX instructions
Bulldozer 2.0. Useless.

>>60839009
>RR will have more perf with the same battery life. And AMD iGPU is ages ahead of Intel.
Yes, in theory RR will utterly smash it.
But OEMs have a good history of completely fucking it up and jewing out on the chassis they put AMD chips in. No real cooling solutions or engineering gone into it.
AMD need to maintain a dominant position and keep control of what happens with their products.

>>60838992
Ryzen APUs should be fucking awesome though.
Undervolting Ryzen chips show about [email protected] for most chips.

>>60839026
RR targets "premium mobile" market first. Anyway, we'll see. The true shoah begins on 20th of June.

>>60838992

I don't think anything AMD does will help capture the mainstream market at this point. I have an all AMD system and got a r7 1700 in March. I was trying to tell some coworkers about it and they specifically said they don't care how good AMD ever is they'll never buy it because 10 years ago they did and had nothing but problems. It might be too late for those people.

>>60839037
I'd love to see AMD just open up and just name drop companies that have preordered a significant number of EBYN chips. Just to fuck shit up.

>>60839045
That'll happen in 20th of June, along with OEM rack designs.

>>60839038
Agreed.
And btw, a chink CEO, and indian Graphics department handle don't really scream confidence, even if their competence is real.
It just feeds the 'poo in loo' banter.
It's sad, but that's what it is.

>>60839038
And with recent developments from Intel, all the ME bullshit going public, public opinion of Intel around here isn't too high right now.
People have to figure out for themselves that one is better than the other. You can tell someone anything, but until they actually sit down and start to use it, to feel the benefits, they'll never accept it.
This obviously doesn't apply to the server space and workstation markets.

>>60839074
How's nVidia's chink CEO is better?

>>60839074
>One of the biggest names in semi conductor research, development, and engineering as CEO of a semiconductor CPU company
>Not confidence inspiring
Lisa Su has done massive things for AMD since being there. Opened up so much more of the market.
Rory Reed fucked it all up mayng

>Indian GPU lead
He has a masters in EE, that's no easy feet to achieve.
He has an established history of development lead at ATI.
Worked in apple as GPU engineer for some years before returning to AMD.
while he may be a shitskin indian, he is most certainly qualified. Which is more than can be said for 80% of Intel's employees lately.

Btw /g/, I don't really feel like starting a thread about it, but here's the situation:
I've got 170€ worth from returning ram that didn't work with my motherboard/ryzen.
What the fuck do I do with it?
I had to purchase actual ram from another site, because these fucks didn't have anything samsung b-die.
I already have crossfire rx480s, oculus rift, force-feedback wheel and joystick. Also a 4k freesync monitor and 10TB of storage.
I have just no idea, what to do.

>>60839106
Leather jacket

>>60839134
Well...buy RAM that will work obviously.

>>60839140
Pork shoulders.

File: AMD vs Nvidia vs Intel.gif (433KB, 630x206px) Image search: [Google]

433KB, 630x206px

>>60839038
most chips are sold through OEMs anyway
it's not like normies are going to care about cpu/gpu brand

>>60839141
That's the thing, I can't.
It's like they deliberatively single out models that will work with Ryzen, ((for some reason))

File: flat,800x800,070,f.jpg (80KB, 800x569px) Image search: [Google]

80KB, 800x569px

>people born after pentium 2 shitposting about netburst

>>60838724
The original Nvidia housefire and the original Intel housefire were a match made in heaven.

>>60839177
Netburst was fine.
Sure it ran hot, but it did hyper-threading and shit, before dual cores appeared.
I owed one, and it was bliss how the system didn't just hang on fucking background tasks.

File: 1485368059242.jpg (60KB, 753x1024px) Image search: [Google]

60KB, 753x1024px

>>60839177
>pentium 2 was released in 1997
>20 years ago
shit I feel old now

>>60838297
Lmao Intel fags are turning Skylake into the next Bulldozer.

>Just overclock it moar :-DDDD

>>60839220
>next bulldozer
Anon Intel invented housefires.

>>60839177
>Having to put a fan on your CPU
REEEEEEEEEEEEEEEEEEEEEEEE

File: 1496657594176.png (241KB, 900x900px) Image search: [Google]

241KB, 900x900px

>>60839156
this, majority of normies i met don't even know what OS they are on let alone the chips inside. It's like buying a car and not knowing what engine it uses, it makes me reeeeee so much on the inside.

>>60839177
>he's born after P5
Fucking kids get out

>>60839184
No, it was a match made in hell.

>>60839148
Alpha tribal nvidia tatoo

Who here used to Cyrix?
My god were they ripping off their customers.
Good thing they're gone.

>>60839283
1.7% yields.

>>60839269
I had a friend in Australia who had both. No idea how it didn't start a bushfire.

>>60839235
>he didn't spend years studying metabolic pathways and atp signaling chains before starting to eat food
get out normie

>>60839087

No these guys don't give a shit how it runs they just refuse to ever buy AMD cause of past experiences.

I have never had the problems they talked about other than knowing it did not perform at the same level as Intel when I bought it but I bought it because it was cheaper and got the job done. Now I can afford to get either and I still chose an r7 because I need to be able to run my VMs and game a bit.

