Official Ryzen Thread - Faster than 7700K, Match with 6900K

>>58899949
>8C, 16 Thread CPU beats a 4C, 8 Thread CPU
It'd be shameful if it didn't.

>>58899949
Intel confirmed for suicide watch

>>58900000

>>58900029

>muh cores

One costs $400, the other is $1,000... the $1000 barely beating the $400 is what you should be impressed by, faggot

>>58900072
Irrelevant.
You can't SLI/CFX CPUs on a single socket motherboard.
Best CPU possible is the only way to get performance.

AYYMD IS FINISHED & BANKRUPT

>>58900103
If you're deployed a ton of them and it actually does have less power usage than intel that's also an important factor

>>58900127
Example
Retard buys FX8350 because muh cheap
Gets 48FPS
Smart person buys i5 K-series gets 60FPS+

The retard says something stupid like: buh muh price is cheapur

The smart person replies with: Cheaper ain't giving you 60FPS, so your shit CPU is irrelevant.

File: 1485888249472.gif (64KB, 720x403px) Image search: [Google]

64KB, 720x403px

>>58900103
>You can't SLI/CFX CPUs on a single socket motherboard.

File: 1477556274138.png (30KB, 154x180px) Image search: [Google]

30KB, 154x180px

>>58900103
>You can't SLI/CFX CPUs

>>58900212
>>58900226
Glad you managed to get the point.

Limited to one CPU= Pick the best CPU, not some shit bottlenecking AMD trash.

>>58900162
>>>/v/ is that way

>>58900248
Yes, you know your way to /v/, now stay there.

File: 1486004450001.jpg (76KB, 723x657px) Image search: [Google]

76KB, 723x657px

>>58900244
>SLI/CFX
>CPU

File: monitors.png (357KB, 682x649px) Image search: [Google]

357KB, 682x649px

What is a dual CPU motherboard for a thousand, Todd?

>>58900072
That's irrelevant for anything other than video encoding, which scale fine per core.

For actual real life applications, single core performance matter. If the CPU needs 2x cores to win in an application that scales fine per core, then in single core performance it is half as good as the other.

Shit is gonna be Bulldozer all over again.

>>58900385
Ryzen has Skylake IPC.

File: dual cpu board.jpg (480KB, 1200x890px) Image search: [Google]

480KB, 1200x890px

>>58900127
>>58900212
>>58900226

>>58900385
>being this delusional
it's pretty much confirmed ryzen has broadwell IPC and there's a chip boosting to 4.0ghz. we don't know the OC headroom but even 4.2ghz is good shit even for your dumb ass babby games.

>>58900396
I think they meant that it wouldn't be called sli/cfx

>>58900162
But the FX 8 core will out perform the i5 in streaming and rendering tests

File: AMD intel hybrid board.jpg (101KB, 500x354px) Image search: [Google]

101KB, 500x354px

>>58900396
GET ON MY FUCKING LEVEL

>>58900447
>not using a term that /v/-tards understand

>>58900385
funny how battlefield gets huge benefit from extra cores last time i checked that wasnt video editing

>>58900646
DICE Stockholm has actually decent coders.

>>58900669
This.
AMD is shit but the game scales wonders through the half cores AMD uses (modules).

>>58899949
>it's not even the top ryzen
wew

>>58899949
I JUST BOUGHT A 7700K FUCKING DELETE THIS NOOOOOOOOOOWWWWWWWWWWWWW

>>58899949
DELET THIS

>>58900861
just delid this 7700k shit and stop complaining, you wont need to upgrade for 5years at least

and here´s your
>you

File: AMD-Ryzen-Chart-6.png (13KB, 1214x745px) Image search: [Google]

13KB, 1214x745px

This shit does not make sense.
i7-6900k faster than i7-6800k and i7-6850k despite being clocked quite lower.
Beside that, I hope this is fake because "Physics" and "Prime Numbers" tests look like real dogshit.
Vishera was fast at some things too (compression for example) but shit in a lot of other benchmarks, this is some kind of bad deja vu.

>>58900962
More cache memory makes for more bandwidth.

>>58900962
I'll answer to myself: 6900k is faster because of higher single core turbo (4 GHz vs 3.8 GHz of the six core 6800/6850k).

>>58901003
That's not how it works.
For example 980x was not faster than an i7-920 clock to clock despite having more cache.

File: 1486094203993.png (83KB, 328x408px) Image search: [Google]

83KB, 328x408px

>>58900103
>SLI/CFX CPUs

The physics tests use AVX so no wonder Intel's lead here with a much bigger FPU

File: singleCore.jpg (138KB, 1716x962px) Image search: [Google]

138KB, 1716x962px

>>58900962
yeah i dont get this either

>people willfully missing the point because he used proprietary GPU terms to describe typical desktop CPU setups
Anything to feel smarter, it seems.

>>58901155
Kill yourself tech illiterate scum

>>58901178
I'm not even him. I just get tired of people being obtuse for the sake of it. Congrats, you ignored the conversation to be autistic.

File: 1472574749467.jpg (27KB, 275x274px) Image search: [Google]

27KB, 275x274px

>>58900498
holy kek what the fuck is that

File: 2017-02-11-121522_1024x768_scrot.png (155KB, 1024x768px) Image search: [Google]

155KB, 1024x768px

ryzen is actually a woodscrews

>>58901272
>sponsored deals
>6700k
lmao

>>58900962
i7-6900K turbos up to 4.0GHz on a single core. The i7-6800K only turbos up to 3.8GHz or something. That accounts for the differences despite baseline frequencies.

>>58900498
Lol

>>58900646
That's one game versus millions of other programs dating back to the 90s that won't get benefits from extreme multicores, other than the ability that you can run one instance on each core (which is most useful for servers, virtualization, and encoding).

>>58901847
these chips already put out more than enough per core performance for those games though

>>58901875
s/games/programs

>>58901847
imagine the performance of pong without multi core support. ryzen is finished

>>58899949

>Official Ryzen Thread - Faster than 7700K, Match with 6900K

>Official Ryzen Thread - Faster than 7700K, Match with 6900K

>Official Ryzen Thread - Faster than 7700K, Match with 6900K

a paid shill at work

>>58901954
and it's (You)!

File: AMD-Ryzen-Chart-2-1000x612.png (61KB, 1000x612px) Image search: [Google]

61KB, 1000x612px

>>58901954
This is the 1700x benchmark. Imagine the 1800x...