File: 1496762224056.png (3MB, 1944x1005px) Image search: [Google]

3MB, 1944x1005px

>>60839333
kek

Well at least, we know X299 boards can handle enough power for 18 cores@3Ghz.
At least enough to take a screenshot.

Have a friend who is talking about this vid like it's the next coming of Jesus buttfucking Christ. What do I tell him to carefully explain how much a huge faggot he is without flat out telling him he's a dumb Intel fanboy?

BTW OP description is wrong, he does delid it in the video, you can even see it with the lid off.

>>60839495
>2x8-pin CPU power connectors
Of fucking course it'll have enough power. But it's as hot as the sun. 95C on 280mm AIO.

File: intel circumcision tool.jpg (139KB, 954x720px) Image search: [Google]

139KB, 954x720px

lmao so this guy is just a hardcore Intel shill now?

>>60839514
He's the same guy that desperately defends jewish shit TIM.

>>60839514
hes not an intel shill per se he just sells the fucking delidding kit so of course he's going to pretend it's not shit

File: delid_this_logo.png (334KB, 679x702px) Image search: [Google]

334KB, 679x702px

Remember goys, delidding your CPU is now idiot-proof! So delid away! What could possibly go wrong!

http://www.pcgamer.com/meet-the-overclocker-who-made-delidding-your-cpu-idiot-proof/

>>60839512
Connectors are a thing.
Board wiring, manufacturing and VRMs to handle power lines like that are another.
It's such a clusterfuck of power and data lines near the CPU, it gives every electronics student a nightmare thinking about it.

>>60839546
What the fuck.

>>60839546
>http://www.pcgamer.com/meet-the-overclocker-who-made-delidding-your-cpu-idiot-proof/
Voiding your warranty is now idiot proof.
Not that intel's warranty was worth anything to begin with.
When my P4-E 3ghz died for no reason, I tried contacting them.
I couldn't make out 1 in 10 words the fucking indian fuck responding to me was saying.
Decided I'd better take my loss and went A64.

File: Capture.png (411KB, 1103x1208px) Image search: [Google]

411KB, 1103x1208px

>>60839514
>>60839526
his job is selling delid tools and delidded, pretested CPUs for huge markups.
He would defend Intel even his sister got gangraped by the Intel board of directors.

File: Mysterious Merchant.gif (4KB, 452x523px) Image search: [Google]

4KB, 452x523px

>>60839514
>PLS BUY INTEL FACE

>>60839604
>someone's selling delidded Intel CPUs for huge margins
Motherfucking god someone stop Intel pls I'm dying of laughter.

>>60839556
Well, the way i understand how its set up, the DRAM Phy is the biggest consumer of connections, while the power feed is typically limited to a single blob on one side of the package. Routing the power feed is relatively easy compared to routing several thousand traces for the DIMMs and a couple hundred more for PCI-E and other chip I/O.

There's also the fact boards with that much shit near the socket typically have more layers than normal.

>>60839637
Imagine the pain of routing 16 memory channels of 2S EPYC board.

File: 1448975922262.jpg (36KB, 400x460px) Image search: [Google]

36KB, 400x460px

>>60838297
>in the end he says its pre-tested golden sample

>>60839649
Board layers. Lots of board layers. At bare minimum you're looking at 6 layers (4 data, ground, and power), and will likely need more than that for signal quality and easier routing.

File: Threadripper_vs_Coffee_Lake.png (1MB, 1255x797px) Image search: [Google]

1MB, 1255x797px

>tfw /a/'s hand-me-down memes are a perfect fit for /g/mouto

>>60839671
Every layer is more cost.
It's also not absolute shielding.
You still have to consider power lines 2 layers below/above.
Or consider error correction mechanisms in your cpu.
Especially when the thing is drowning 300W.
I'm even impressed this thing exists at all.

File: amazaryzen.png (68KB, 766x668px) Image search: [Google]

68KB, 766x668px

>>60839687

>>60839713
2tb of memory per socket are too sweet to not design board for.

>>60839767
384GB DIMMs when?

File: 1496954629640.jpg (103KB, 1280x720px) Image search: [Google]

103KB, 1280x720px

>>60838297
>PROJECTING THIS HARD

>>60839514
>defending RFID chips

this nigger i swear

>>60839782
You want even more memory per socket?

Anyway, I'm predicting everything 12 cores and up Intel will be downclocked to oblivion.
Actually making more coars irrelevant.
Performance will just be linear with the max output of the motherboard's VRMs.

>>60839286
I got one for the company I worked for. Huge mistake. Slow pile of shit.

>>60839809
Why not?

>>60839286
Got one for free a long time ago.
Introduced my young mind to the concept of bestiality.

>>60839830
Ebin two bingbuses are shit.

>>60839604
What the fuck

File: 1493663366636.jpg (48KB, 410x416px) Image search: [Google]

48KB, 410x416px

>>60839865
It's the best CPUs in the world, now delidded, goy.

>Admits it's a pretested golden sample in the video
>Sells delid kits and delidded CPU's at ridiculous prices
>Is known for overclocking on LN2
Basically if you believe you will get a CPU from Intel that runs as high as this without considerable knowledge and voiding your warranty you are pathetic and delusional.

>>60839957
You forgot
>90C on 280mm AIO

Hey goys.
Heres our new shit.
Notice how it has more coars than competitor.
What, they run slower?
Don't worry.
Here are some Quake 2 benchmarks that will show you how much better it is.
What do you mean it's not a gaming cpu?

People get way too mad when a company tries to max out the profit margins although it's a normal business practice. You shouldn't exaggerate it of course since it will alienate your customers. But still, I wouldn't be surprised if the production of Intel's best chip is maybe just 10% more expensive than the production of their G4560 dual Pentium chip. The difference in prices just being diversification like every company does just like when Microsoft releases a basic and a pro edition of their operating system.

File: gay meme.jpg (267KB, 821x1208px) Image search: [Google]

267KB, 821x1208px

Anyone think of anything to add to this?

>>60839987
That's not how it works. Intel's process is outdated and thus their chips cost a lot if they have more cores. AMD hit the jackpot with their new method.

German Jedward over there is a competitive overclocker and supplier of extreme cooling parts. He gets all his shit for free and literally couldn't care less about economics or competitive positioning. That's why he "can't understand" the negative press.

>>60839977
On an open air test bench too.

Notice how all leaks of skylake are overclocks? Stock perf must be really disappointing compared to raisin and broadlel.
And most HEDT buys don't give a fuck about running a sun in their case

>>60840047
What a fucking autismlord.

As I predicted long ago, as soon as competition happened, Intel just went and dropped out its iGpu bullshit.
I wonder how many people in history that bought a K processor ever used that shit.

>>60840056
https://www.hardocp.com/news/2017/06/05/intel_core_i9_skylake_e_clocks_no_higher_than_broadwell/

Pretty much already confirmed.

>>60840056
Intels own slide put it 10% faster than a 6950X, the 10core SKL-X

So yeah, fucking nothing.

>>60840028
So they really cost more to produce if they have more cores. Yeah, I could be wrong then. I just watched a video how processors are made on YouTube and when I thought about it I couldn't really see how a less powerful chip is created differently than a more powerful chip. They all go through the same production process on the same machines, some take maybe longer but there still wasn't much of a difference. But still, I can't imagine that a 1000$ intel processor is really 5 times more expensive to produce than a 200$ chip. I could imagine that the production costs are much closer.

AMD actually invented crossfire CPUs.
Nothing to stop them now.

>>60839989
A huge cock in mouth

>>60840149
>AMD invented MCM
Wat.

>>60840161
The 'revolution' is that it's on a single socket.
Before you would have needed multiple sockets for that shit.

>'competitive' overclocking

so uh is this like a visceral heartpounding spectator """"sport"""" or what

File: intel-core-2-quad-q6600-cpus1.jpg (26KB, 375x280px) Image search: [Google]

26KB, 375x280px

>>60840149
They didn't invent it, they just made it work well.

>>60840179
>AMD invented MCM
Dude you're just retarded.

>>60840130
The cost is mostly for the area on the wafer and the possible defects. They can make more 4 cores from the same material that work fine than a few 22 cores. Watch Adored's recent video on this.
Basically AMD is now making 4 core chips and can build any number of cores from them like it's lego while Intel still fucks itself with the full cpu on one silicon method.

>>60840200
AMD is making 8core chips, though. Zeppelin is ~195mm^2.

>>60840195
>>60840193
I stand retarded.
Nevertheless, what prevents intel from doing the same.
Performance looks quite linear on AMD's side.
I suspect they've got something here.

File: Power5.jpg (64KB, 800x531px) Image search: [Google]

64KB, 800x531px

>>60840179
hmmmMMM?

>>60840185
There are online leaderboards (a bit like with speedrunning), but yes it is also an "esport" where people try to get the best benchmark score against the clock.

https://www.youtube.com/watch?v=b_k3uXRkBsI

>>60840200
That's interesting to know. Thanks. I will also watch the video.

>>60840217
Whatever, it's the same stuff.

>>60840218
>Nevertheless, what prevents intel from doing the same.
No suitable fabric.

>>60840218
It's called infinity fabric.

>>60840218
Intel *could* do the same but their interconnect technology is somewhat behind AMD.

>>60840227
Doesn't this just come down to how good of a CPU you have? Seems pretty retarded.

>>60840249
More like how well you know your way around the bios and memory settings, as well as how well your OS is set up.

>>60840246
QPI is ages behind IF and Omnipath a shit.

Honestly speaking, I went from i7 3770k to 1600x, and the only thing I noticed is more graphs in task manager.
More coars is a meme for now.

>>60839792

>"why does my processor box look like it should contain tampons comrade?"
>"answer is easy my friend, both products are for pussies"

>>60840287
>I have no need for better CPU yet i've upgraded
Why?

It's funny how Intel is throwing stuff out there to try and hype their shit up whilst AMD are keeping their shit close to their chest. It's almost like Intel are in a rush or something.

>>60840303
CPU died a month after I oc'd it to 4.5 Ghz.
Not even 80°C and 1.25V.