>>58901987
>>58902006

a very desperate paid shill

File: 1670b00cf2535c4663b86b366e11287fd698a37833784e1eff6783cfc201e295_1.jpg (71KB, 714x820px) Image search: [Google]

71KB, 714x820px

>>58899949
INTEL BTFO
N
T
E
L

B
T
F
O

>>58902006
The integer tests are the only ones that web server providers will give two shits about (assuming restrained power consumption at least).

Intel wants a single line of Xeons to do everything from integer loads to moderate numeric processing, and they've left the door open for AMD to release cheaper parts that skimp of AVX to just focus on integer performance.

I don't know how Intel will do with their Knights Landing AVX-512 monsters, but they got caught flat footed with Zen it seems.

>>58902104
THe integer tests are by far the most relevant since integer performance at Intel has been at a stillbirth for 6 years and game engines are mostly integer.

>>58902048
Can you elaborate?

File: image.jpg (103KB, 1500x1500px) Image search: [Google]

103KB, 1500x1500px

>>58900103
>sli/cfx CPUS

>>58902133
Besides physics but game engine physics aren't really 512bit unless the developers are complete fucking retards.

>>58900498
Hey stop stealing my 9 year old ideas!!! I thought of 3D motherboards before it was cool.

>>58902226
Time for you to think of 4D mobos now.

>>58902236
>Where do I put the CPU?
>"Where?" Don't you mean "When?"

>>58899949
>>58902006
fuuuuuck. as long as you don't give two shits about AVX at least, Ryzen is making the 6900/6950 look like straight scams.
$400/$500 chips shouldn't be anywhere near this close to $1k/$1.5k ones.

I really hope that AMD puts out a 16c/32t workstation platform soon in either 1S or 2S.

>>58902325
>near this close to $1k/$1.5k ones.
Because Broadwell-E are not that costly. You're just too used to Intel's kikery.

>none of these chips only support Windows 10
remind me why anyone gives a fuck?

>that integer slaughter at lower TDP

Oh the server market is gonna have a field day with this.
Especially the tele/SIP, data/web, proxy, media, etc, etc too many to fucking count.

DELET

What's the i7-6700K equivalent and how much less will it cost?

>>58902363
This is just the eight-core version. Just imagine the slaughter with the 32-core Naples, let alone TWO of the motherfuckers.

>>58902363
Does Intel have anything like presumable top Naples chip?

>>58902420
Some undisclosed 32 core skylake at 205W TDP kek

>>58902420
Intel Xeon E7-8800 v4s.

>>58900029
the 6800k is 6c/12t you retard

>>58902433
That's only 24c.

Well Intel... time to get working. Bumping up the performance by 2% every year won't do it anymore.

>>58902439
>Well Intel... time to get working.
They started, by firing 12k workers.

>>58902433
There's no way a 24 core at 165W TDP is going to come close.

>>58902436
>>58902444
That's my point :^)
But what's this I'm hearing about Naples being 150W-225W TDP?

>>58902493
Supposedly so, it's 4 Zeppelin dies in MCM with lower clocks and the best possible binning.

File: 1457804763369.gif (2MB, 290x260px) Image search: [Google]

2MB, 290x260px

>>58902443

>>58899949
>Sorting
what kind of sorting?

>>58902493
Didn't hear about 225W, but I guess those will be specialized with ultra-high turbo for ST performance in serial parts of the workload.
Expect them to cost more than Beyonce's bag.

File: 1451862698912.jpg (48KB, 500x655px) Image search: [Google]

48KB, 500x655px

>>58900103
>SLI/CFX CPUs

>>58902493
>>58902530
Some anon said nipples was going to be ~180w. current xeons go up to 165w, but on their new 3000pin socket it'll probably go up as well.

>>58902506
the beauty of it is that amd can sell all the high leak current dies as a "full power" consumer parts and only use the lower power draw dies to make their quad-die mcm monsters. 3.4ghz 32c at 225W is fairly realistic rumour

File: 9d7312ab91d7c638c9760dd5aec3f46ef3c2aa4c300b9ea9b2fb61769c435311.jpg (76KB, 680x510px) Image search: [Google]

76KB, 680x510px

>>58899949

Check this, fags:

File: 145e38f6d04423e49901763c98e36396ffce2b5a50dd64374a94908bd28dc79a.jpg (410KB, 1280x960px) Image search: [Google]

410KB, 1280x960px

>>58902676

>>58902664
>3.4ghz 32c at 225W is fairly realistic rumour
That kills Intel's server market.

>>58902676
>>58902688
So the PRO variants will ship with worse thermal solutions, but be rated for higher operating temperatures?

>>58902676
>all of the quadcores are stuck at 65W TDP
That ain't good...

>>58900103
A-are you clinicaly retarded?

>>58902801
If you want a low power chip, you'll have to wait for Raven Ridge APUs. Ryzen is the "enthusiast" line aimed at performance.

>>58902864
No, that's my point. The TDP is way lower than I'd hoped. I have a 65W TDP i7-5775C and it can't go beyond 4.2GHz. This shows that the quad cores will not overclock as well as the full Ryzens.

>>58902894
>No, that's my point. The TDP is way lower than I'd hoped. I have a 65W TDP i7-5775C and it can't go beyond 4.2GHz
Are you retarded?

File: sddefault.jpg (35KB, 640x480px) Image search: [Google]

35KB, 640x480px

I want a chip to replace my i5 2500 for gayming that's sub $200. Will an upgrade finally make sense?

>>58902913
im in a same boat with my 3570k. It might be that Ill just buy better cooler and overclock some more. Fucking hell

>>58902910
Are you? Lower TDP means a more aggressive thermal management (read: thermal limitation) setting.

>>58902928
I've got the non k. It was off a Dell build :[

>>58902938
no. the 5775c was a poor overclocker because the way the chip was built, edram and all.
Thermal Design Power has nothing to do with overclockability.

http://valid.canardpc.com/0i1fr6

>>58902521
Sorting of rare pepes for ultimate shitposting.

>>58902894
TDP doesn't matter when overclocking.

>>58902676
>Ryzen 3 1200X
Wonder how much this poorfag part will cost

>>58903036
Cheap as peanuts, it's wasted silicon otherwise.

>>58903047
100$ Ryzen?

>>58903052
Maybe. I'd say 120-130$ for cheapest quadcore.

>>58903064
Will there be any hyperthreaded dual cores ? I want to build a CPU+Mobo+Used RAM+GPU combo for under 300$ this year.

>>58903106
>Will there be any hyperthreaded dual cores ?
H2 2017 with the release of Raven Ridge.