Well I'm lying, it still works.
But at 3.0 Ghz.

Remember kids, oc brings the risk of gimping your cpu.

>>60840325
Remember to not OC your K series unlocked overclockable Intel cpus, according to Intel.

>>60840310
There's no 12-18 core SKL-X parts at all. They were paperlaunched. AMD can take all the time in the world with kikeripper.

>>60840347
Funny thing is, I did it to convince myself I didn't need to upgrade.

>>60838811
You're humorous joke aside, you don't know what millennial means.

This site and it's predecessors were built and populated entirely by millennials. People born in the year 2000 and after are not millennials.

>>60840149
>AMD invented crossfire CPUs
>AMD invented MCM
>calls MCM crossfire

I know what you meant, and I know you've been corrected and admitted to the mistake. But Jesus wept, child; fix yourself.

>>60840496
Nobody's perfect, and nobody will be ever.
What would be the point of exchanging banter if that wasn't the case?

>>60838409
Bingbus - X strikes again, FPBP

i got a i7 6700 last year. im no enthusiast but since i've got a job months ago i have the urge to buy something big.
I want to game while i do shit, maybe have vegas rendering or baking something in blender, that shit possible yet? or should i wait some of these 16 core cpus?

File: 1459629113233.jpg (47KB, 531x471px) Image search: [Google]

47KB, 531x471px

>>60838409
> 500 single thread performance
WHAT THE SHIT MAN

>>60840598
I'm no expert but if you already have a caliber like the 6700 you should wait for the 16 cores. Current stuff is not that big of an upgrade.

>>60840608
Welcome to >6 core Intel CPUs. Only their quad cores have a significant single threaded advantage.

>>60840645
Well, yeah, I would never drop so much money (HEDT), so the 7700K is fine. I expected some loss due to the core number and heat, but THIS FUCKING MUCH?

That is just horrible.
Not attacking any company here, we don't know what Zen will be capable of with Threadripper, just saying that this shit is unacceptable.

>>60840673
Imagine the riot if people learned there's a NSA mole in their intel stuff.
Wait, they know and don't care.

>>60840608
CPU-Z benchmark scores got updated so that all CPUs have lower scores now. The old scores are no longer relevant.

>>60840673
Isn't threadripper practically out? I mean people have samples and we know how is it. Mostly.

>>60840701
Son I use Chrome. And Windows.
I do work to earn money.

However, I cannot fucking work if a CPU scores 500 fucking hundred points ONLY. For fuck's sake.
One step forward, ten back? What the fuck?

>>60840716
Until we get scores, there is only shitposting and shills really. We will see. Not saying it will be bad, or anything, it may be the best thing since sliced bread. We will see.

>>60840734
>2 Zeppelins with near linear scaling
It's already killed Intel.

>>60838974
Nobody really likes Intel, everyone just bought them because FX was garbage. Nvidia has a massive amount of fanboys and much better marketing than ATI/AMD, even when AMD beat Nvidia is performance or price no one cares.

File: nvidia graph rollercoaster.png (24KB, 678x381px) Image search: [Google]

24KB, 678x381px

>>60840768
I fucking LOVE nvidia's marketing.

File: fucking_joke_processors.jpg (42KB, 407x404px) Image search: [Google]

42KB, 407x404px

Can't believe there is just no progression on the desktop CPU market whatsoever.

>>60840598

Thread Ripper is what you want. Along with dual GPUs. One for pass through to a VM you use for gaming the other for the host.

>>60840734
I thought we had some rest results already but my memories might be failing. Things happen fast for my old brain.

>>60840824
It's okay, it's more like we have so much shilling and misinformation (and shitposting) on /g/ that it is really hard to filter out the real information.

>>60840673
>I expected some loss due to the core number and heat, but THIS FUCKING MUCH?
Welcome to ringbus city.

>>60840673
>we don't know what Zen will be capable of with Threadripper
We pretty much do already. https://youtu.be/NeuamYdfU4w?t=16m

The time an 1800X usually gets on this blender test is 23 seconds, so that's pretty close to 1:1 scaling.

>>60840865
Clocks and setup unknown, for all we know AMD could be sandbagging again. Seems like it's become their staple

>>60840815
CPU-Z's benchmark is garbage. http://www.guru3d.com/news-story/new-cpu-z-upgrade-lowers-ryzen-performance.html

>>60840768
>even when AMD beat Nvidia is performance or price no one cares
>the state of the delusional AMD fanboy

>>60840905
You must be 18 or older to use this website.

>>60840905
I'll skullfuck you with my 5850.

>>60840904
So what, did Intel pay the bench company to fake amd data this time?

>>60840958
Ryzen's preprocessor was throwing away the "busywork" the benchmark was doing as it could tell it was useless. So they just nerfed AMD across the board, because they're too stupid to write a functional real-world benchmark.

>>60840734
>I use Chrome. And Windows.
>However, I cannot fucking work if a CPU scores 500 fucking hundred points ONLY
You sound like a completely clueless retard. Go to reddit or something, grandpa
see >>60840712

>>60840958
Ryzen branch predictor was too smart for the bench and cpu-z devs "normalized" the performance "inline" with other benchmarks trying to make some average

In conclusion, mostly horseshit.