>>58903106
We'll probably see them when the APUs get released

>zen cpu/polaris gpu laptops
CAN'T WAIT

>>58903212
Don't forget possible HBM2 on die.

>>58903106
>dual cores
there is no 2c planned

>>58903265
For Zeppelin. We are talking about zen-based APU line here.

You people do know that a difference of 10% IPC in synthetics barely translates to 3% in games?

Because surprise, surprise, no game out there runs the entire frontend of the core at 100%

>>58901888
kekd and checkd

>>58903413
Real world doesn't run on gaymes, kiddo.
>>>/v/

>>58903451
He's just using one example of a real program vs synthetic dick measuring

>>58903451
Oh then Intel is finished because it's getting slaughtered in integer loads which are around 85% of all applications on the planet and at a lower TDP.

Zen looks amazing.

>>58903467
>Zen looks amazing.
>looks
You said it yourself buddy. This is going to be another AMD upset.

>>58903467
>Oh then Intel is finished because it's getting slaughtered in integer loads which are around 85% of all applications on the planet and at a lower TDP.
Why has intel been so focused on floating point the last few years?

>>58903526
I'm usually pretty happy with AMD

Fury wasn't as good as hoped and polaris is taking longer than I've hoped

>>58903536
Because Intel knows better. Most programs are still FP demanding these days, this anon is full of shit.

>>58903536
Because integer performance has been impossible to improve the last half a decade due to their architecture, since those ALUs are already underutilized and are practically sitting on their asses outside of synthetic benches.
And as FP is the only x86 metric left, they have to improve that.

When it's coming? I wanna buy cheap i7's

>>58903592
This. Nobody fucking cares about servers. I want to upgrade my desktop PC. For fuck's sake why don't you start talking single user scenarios for once.

This board is complete shit. It's not giving me any useful information on home computing.

>>58903646
Video games belong on >>>/v/

>>58900103
DAT logic. Only dumbshit Intel nigger would say that. Enjoy burning your cash for diminishing returns.

>>58902664
I want one. It would be lovely upgrade from my dual e5-2670 vm lab machine.

>>58903673
And you belong on a tree, with a rope fastened around your neck, shitlord, with all your buddies who can't fucking talk about anything that isn't enterprise level shit!

>>58903718
I hate you /v/.

>>58903673
Video games are tech, just deal with it already.

>>58903673
there are tons of non-server usecases that aren't games. just because servers are the only place where Linux has a decent marketshare doesn't mean everything else is irrelevant.

>>58903810
You sure you're on the right board, mate?

>>58903810
There's only Desktop(laptop), Workstation and Server.
And recently IoT if anyone cares.

>>58903842
Isn't IoT just another hip buzzword for embedded?

>>58903842
workstations often use consumer chips with better I/O on the boards.
also IoT is literally botnet now

Waiting to see the war between Intel and AMD like when the K6 was released. Imagine the scramming they'll have to do to core arch in order to catch back up to AMD again. Tsk tsk, sitting on your ass was a bad idea Intel.

File: 1457929666775.png (117KB, 649x480px) Image search: [Google]

117KB, 649x480px

>>58903925
>Imagine the scramming they'll have to do
more like scamming

>>58903925
> Tsk tsk, sitting on your ass was a bad idea Intel.

Not at all. Milking the consumers without any investment is an absolute smart thing to do. Now they got the money and a clear goal to achieve, instead of hunting shenanigans.

>>58903892
12-16 core Zens seem perfect for workstations, two dies should provide at least 50 lanes, which is enough.

>>58903996
Intel just fired 12k workers. Something BAD is about to happen with Intel. Or they'll neck Krzanich.

Quad core Zen benchmarks when

File: 1396929868287.jpg (143KB, 638x680px) Image search: [Google]

143KB, 638x680px

>>58903996
>Now they got the money
Intel. Pay the 1 billion dollar fine you owe the EU so they can discount AMD laptops.
The longer you wait, the better the laptops will be. Giving AMD a greater market share.

>>58904018
everyone is jacking off to the 8 core ryzens, i bet the 4 core versions wont leak

>>58904018
It's current year and you want quadcore?

>>58904069
What if he wanted to upgrade Grandpa's block of rust?

>>58904085
Get a dual core pentium.
One core for solitare, the other for telemetry.

More than enough

>>58904090
But Win10 is already demanding two more for telemetry

>>58904069
It's not that we want quadcores, but given the yet to surface prices of the eight core a lot of neets will only be able to afford their quad core offerings.

I mean for gaming an i5 doesn't bottleneck shit, I suppose a 4 core ryzen CPU should be on par with that.

>>58904085
>What if he wanted to upgrade Grandpa's block of rust?

Bullshit, the only legit excuse is being a poorfag and being cut already by am4 mobo.

File: thCSaDS.gif (3MB, 500x281px) Image search: [Google]

3MB, 500x281px

>mfw about to drop cheap octocore chips and BTFO (((Intel)))'s profit margins

>>58899949

wait wut the 6700k>7700k?

>>58904131
Not yet. But Naples is around the corner, and Intel have no answers for price and sheer integer perf of Zen.

>>58904103
Overclock it.

File: 1468556194055.jpg (213KB, 1024x1024px) Image search: [Google]

213KB, 1024x1024px

>>58904150
>thread 6 hours old
>had 150 posts
>thinks he's the first to notice that
READ THE GOD DAMN THREAD BEFORE POSTING, YOU GOD DAMN FAGGOT

>>58904191
Are those some kinda mascots I missed or just some random AMD fan work?

>>58904219
tech companies often make Japan exclusive mascots, wouldn't be surprised

File: BULkKRjCcAAyH46.jpg large.jpg (190KB, 1024x1024px) Image search: [Google]

190KB, 1024x1024px

>>58904219
I'm not sure. But it appears that moar cores = moar whores.

>>58904259
>that smug hipster on the bottom left

>>58900103
I purposely turned on my mom's boyfriend's shitposting machine just to tell you to kys yourself to peaces.

>>58899949
verdict according to these leaks
>single core perf, while a major leap ahead for AMD, is behind *lake
>perf in heavy AVX or memory IO tasks is lower than the X99 stuff, which is to be expected

since this isn't the flagship ryzen and the chip was held back by lack of boost, slow memory and possibly crappy VRM, we can reasonably expect the 1800X to perform much better than this across the board. I predict by a factor of 10% or more.

basically, the 1700X with boost disabled puts itself exactly where I expected Zen to perform a year ago. I would have been content with these results if they were as good as it gets. they aren't, and that's pretty exciting.

and seeing as this thing will most likely sell for 350 bucks, it's an absolute fucking monster. Intel's mainstream chips now have ONE thing going for them, and that's single core perf, the benefits of which are nearing their timely end.

>>58902676
>>58902688
>tfw the highest end Ryzen 8c is the one i want
>tfw it only supports 60C TCase max
why is this world unfair, /g/?

how does jim keller do it?

how did he single handedly make intel his bitch?

File: certified shit wrecker.png (400KB, 816x1020px) Image search: [Google]

400KB, 816x1020px

>>58904502
It's like gravity; nobody can explain this shit.

>>58904502
He is a certified Shitwrecker that's how

>>58904502
Twice.

>>58904559
You mean thrice with Apple's A5 and whatever designs he left for them after that.

>>58904577
Not Intel in that case as much as all the other ARM competition.

>>58904592
Trust me Intel felt it the most.

>>58902339
>Broadwell-E are not that costly.

6900k launched at $1100, and 6950k at $1710, m8.

>>58904690
Not that costly to produce, fella.

>>58904715
Yes, but Intel never drops prices, even if they're straight up outclassed

How is the single threaded performance?

90% of my gaming is emulators, so that matters more than overall performance using moar cores.

File: 1486812876914.png (186KB, 1920x974px) Image search: [Google]

186KB, 1920x974px

>>58904766
>>58904715
This is their defense to Ryzen

>>58904766
Then they are doomed to fail.
>>58904788
Bretty good, like Broadwell-E levels at very least.

>>58904805
So still 2 or more generations behind Intel?

>>58904800
>Optane
It's still alive? Dammit.
>>58904836
Sweet child, do you know what Intel's """""""""""""""""""""""""""""""""""""""""generation""""""""""""""""""""""""""""""""""""""""" means?

>>58900029
By that logic, the 8350 should be better.

>>>/v/

>>58904069
Would be good for an HTPC/emulation build

File: 1381760132451.png (71KB, 191x226px) Image search: [Google]

71KB, 191x226px

>>58900103

>>58904788
its good, but can't seem to beat the 7700K

AMD looks like it has a winner on their hands, it's doing much better than their previous stuff but like previous AMD products it excels in things that are heavily threaded but intel will likely retain the crown for best over all performer.


I will wait and see until proper reviews are released.

>>58904897
No iGPU.

File: 1484450077402.png (783KB, 1300x1080px) Image search: [Google]

783KB, 1300x1080px

>>58900103
>You can't SLI/CFX CPUs on a single socket motherboard
You can't do that on any board.

>>58904836
Yeah, only 5% to match Intel's chips.

>>58904954
I wont need one :~)

>>58905062
Then you're good to go.

File: Society-of-the-spectacle.jpg (71KB, 466x600px) Image search: [Google]

71KB, 466x600px

Why are Intelfangays so fucking upset they're going to have cheaper CPUs?

File: amadaexcite.png (38KB, 422x537px) Image search: [Google]

38KB, 422x537px

>>58904951
They don't need to beat Intel's best CPU. They only need to beat most of them on price and come in equal on performance.

>>58905115
Because they are not going to have cheaper CPUs unless they go AMD.

>>58905115
Because they already paid out the (((nose))) for the overpriced ones.

>>58905115
>Intelfangays
A true master at the English language there, pajeet

>>58900029
>Moving the goalposts

>>58905149
(((you)))

>>58900103
>SLI
>A CPU

What? I don't know what the fuck you're trying to say but no. Whatever you're ACTUALLY thinking of isn't true. SLI and CrossFire has nothing to do with the CPU, and most users don't use SLI or cross-fire on high-end cards since price / performance is crap. The only people who do it are the ones who like to throw their money away.

File: 1486791376899.png (168KB, 455x438px) Image search: [Google]

168KB, 455x438px

>>58900103
Autism

>>58900162
What about how the 6900k only provides a 10% performance increase based on score but costs literally twice as much?

File: 1484353081022.png (402KB, 588x598px) Image search: [Google]

402KB, 588x598px

>>58900103
>He doesn't SLI his CPU and GPU

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

224KB, 799x540px

>>58905424
Don't worry, Intel will give you a 25% off coupon. Valid 1 week.

/People hate it when something cheaper comes along and blows the fuck outta what they paid through the nose for the thread.

>>58900962
People on reddit are saying that the person who made the chart is claiming that some of the Intels were overclocked and that isn't shown on the chart.

I think these charts are just lies, desu.

>potentially dual 1800Xs in a $2000 pc
jesus

I'm planning to do a build with AMD CPU, AMD GPU, AMD memory and AMD SSD.

Too bad they don't make AMD brand motherboards.

>>58906589
Too red.

>>58906604
AMD's color has always been forest green, not red.

>>58906604
Which motherboard is reddest, MSI?

What percentage does /g/ estimate Intel will reduce their prices by?

>>58906634
Zero.

>>58906619
And now its black and white.

>>58906634
jvvst enough

>>58906634
haha

>>58906634
You're too young if you think dropping prices is what Intel will do.
Intel doesn't drop prices, period.

>>58906634
Zero. Intel never lowered prices before, even during K8 times when AMD was raping them in the ass with no lube applied.

>>58906641
You lack business sense.

>>58906664
You lack basic industry knowledge, and my opinion has no impact on Intel.

>>58906674
>if it didn't happen before, it can't happen

Yeah yeah, just wait for that Intel Price DropTM in a few weeks, I'm sure you'll enjoy your discounted $350 i7.

>>58899949
How good is the single core performance?

>>58906659
cool story but that is completely wrong

intel is already preparing for ryzen by adding hyperthreading to pentiums and later to i5 models so their cpus look better than they do now on paper

>>58906685
I'll direct to you to the famous 60% Intel margins, and price drops can't maintain those margins.
Ergo no price drop.

>>58906699
Linking for downs toddler, btw you need to click: >>8901118 to see the picture where it says "Single Core Performance." I hope this helps, also please get your parents permission before trying to build a computer. In fact I recommend you go pre-made since you seem to have trouble with the search functions.

>>58906720
>preparing for Ryzen with a dual core chip

Uh yeah.
Great going Intel

>>58906699
Some friend of mine is selling me 32gb of ddr4 ram at 2133mhz at a good price.

Is the ram speed compatible with the new AM4 motherboards?