>>60838951
>Capsaicin
Fucking curry niggers

>>60840325
Did you by any chance have LLC enabled?

File: 5850_Vapor_2GB.jpg (99KB, 1000x998px) Image search: [Google]

99KB, 1000x998px

>>60840935
Muh brutha

>>60840985
>>60841058

>bulldozer
>shit branch prediction
>this is fine
>ryzen
>amazing branch prediction
>SHUT IT DOWN

>>60841532

It's like they aren't even trying to hide how much they've been trying to dick AMD up the ass. And it's not just CPU-Z either.

Time Spy is an absolute joke of a Direct X 12 benchmark that had to be neutered because Nvidia couldn't run it.

Sysmark has a LONG history of fucking AMD over as well.

>>60838297
And it'll still lose to AMD's 10-core because Intel's ringbus scales horribly.

>>60841697
AMD should just lump them all into a lawsuit and sue them all at once. Userbenchmark flat out claims they're biased. People make poor purchasing decisions based on this shit.

>>60841735
AMD has no time and no money for silly lawsuits.

>>60841767
I'm getting tired of having to tell retards on here and other sites that passmark is garbage that nobody should be using for anything.

who cares, you're all being conditioned to buy new cpu's morons

>>60841783
They are using pissmark to shitpost. No one really takes pissmark seriously.

>>60841790
>hurr durr core2quad should be enough for anyone

>>60841797
Plenty of idiots still do. I've had to convince multiple people IRL not to use it for anything.

>>60841804
B-but 2 moar corez goyim.

>>60841804
Shouldn't be using CPU-Z for anything desu. >>60840904

>>60841735

Lawsuits take a lot of time and money. It's just not worth it.

But it's fucking hard to be AMD. We won't even go in to the medias skullduggery who kicks them in the balls because they don't buy advertising.

OEMs relegate AMD to second class status even when they have competitive products like giving single channel memory in APUs.

Then they have suppliers that throw all their manufacturing behind the Z270 and leave Ryzen mother boards in short supply then say AMD gave them no time.

They can't catch a fucking break.

>>60841800
core 2 quad came out a decade ago and intel literally sells a quad core in the 200+ range when it should've already been a 8 core in that price range. fuck intel

>>60841835
It's like everyone in the x86 market actively wants it to die.

File: 1493430222289.png (15KB, 417x414px) Image search: [Google]

15KB, 417x414px

>>60838409
Neat

File: Screenshot 2017-06-10 13.21.43.png (2MB, 1920x1080px) Image search: [Google]

2MB, 1920x1080px

>>60838409
Not even 5% faster than my $300 R7 1700 LOL.

>>60841863

Nice 2p workstation.

why the fuck doesn't windows 10 have encarta?

>>60841797
Unfortunately there is a sizable group of people with the brainpower to build their PCs like a packaged LEGO kit, but not enough to understand which benchmarks are BS.

I am friends with several of them. It makes my head hurt, but I'm not autistic enough to go off on them.

>>60841890
>627 vs 430
>5%
What?

>>60841951
Let loose your inner autism.

File: jews.png (116KB, 830x866px) Image search: [Google]

116KB, 830x866px

>>60840815
>>60840958
>>60840904

>>60841697

Timespy is a fantastic one is the original reason why the various unigine benchmarks cranked up tessellation was "to push gpus to their limits". Then they hype up the asychronus nature of timespy and use relatively little because (as per a long buried post from a futuremark employee on anandtech's forums) "it has to run on everything from intel's igpu to dedicated gpus and a heavy asychronus load would crash certain hardware".

In other words fuck AMD precisely because they have hardware for the task nobody else can match.

File: nice ringbus faggot.png (21KB, 722x125px) Image search: [Google]

21KB, 722x125px

>>60841979

>>60842018
Core speed is different.

>>60841979
That's a 527 not a 627.

>>60842066
when the CPU isn't doing anything it clocks lower. its the same speed.

>what is XFR

>>60842025
>13% = 5%
Tard

>>60842018
3.2ghz vs 2.55ghz

Are you even trying?

File: 1463092001798.png (82KB, 300x191px) Image search: [Google]

82KB, 300x191px

>>60842165
Multicore score baka~ Literally the only thing that matters on a HEDT platform.

File: 102cee2d7faf13b5744942deda834ad9.png (224KB, 799x540px) Image search: [Google]

224KB, 799x540px

>>60842196
>>60842066
>what is Cool and Quiet

I always knew /g/ was retarded but come on. CPUs have been doing this for over 12 years now.

>>60842237
>muh multi score
AMD fags everyone.

File: 1481915065259.jpg (52KB, 932x576px) Image search: [Google]

52KB, 932x576px

>>60842288
>what is HEDT?

kym senpai

>>60842288
>muh single score
Intel fags everyone.

>>60840904
>>60840815
>>60842018
Well since we both tested on the latest version, the scores should be the same. I mean my cheap ass laptop CPU is only a tiny bit slower in single thread. (feels kinda good)

Btw I use Passmark mostly, but it's a shitty trial version... we could test that too.