>>58906720
Pentium is okay I guess, but HTed i5 for fucking 2066 is absolutely hilarious attempt at damage control.

File: canvas.png (9KB, 300x500px) Image search: [Google]

9KB, 300x500px

>>58899949
WHAT DID JIM KELLER MEAN BY THIS
http://en.genzu.net/yurai/?n=Jim%20Keller

File: 1465887734989.gif (480KB, 493x342px) Image search: [Google]

480KB, 493x342px

>>58900103
>SLI/CFX CPUs

>>58906730
It could do some damage against the 2 core APUs I guess.

>>58906744
His arch is about to penetrate Intel's anus again.

>>58906754
Said APUs would probably be more efficient and they have better iGPU and that probably matters for cheapo boxes.

>>58906735
Why for the HEDT socket and not the mainstream one? What's the fucking point unless they don't want to invalidate their entire lineup with a cheap OCable i5 and make people pissed they bought something twice as expensive with the same performance.

>>58902286
https://en.wikipedia.org/wiki/Four-dimensional_space

>>58906791
They need something for new platform and Skylake-(e)x is ahead. Basically Intel thinks usual jewery will werk.

>>58906634
You mean

>What percentage of the desktop manufacturers does /g/ estimate Intel will bribe?

>>58906814
It'll be slightly more difficult this time but possible with Trump's help.

How well will these guys sell

>>58906814
Noooot this time. Zen-based APUs are too good to skip.

>3.4ghz Ryzen handedly outperforms 5ghz Vishera
>massive int performance
>perf/clock basically neck and neck with Broadwell
>even the floating point perf is massively improved
>even with gimped memory

Whats going to happen now is that AMD and intel are both going to start shitting out incrementally higher chips, at very fast intervals. Just like they used to do. They can't rush out new arch, so they'll focus on trying to get yields higher so they can release progressively higher binned chips.
We could see the average high end desktop chip with clock speeds of 4.5ghz within a year or two. Zen+ is going to be fucking wild.

Just like with the old Pentiums and Athlons. A new SKU every other month.

>>58906828
>Vega+Zen APU
>$200 for i5 7600k+gtx1060 performance

>>58906870
Surely not 1060, but ~1050 is possible.

>>58906733
PLEASE SOMEONE ANSWER!
I WILL HOLD HANDS WITH YOU!

>>58906733
>>58906902
2133mhz DDR4 is the lowest speed of the specification. It works in literally everything that supports DDR4.

Its probably a really low end junk kit though.

>>58906902
Yes, of fucking course AM4 supports DDR4 of said frequency.

>>58906914
>>58906921
Thanks I did think that IT only supported ddr4 at 2400mhz.

File: 6543528721873.gif (3MB, 410x307px) Image search: [Google]

3MB, 410x307px

>>58906914
>>58906921

>>58906685
Intel has been in the same position before and never dropped prices.
They will never drop prices.

File: 1405939015993.png (302KB, 450x488px) Image search: [Google]

302KB, 450x488px

>>58906985

>>58904069
Personnally I don't need more than an i5 so I won't buy those 8c ryzen. I'm waiting to see what they are going to release for "poorfags"

Useful info:
http://semiaccurate.com/forums/showpost.php?p=281929&postcount=6391
Memory latency on Passmark scores

##1066 10-11-10
CPU Mark This Computer 9230
Integer Math This Computer 19408
Floating Point Math This Computer 8121
Prime Numbers This Computer 19.8
Extended Instructions (SSE) This Computer 225.8
Compression This Computer 14193
Encryption This Computer 2024
Physics This Computer 359.7
Sorting This Computer 8723
CPU Single Threaded This Computer 2370

##1066 7-7-7
CPU Mark This Computer 9932
Integer Math This Computer 19862
Floating Point Math This Computer 8305
Prime Numbers This Computer 25.5
Extended Instructions (SSE) This Computer 227.5
Compression This Computer 15206
Encryption This Computer 2078
Physics This Computer 413.9
Sorting This Computer 8589
CPU Single Threaded This Computer 2327


##2000 10-11-10
CPU Mark 10673
Integer Math 19504
Floating Point Math 8336
Prime Numbers 31.5
Extended Instructions (SSE) 219.1
Compression 14806
Encryption 1944
Physics 597
Sorting 8577
CPU Single Threaded 2358

The aggregate CPU Mark score, Prime Numbers, and Physics scores are significantly effected by memory performance.
The Ryzen system tested was equipped with 2400mhz 17 17 17 39 DDR4. The latency is really poor, and thats lowering performance in those metrics substantially.
Overall the performance of Ryzen looks incredible.

>>58907395
I hate Semi Accurate but this post is legit

File: Drake - Hotline Bling.webm (3MB, 1920x1080px) Image search: [Google]

3MB, 1920x1080px

Who are future /Ryzen CPU + Nvidia GPU/?

>>58907515
heresy

>>58907515
>nVidia
It's like you hate open source drivers or something.

>>58907584
I don't get it, because Radeon are better?

>>58907561
AMD tested RYZEN with a TITAN X at the CES...

>>58907601
amdgpu is the driver released and supported by AMD. The radeon driver is deprecated and all development is now on the amdgpu driver.
They are also open sourcing various debugging utilities for monitoring API calls and utilization.
AMD is going pretty hard on the FOSS right now

>>58905203
>thing I don't like
>autism
pick two and only two

>>58907623
Well duh, they wanted the most powerful gaming GPU that demonstrated that Ryzen is good enough that it won't bottleneck even under the most CPU intensive gaming loads.
They didn't use Vega because it's not a stable platform yet, and they didn't want to give too much info about Vega at CES.
Vega is still in very early R&D, and it wouldn't make sense to show off Ryzen with Vega, as it's just another variable in the mix. Apples to apples should be done with well known and tested hardware, which is why the Titan XP was chosen.

>>58900103
Nice one.
Haven't seen a good troll in a while.

>>58901309
>5%
marginal gains in late capitalism

>>58907676
So use a the 1700x/1800x with my 1070 will be a good idea

>>58907469
It'd be decent if a couple posters were banned, but it'd also be totally dead. The main site is just so-so.
Charlie's speculation articles aren't that great, and he usually only has a tiny bit of insider info which he far too heavily extrapolates from.

>>58907728
A very good idea.
The CPU shown off at CES is probably comarable to the 1700 (not X), so the 1700X should not bottleneck anything up to the Titan XP.

File: Coping with excessive sweating.jpg (27KB, 463x265px) Image search: [Google]

27KB, 463x265px

You guys need to calm down and ignore all benches except AMD's. What we do know is that AMD's 8 core can perform just as well if not better as a 6900k on handbrake and blender. The other stuff is just noise to me. It'd be amazing if they somehow have Broadwell-E single core performance since that is basically Skylake performance, but now with 8 cores. All I want for them is too succeed and make a gaping hole in Intel's ass so I can finally upgrade the 3570k. Beat it at single core, and I'm switching immediately.
>mfw Zen+ might be 7nm
>mfw possibly building a Zen router
>mfw definitely building a Zen NAS
>mfw similar performance to a 6900k at lower prices

>>58907871
Greyhawk (gpu arch) is confirmed for 7nm in 2019.

>>58902764
>>58902688
>>58902676
>>58904335
The X variant chips have XFR enabled.
95w XFR enabled chips have a lower tcase because of how temperature effects leakage and switching characteristics of a transistor. The lower leakage is, the more thermal headroom they have to exploit. Its not anything about the silicon that prevents it from hitting higher temps, the dies are all the same, this segmentation is about preventing a marketing snafu. You just need to keep the die in a lower temp range to exploit the XFR feature.

Pro variant chips are business SKUs, exactly has been in the past.
Pretty simple.

>>58907906
so knwoing amd from past experiences, we could expect 7nm by 2025

Could someone possibly dumb down why Intel's Trigate is so bad and troublesome compared to everyone else's processes for laypeople like me?

>>58899949
INTEL ON SUICIDE WATCH

>>58907961
Global Foundries is bringing IBM's 7nm node online later this year. The ramp up to risk production will be in 2018, they'll have parts on the market in 2019.


>>58907980
Intel's Trigate FinFET processes have been very good, for its time their first gen 22nm node was actually ridiculously high yielding, and very early on. They didn't have any issues until their 14nm node. From their 22nm process node they basically squeezed in a node and a half worth of area scaling so they could outpace other foundries. That immense leap in area scaling, having extremely high density, led to very poor yields, even with small die parts.
They first launched the process with Broadwell CoreM which went into production mid 2014, and was available on the market Q4 2014. The product lineup was a total mess though. There were dozens of of SKUs with the only difference between them being incredibly small clock bumps, the type of thing you see when a foundry can't produce enough dies that hit target clocks at a given power level. They had non stop issues after that for a solid year.

The actual characteristics of their 14nm process are very good, and by now they have no problem producing high volumes of high clocking dies. If you look at the power per core used at a given clock speed it shows extremely well for them. They just bit off far more than they could chew, and it took them a long time to sort out. Unfortunately they've done the same thing at 10nm though.

It looks like a prime example of an executive setting unrealistic goals for engineers.

>>58907933
Link to source of that? I'm really curious because the biggest reason I was going to invest was for XFR and if it's only going to be enabled on certain chips I'd like to know.

File: 11.jpg (296KB, 720x540px) Image search: [Google]

296KB, 720x540px

>>58908078
https://arstechnica.com/gadgets/2017/01/amd-ryzen-motherboards-hype/

File: AMD Vs Intel Updated for Ryzen.png (71KB, 1605x295px) Image search: [Google]

71KB, 1605x295px

It's done

Seriously can't wait to upgrade my FX 8 core to a Ryzen chip of some kind, not sure I need 8 cores/16 threads but I should just go that route anyways shouldn't I...Really what I think I'll get is the higher clocked 6 core Ryzen chip, rather than a lower clocked 8 core that may not overclock very well

>>58908158
This doesn't seem to mention XFR being locked to specific models, just how it works.

>>58908207
I kind of want to get the 1800X just because
>muh cores
for
>muh VGApassthrough
But, I also want to use either my NH-D9L, or get an NH-L12, so a 6-core might make sense from the thermal perspetive

>>58908276
a 240mm AIO is only around $100 man

>>58908265
XFR can only exploit remaining thermal margin if remaining thermal margin exists. It doesn't exist if you have a 95w chip with a mediocre cooler.
Those 95w X SKUs have the HS81 (F) designation, showing a lower tcase, because they need to be kept cooler for XFR to work. The only other X SKUs are 65w parts, significantly easier to cool, less cores active.

>>58908296
Yeah, nah, fuck watercooling.
I want to throw this together and not pull the cooler off for 5 years, and I can't be fucked dealing with the risk of AIO's fucking themselves up because of chemistry.
Also, I like the L12, because the downwards facing airflow goes to all the other components on the motherboard, so I don't need to worry about VRM and Chipset temps. I've fried an AMD 970 because the heatsink wasn't getting any airflow and was getting to 130-140'c

>>58908351
Except you're throwing hot air at those components...

Also is noctua shipping AM4 brackets for your units? because AM4 uses a new socket

File: 1469189275455.jpg (57KB, 1000x667px) Image search: [Google]

57KB, 1000x667px

>>58908351
If ya going ryzen you want water, or a massive fucking heatsink. Like a block of copper. Ryzen auto overclocks with thermal overhead in mind.
On air you'll keep throttling it. But water has a high amount of material that can absorb heat before peaking die temps.
Ryzen will BTFO air cooling with 'burst' overclocks.

File: HR 22.jpg (150KB, 602x635px) Image search: [Google]

150KB, 602x635px

>>58908351
>>58908409
He should just get an HR-22 or Le Grand Macho instead

>>58908371
new bracket*

>>58908067
>7nm
The final shrink we'll ever see

>>58908371
>hot air
Airflow is airflow
~50'c flowing air is better than ~90'c stagnant air.

>>58908409
I am aware of this, I am not looking for peak possible performance. I just want great performance and an easy to manage system.
Air is fine. Look at how well you can run the 6700K on the NH-L12.

>>58908418
>ITX build

>>58908418
You don't understand. That block is tiny. It'll choke Ryzen. You're gonna need a block that will stay cool for a long time before peaking under load.

>>58908471
noctua nh-15

>>58908459
Nah, foundries will probably squeeze out 5nm and 3nm GAAs on silicon before switching to something else in the 2020s.
No clue what Global Foundries will do though. IBM would have had more IP developed beyond their 7nm SOI since that was already a completed process with validated test chips ran. Whatever they have in the works is still closely guarded secret.

>>58908467
>ITX build
the Core 500 from Fractal supports 240mm AIOs in an ITX rig, it's the best cooling you're going to get in that type of rig

>>58908512
>240mm AIO
It's like you can't even read.