File: 4770_hq_nooc_all_100.png (70KB, 800x600px) Image search: [Google]

70KB, 800x600px

>>60842365
Here is mine.

>>60842347
Wut? Single thread is 99% of desktop use mate.

>>60842023

Watch them revisit the time spy after Volta is released.

>>60842378
>implying hedt is for normal desktop use

>>60842347
Because single thread performance is literally what makes a computer "fast"

Do I need to bring up the 9590? On paper, it's 8 cores we're faster then the 4770ks 4 cores, BUUUUT...

File: here.png (197KB, 1929x354px) Image search: [Google]

197KB, 1929x354px

>>60842196
>>60842101
>>60842066
>>60842018

>>60839989
>>60839604
>somebody actually bought and rated a €520 7700K

>>60842390
No one has an idea what is it for yet.
There is desktop, high-end desktop, servers, workstations (epyc, xeon), and there is this HEDT thing.

All we know is that it will cost over 9000 bucks.

>>60842101
XFR is not generic name for frequency management.

>>60842423
>stock speed is 3.0, but runs at 3.2 most of the time

>>60842365
Pissmark is also trash.

File: muhsinglecore.png (453KB, 1920x1080px) Image search: [Google]

453KB, 1920x1080px

>>60842403
Zen is not Bulldozer, so your retard comparison makes no sense.

>>60842416
HEDT literally is what Intel calls a high end desktop.
It's nothing special at all.

File: 1457293195161.jpg (48KB, 243x290px) Image search: [Google]

48KB, 243x290px

>>60842416
Are you okay?
Do you have brain damage?

>>60842416
HEDT is high end desktop. Literally. High. End. Desk Top. HEDT.
Anyone who gives a shit about single core on HEDT is retarded btw. The only exception is if your single core so so astronomically bad that it affects performance negatively.

>>60842508
Shitposting is never acceptable.
Back to /a/.

>>60842378
>Single thread is 99% of desktop use mate
Not only is this a load of BS, but even the highest end i7-7700K is only about 25% faster in single core over Ryzen anyway. Programs that gobble up one core's worth of CPU and just sit there hogging it are a rarity. Even browsers are multiprocess now.

File: Cinebench426sing.png (34KB, 322x817px) Image search: [Google]

34KB, 322x817px

>>60842483
>muh it was the past, it don't count.
Kill yourself.

>still digging into this multithread thing.
My dual xeon from 2011 beats your brand new ryzen then, according to your views on performance. your point?

>>60842561
Ryzen offers a better balance of single and multithreaded performance than your old-ass Xeon.

>>60842579
and less power consumption
less heat

>>60842407
The core clock is slower.
That explains the score change

>>60842579
>moving goalposts
So now singlethread matters suddenly?

>>60842617
>having this much autism

>>60842557
Sadly, your browser example does not stand. They indeed tried to thread things in the past years, but execution still runs on one thread. So indeed even Firefox has a separate thread for the GUI, but only one of your cores will do the heavy work.

Same on Chrome, the javascript parsing for example will only affect one single core. And the faster that core is, the less time you have to wait.

Same if you start Visual Studio, if you build a project that does not have a build system (so almost any project except Gentoo/open-source programs). Games usually (sadly, still 90% of them use 1-2 heavy threads.)

Then again, some people encode video all day long, or render in blender all day long, or like issue "emerge world -avuDN" every day, or something like that. In that case, having a better threaded perf is better indeed.

>>60842617
Stop shitposting.

>>60842631
>losing so bad that you result to attacking the person on the other side of the argument instead.

>>60842640
Can't it run multiple concurrent instances of Javashit?

>>60842655
Nope. If you would have several pages open, with all Javascript running that would be threaded. Servo tried to change this (by Mozilla), it's a new browser engine, but so far there is not much progress there.

Looked at the current progress of it, it's still 2-3 years until it becomes "something".
But as I mentioned, there are parallel workloads like rendering, or video encoding where so many cores or threads would be beneficial.

>>60842648
>Old Xeon: Shit balance, runs on old server hardware that is hard to maintain and find parts for
>Ryzen: Good balance, new, has support for latest stuff, can be upgraded, inexpensive
>New Intel: Shit balance again, on a dead socket that can't be upgraded to Coffee Lake

>>60842613
When the cpu is not in use it lowers the clock speed

File: Screenshot 2017-06-10 14.21.19.png (1MB, 1920x1080px) Image search: [Google]

1MB, 1920x1080px

>>60842561
>dual xeons barely managing 61 points above a single socket consumer cpu in multithread

>>60842579
But this is HEDT Goy, single thread doesn't matter. The old ass xeon is 9% faster then the 1800x

>>60842725
None of those are HEDT. One is old server shit, the other two are standard desktop CPUs.

>>60842579
Different anon, but those old X5600 series Xeons do a pretty gud job when overclocked properly. Part of the problem is that there are several ways to OC them with different results.

>>60842598
This is true though. They inhale power like Kirby at a banquet, and regularly hit 75-80C if you're not running on liquid.

>>60842708
Then why would you take a screenshot of it at the lower clockspeed?
That doesn't make any sense for someone who's trying to record scores

>>60842725
The xeon also has fewer PCIe lanes, which matters for multiple GPUs and ssds.