>>58908467
I still suggest an 240mm AIO. The water will stay cool for when you need to dump heat and net you a 'free' overclock lasting for 5mins.
I know it's scary to put a cooler in your expensive electronics, but they've put a lot of work into making sure they don't leak. Because if they do the company is responsible for damages.
>>58908500
You're still not of the understanding.
The actual block on the cooler needs to be large to give a thermal overhead. Around 50x50x50mm of copper. But it'll still fail in comparison to water loops.

>>58908523
Why would you want inferior cooling for your $500 CPU?

File: 3.jpg (136KB, 720x540px) Image search: [Google]

136KB, 720x540px

>>58908158
>l2/l3 cache have independent voltage supply from other parts of the core and die
I wonder just how the NB will operate. It'd be amazing to clock caches higher without breaking USB/SATA/PCI-E functionality of the FCH.

>>58908542
Get the fuck out of here, retard.

File: 1324451596202.jpg (33KB, 480x360px) Image search: [Google]

33KB, 480x360px

>>58908584
Hot goes into cold faster than it does into warm. What don't you get?

>>58907741
Charlie is THE cancer

>>58908542
I'm not too concerned with leaking. It's more to do with pump failures, or corrosion.

>>58908597
Things aren't black and white.
A 150w rated heatsink+fan sitting atop a 95w rated chip is providing tons of headroom. Saying an enormous 200w+ rated heat sink is too small is simply trolling.

>>58904116
I'm thinking about getting a cheap quadcore and a cheap mitx/uatx mobo and building a small NAS with a buttload of drives. Use a node 804 w/o a window if possible (haven't looked). I only think the 804 would be good since I know it holds a shitload of drives.

File: 203c3b5fa61698069c2413c5b7ce276f51c32f6d3f65bfb1ee14b4e095e5e5da.png (160KB, 800x310px) Image search: [Google]

160KB, 800x310px

thoughts on the stock coolers?

>>58908658
Corrosion shouldn't be an issue if the loop is all copper, it's mostly a problem when you start mixing metals, like copper and aluminium IIRC. I've had the same loop for like 7 years and have had 0 issues with it.

>>58908658
Pump failures do happen, but a shutdown will happen the moment thermals run away.
This is why I've kept the stock cooler I've got to hold in mean time should I need to RMA.
As for corrosion, that's not an issue in newer AIO's. to my knowledge.
>>58908672
But you're still looking at this for keeping it within a range, not keeping it as cool as possible for as long as possible. These chips automatically overclock. So they can attain speeds that would normally BSOD you for a temporary amount of time, PROVIDING you give it the ability to stay cool. Not 70c cool, but as near room temp for as long as possible. Overclocking scales with temperature.
Say at 20c it'll hit 4.3ghz. but as temps raises to 40c it'll go back to 4ghz and as it hits 60c back to stock speed.
An air cooler is only capable of removing so much heat at a time. But with air, it'll actually store that heat over a wide space allowing you to stay at a cooler temp for much much much longer.

>>58907676
Vega isn't early R&D, its literally about to enter production for a may launch

File: guy.jpg (322KB, 600x858px) Image search: [Google]

322KB, 600x858px

>mfw not even getting full use of a stock 3770k

>>58908763
MEME RGB LED-tier
But damn, a 90mm fan for a 65W TDP heatsink sounds pretty good to me.

>>58908774
>Vega isn't early R&D
HAH
It might be now, but the CES board was pre-production, and not stable.

>>58908763
Always replace stock coolers / 10

>>58908776
Good for you.
Some of us actually use our computers for more than facebook and CS:GO

>>58908773
>>58908672
>But with air, it'll actually store
But with water***

>>58908795
Well AMD says 1H 2017, guess we will wait and see. Things are looking in line for ryzen to release March 2-3, just like AMD said before...

>>58908805

I'm mainly using photoshop and 3d modelling tools

>>58908798
But warranty only applies when using stock.

>>58908542
>You're still not of the understanding.
Fuck AIOs. Custom loops or big ass heatsinks.

File: 1353242491957.png (148KB, 300x300px) Image search: [Google]

148KB, 300x300px

>>58908854
Stop telling them you do the things.
>>58908861
What are you basing this fuck AIO's off? Did your mum not let you have one?

File: Noctua fan on a PS4.jpg (258KB, 882x1052px) Image search: [Google]

258KB, 882x1052px

>>58908873
>What are you basing this fuck AIO's off?
AIO is half-assing it. If you're going to AIO might as well get a big ass heatsink. Custom loop is superior even if it is pricey.
Speaking from personal experience, faggot.

File: 1463649555491.jpg (71KB, 600x450px) Image search: [Google]

71KB, 600x450px

>>58908898
Then you haven't read the reasoning why an AIO is better in this situation.
Sure a custom loop would be even better because it'll have more water to soak heat giving longer boost frequency's. But air is inferior because these chips auto overclock based on thermal overheads.
Air's fine for traditional chips and overclocking, but these are very different in how they clock.

>>58901089
Disable HT on a 6700K, clock it to the same speed as the 6000K.
Compare them.
There you go.
The i7 now wins because it have more cache.

>>58908552
>per core voltage and frequency control

So all cores can run at different frequencies??!?

>>58908962
They're capable of it.
Zen can dynamically clock single cores, or an entire CCX depending on the workload and available headroom.
A chip with a 3ghz base clock could hit 4ghz on a single core, 3.9ghz on two cores, 3.8 on four cores, 3.6ghz on all cores, etc.

this is the first time i've been remotely excited about AMD in over 10 years

considering those benchmarks in OP are done with absolutely dogshit RAM and probably supporting mobo this is looking really good. If the turbo wasn't actually working I could see it being close enough to 7700K on single to be an easy replacement considering the multithreading blows it away. and then this wasn't even the highest chip..

File: 1b70c8cd-73d1-4cab-bc44-903e46d0351a[1].png (105KB, 255x345px) Image search: [Google]

105KB, 255x345px

>>58909017
That's going to rekt intels shit

>>58908962
Intel does the same, but to a much less flexible degree

>>58909017
>>58909062
Intel already does that, the turbo boost you see advertised is actually only single core boost

>>58909075
Intel's Turbo Boost 3.0 will pick a single favorite core, and that will hit the peak turbo. Below that they have a set all core turbo, then the base frequency. They don't have a scheme for clocking only the cores are that being loaded based on the workload at hand.

Zen could load 4 cores to take up the entire available TDP, intel has literally nothing similar.

>>58903646
head room.
an i7 may out perform it single core, but when I have something going that's using all 4 cores, I don't want a background application to suddenly demand 25% of the cpu and grind everyfucking thing to a hault.

That headroom is nothing to joke about, we seen it in intels 6 and 8 cores for god damn ever, but no one was paying what intel wanted just for that headroom, now we get a choice.

File: 1440421936208.jpg (58KB, 640x560px) Image search: [Google]

58KB, 640x560px

>>58909017
>>58909070
>>58909075
Don't cores get completely disabled on the highest steps with their dynamic frequency scaling?

or have things changed since sandy bridge?

https://en.wikipedia.org/wiki/Intel_Turbo_Boost#Example

>>58909092
This would be interesting for virtualization, methinks

>>58909092
I still wish AMD had someone invented Reverse Multi-threading

where they could turn an 8c/8t chip into an 8c/4t chip, effectively double single thread performance, would be very important for getting games to run at over 144hz and legacy single threaded stuff

Then when they get to 16c/8t you can beast through any workload

>>58909126
http://www.anandtech.com/show/10337/the-intel-broadwell-e-review-core-i7-6950x-6900k-6850k-and-6800k-tested-up-to-10-cores/2

This page explains the nuances of their latest iteration.

>>58909126
I think with the older versions of turbo introduced with Nehalem/Westmere as the speed bins got higher the CPU would shut down the idling cores.

Not sure if that has changed with the newer chips.

File: VISC-CSMT.jpg (201KB, 980x643px) Image search: [Google]

201KB, 980x643px

>>58909157
Early into the development of Zen people were expecting the CCX design to eventually evolve into just that. A flexible virtualized single core that could massively increase serial throughput. It might be what they're aiming for, but thats years off. Still AMD does have patents on that, a couple companies do, but no one has ever demonstrated a working concept apart from Soft Machines with their VISC architecture. Soft Machines was recently acquired by intel, so take that for what its worth. Could be intel trying to not let AMD get ahead with a new feature, could be them trying to get a trump card for a new ISA, could just be locking away patents.

>>58909157
>Reverse Multi-threading
It's been a while since I've seen someone mention that.
No, it's not a thing. It will never be a thing. If you think it might ever be a thing on CPU's you're naive or retarded.
The whole "use multiple resources to solve a single computation sequence" is what hyberbibelines and branch prediction were invented to do.

>>58904107
i5 bottlenecks the hell out of most games, you just either don't have a monitor good enough to show you, or hardware good enough to push it to the bottleneck.

>>58904116
>>58904069
I want to make an emulation box, something nice for my dad who really only gamed up till super nintendo, quadcore gives it headroom to be a real computer along with more then enough for emulation.

>>58909211
Settle down there, uninformed tough guy.
http://www.anandtech.com/show/10025/examining-soft-machines-architecture-visc-ipc

CSMT is a real concept, it exists in silicon, and it was real enough for intel to acquire the company for whatever reason.
If you don't understand a topic, don't post about it. Simple.

>>58906658
intel has dropped prices in the past, the pentium 300 that came out for 2000$ but once yeilds stop sucking all the dick they could put their mouths around, the price came down.

>>58909252
It's the same sort of thing as
>memristors
>ternary computers
>general purpose quantum computing

It's not going to be a reality any time soon.

>>58903996
>Milking the consumers without any investment is an absolute smart thing to do. Now they got the money
that's not how economy works
money either produce something or become useless, specially in tech world where R&D is what keeps company alive not profit margins
you don't have next generation you may as well hang yourself, AMD survived by a miracle and competent GPU department
If not for ati they'd be dead for real.

>>58908774
>Vega isn't early R&D, its literally about to enter production for a may launch

look at amds released gpus and see how much they grow in just 6 months.

then look at the vega they shoed off and realise its a very early dev board.

>>58909211
>hyberbibelines and branch prediction were invented to do.
isn't it that just in other word? basically after all this DEEP learning research is collected in years to come CPUs become AIs?

>>58909434
that probably was ES from october or so

they wouldn't show off chips everywhere if it wasn't finishing line already
It will launch for consumer as polaris did early june

>>58909444
have you ever decided to do as look more like?

>>58909455
ryzens worst es was 2.8ghz, and we know its minimum base is 3.4 now. thats a difference of around 17% and thats the difference we could be seeing in vega.

I'm not touching oc, because vega is made for enterprise first, and we won't see good binned ones for months if not till next year.

and sure, drivers aren't there yet, but 6 months before launch drivers... we know 6 months on polaris did what, a 7-14% uptick in performance?

I'm not saying vega will be fucking amazing, I'm just saying that this early on there are many other factors that could be hampering it, and because amd isn't talking, they could be showing off vega in the same way they did ryzen, they know damn well where the gpu is going to stack up, and writing is on the wall for shit wrecking performance, if they manage to not bottleneck the fucker, shit will be amazing.

>>58908550
because 240mm fucking suck, get a d15

now if you got a triple rad, that may be worth it over a d15, but by how much?

>>58909557
>d15
ITX bro, also the air cooler takes up a fuck ton of space around the motherboard

https://youtu.be/51hQUYp40nU?t=338

>>58909576
https://www.newegg.com/Product/Product.aspx?Item=N82E16811133275

That is the case im getting, Im so fucking sick of having no room to work in my computer, no ability to cable manage, and no backplate access.

File: IMG_0284.jpg (44KB, 447x440px) Image search: [Google]

44KB, 447x440px

>>58908763
>matte black cylinder shroud
Reminds me of hotwheels.
I actually quite like it honestly, but I'd still replace it with something more capable.

>>58909593
I'd go for a HAF XB Evo over that thing unless you actually need E-ATX support, the XB case isn't quite as big, and has hot swap hard drive bays

>>58908941
You can still manually overclock the good old fashioned way. XFR is only good if you're too stupid or lazy to do it yourself.

>>58909629
4hdds is the sticking point, I have 5 currently and had 6 till I needed a new disc drive so I had to remove one of the redundant drives.

the added benefit, if ryzen does get heavily bottlnecked on air cooling, ill get a dual 360mm rad setup for the cpu alone and make a custom loop.

I fucking hate water cooling, as its a pain in the dick, but ryzen may give me a good enough reason to bother with it again, barring that, a 360mm aio may be sufficient. but I am planning for a d15 to be good enough to push the thing to 4.0+ permanently.

>>58900861
so did I, gonna sell it for some ayyyyyMD

>>58909670
And you don't want a normal tower PC why?

>>58909728
ITX builds are kawaii as fuck.