>>60842746
Actually, that is a Evga Sr2, the definition of a HEDT in 2011

>>60842716
I hope you realize those Xeons were released in 2010. Are you underage? If you had a dual CPU setup from 2000, you'd get fucking CREAMED by a mid range single CPU setup from 2007.

If this goes to show anything, it's how pathetic technology progression has been.

Intel wants to make the PowerPC 970 but the low yields doesnt let to do that.

>>60842716
Obvious troll.
X5650 is from Q1 2010

File: rz.png (34KB, 895x582px) Image search: [Google]

34KB, 895x582px

>>60842752

>>60842807
>PPC970
That's the ultimate housefire. Nuclear holohoax in your socket.

>>60842408
>€520 7700k
Did you miss the 769€ version right next to it?
Supposedly the heatspreader is 99% silver

>>60842803
>If you had a dual CPU setup from 2000, you'd get fucking CREAMED by a mid range single CPU setup from 2007.
Irrelevant. Moore's law is dead and has been dead for a long time now.

>>60842640
>launch modern video game on my Ryzen 1700
>uses all 16 threads

>>60842868
Oh, so you're glad that you're only getting marginal upgrades as years go by? Consumer whores are the dumbest fucks out there.

Bet you're one of those people who would have been happy with an Intel quad core in 2020. Lucky that AMD got off their arse and made something decent.

>>60842839
This feels like damage control and not an answer.
If I'm running CPU benchmarks, I'm gonna make sure the clockspeed is identical

>>60842904
This. It was proven that the latest games benefit from multithreads. Intel is over.

File: 1471397846624.png (165KB, 578x691px) Image search: [Google]

165KB, 578x691px

>>60842916
Im stating facts m8. Performance increases are becoming harder to come by. Single core performance is DEAD. Only way to scale is multi threaded and AMD is far better at it than Intel.

>>60842928
during testing the clockspeeds were identical. wtf are you talking about

>>60842928
are you retarded or just shitposting?

https://en.wikipedia.org/wiki/Cool%27n%27Quiet

>>60843058

It is amusing how Intel has spent a decade trying ot perfect SMT and AMD has a better implementation on their first try. Naturally neither hold a candle to IBM and their monstrously wide POWER chips (delicious 4-way SMT on POWER9).

If i'm reading Agner's information correctly there is basically no point to disably SMT on ryzen as it is wide enough that you'd just end up with idling resources rather than a potential gain due to resource contention that sometimes affects hyperthreading (i.e hyperthreading giving negative scaling).

File: cinebench451.png (37KB, 316x1301px) Image search: [Google]

37KB, 316x1301px

>>60842716
Shit, my bad, that's an old one,

>>60843089
>first try

funny how recent AMD BTFO breds reeks of the same level of helpless desperation that Intel BTFO breds did one or two years ago.

File: 93c_1.25v.png (735KB, 1546x630px) Image search: [Google]

735KB, 1546x630px

>>60838297
No silly, it's not from Ross. I got it at TJ-Maxx
4.8GHz
>93c
>1.25v
my fucking sides you have to destroy the warranty on your 4 digit silicon so it can reach previous gen clocks.

>>60843089
I thought SMT lowered ryzen performance just like it did with the original i7s.

>>60843234
How does this shit OC worse than broadwell-e? I mean besides the jizz TIM that the guy delidded to take care of?

>>60843251
I still get a boost running CEmu with HT disabled by setting processor affinity to real cores only on my i7-2600K.

>>60843089
POWER8/9 also uses 8-way SMT.

>>60839106
He founded his own company.

>>60843234
I wonder how a custom loop would fare. Closed loops are junk.

>>60838662
I have this on 3DO or Saturn. Fuckin weird game.
>>60838861
>redpilled
what a fgt
>>60843263
this ss was from pre-delid. After they destroyed the warranty it clocked up another 200mhz at the same temps. I think intel has reached the limit of their arch, it's running well out of it's efficiency range at these levels. It's the Hawaii of cpus

>>60843365
>game
It's a short story.

File: smug elf.png (301KB, 468x494px) Image search: [Google]

301KB, 468x494px

>>60843365
>After they destroyed the warranty
>Implying you don't invalidate the warranty just by OCing

>>60843397
You can buy an OC warranty from Intel where they replace the chip once if you fry it.

>>60843416
forgot link: https://click.intel.com/tuningplan/

File: muhsinglecore.png (454KB, 1920x1200px) Image search: [Google]

454KB, 1920x1200px

>>60842561

File: 1481744513695.png (349KB, 716x2588px) Image search: [Google]

349KB, 716x2588px

>>60842640
>Hurr Durr people buy HEDT systems for non computationally intensive workloads.

File: this is what intelfags actually believe.png (63KB, 1618x220px) Image search: [Google]

63KB, 1618x220px

>>60843435

File: sth_ryzen_benches.png (2MB, 973x6115px) Image search: [Google]

2MB, 973x6115px

>>60842725
Sure it is.

>>60840218
>Nevertheless, what prevents intel from doing the same.
They haven't done any real R&D in years.
They have no talented engineers of note.
And they don't have infinity fabric. An AMD technology.

If Intel want to get serious and regain a stranglehold on the market, they need to fuck off with this diversity hiring bullshit, get people who know which pin is negative on an LED, and get their heads down into some good hard spending of all that revenue.

>>60838438

AMDrones post this picture in every thread in hopes of convincing people though I'm pretty sure they can't replicate the results.

>>60843477
>AMDrones
You mean people with a brain? Just look up everyone's reaction to Intel's recent announcements.

File: 1494090152782.png (495KB, 1070x601px) Image search: [Google]

495KB, 1070x601px

>get $1500 10-core i9 off siliconlottery . com
>OC to 4.8 GHz, requires fusion reactor and custom loop to keep temps down
>Still lose badly in multicore to a $1000 16-core Threadripper
>"b-but muh clocks .."
>Threadripper XFR's to 4.1 GHz anyway

>>60843477
Which results? The crappy 18 core Xeon? Or the 1800X? Because I can guarantee you that you should kill yourself, tripfag faggot.

File: 1491166649444.jpg (66KB, 568x612px) Image search: [Google]

66KB, 568x612px

>>60843531
>Still lose badly in multicore to a $1000 16-core Threadripper

Haha buddy you wish this was real life

File: boi.jpg (18KB, 480x538px) Image search: [Google]

18KB, 480x538px

>>60843491

I don't look at reactions, I look at real life performance.

AMDrones are indians who are paid to promote AMD here while they still have enough rupees to fund this before they finally go bankrupt.

How much do Intel shills get paid? Is it a good job?

>>60843545
TR was already shown with near 1:1 scaling to an 1800X. https://youtu.be/NeuamYdfU4w?t=16m

1800X time on that blender test is usually 24 seconds.

>in b4 fake news
>in b4 pre-recorded
>in b4 it was a xeon running that test
>in b4 tripfag acts like even more of a huge faggot

Threadly reminder that the tripfag currently shitting up the thread is some retard from /o/ who showed up spouting the most inane Intel shilling known to man. https://archive.rebeccablacktech.com/g/search/tripcode/%21%211w3SeuwCso0/

>>60843614
At that point I'm really interested if he's falseflagging Intel or falseflagging AMD (because people will think it's a stupid AMDfag falseflagging Intel).

>>60843658
I don't think I'd even believe it if he broke out the "I was just pretending to be retarded" spiel.

>>60843594

When TR is released and it can perform like that I will believe it. As it has been with every AMD product in the past 5 it won't perform as well as it first appears. This time is no different.

>>60843728
Ryzen was better than expected, especially with agesa 1006

File: 1496793008848.png (1MB, 1440x2560px) Image search: [Google]

1MB, 1440x2560px

>>60843592
They do it for free

File: 1fx2km.jpg (71KB, 888x499px) Image search: [Google]

71KB, 888x499px

>>60843728

>>60843376
You used the MS-DOS game box, not the book cover artsypants.

>>60842904
>>60842934
The latest AAA games do. But this was true 5 years ago as well with a few very well written games. Problem is, no matter how many years pass, games remain few threaded.

Only a very few games utilize threads, such as Battlefield 1.
But these games usually run just fine on any other CPU, since if they scale out, an i3 with 4 threads, or an i5 with 4 cores can already run them well above 100FPS. (even BF1.)

tl;dr: Single threaded games have dreadful performance (ie.: ARMA3), and well-threaded games (BF1) runs fine on a Pentium, i3, on literally anything, even on my fucking cat.

>>60844384
(The reason to this is simple. While you have Unity, Crytek (Lumberyard), Unreal... which they all support threads. Making this work requires extra manpower, talent, and money. If you customize your engine - which is like a must for any game, then making the engine stay threaded while customizing it is a real fucking pain in the ass. Yadda yadda, games and programs will be single threaded or benefit most from single core performance for long years. If you render or simulate or something, go for threads/cores.)

There is a reason why Xeon/Opteron had several slow cores for a long time, because mostly it's the web that benefits from many threads. Web servers, databases, even now VMs.

>>60844384
First, we can only hope there's progression and more games will use more cores in the near future.
Second, games don't really matter in cpu choice. You should choose one for other matters. And AMD is better in everything else.

>>60844425

> several slow cores for a long time

That is only a by-product of monolithic die binning and TDP. There is a reason why Intel make a special 5ghz xeon for the NSA.

>>60842378
>Single thread is 99% of desktop use mate.
stop running dos then faggot

>>60844209
>You
Wrong. Anyway it looks very similar to the cover of the version I read.

>>60839074
Raja Koduri designed Evergreen. He also worked for ATI before he left for Apple, and then he moved back to the graphics department in AMD.

>>60845219
We all know how kickass were Evergreens. Too bad Fermi outsold them, lmao.

>Preselected silicon lottery chip
>90C @4.8GHz
>>>>AMD BTFO

hhahaahahahahaahahahahahahhhaahaahaahahahaha holy fuck intel stop it LMFA0 zidez are in 0RBIT

>>60846174
>delidded
>liquid metal
>no avx
>still 90C

>>60846255
Truly a weapon to defeat 9590.