IT'S HAPPENING: AMD Zen CPU & AM4 Socket Pictured

Maybe if AMD ever produced something that wasn't absolute shit Pajeet.

>>56639497
Zen will use the power of Shiva to btfo Intel, mark my words Shlomo

File: pallas-cat-manul-2__880.jpg (116KB, 880x587px) Image search: [Google]

116KB, 880x587px

Its still using fucking pins.

>Still using pins
>40%+ More Pins Than AM3+
YOU CAN'T MAKE THIS SHIT UP

File: 1468602076611.jpg (40KB, 550x512px) Image search: [Google]

40KB, 550x512px

>>56639400
>PGA SOCKET

>Pins
Fucking why

Just fucking why the pins.
REEEEEEEEE

>>56639710
>40% more pins
>40% more IPC

>>56639400
I wish the public gave a shit enough about all the shit foxconn did to make the crap we use to give a fuck but I guess I need to move

>>56639802
IPC - included pin count

>>56639802
>still inefficient as fuck per clock cycle

>>56639804
Foxconn employees have a lower suicide rate, 50% higher pay, and better living conditions than any surrounding area. Your point?

>>56639804
There's nowhere in this world you can move to where people will give a shit about foxconn workers.

>PGA
Shill harder Pajeet.

File: 2623358-Hand-grenade-in-a-woman-s-hand-with-the-pin-pulled-isolated-over-a-white-background-Stock-Photo[1].jpg (74KB, 1300x870px) Image search: [Google]

74KB, 1300x870px

>>56639400
Meanwhile, at AMD...

>>56639814
Lel

how much will this shit cost

>foxcon
nice

File: gigabyte-foxconn-lga2011v3-socket.jpg (477KB, 1024x683px) Image search: [Google]

477KB, 1024x683px

I'm retarded but Intel socket also use pin...?

>>56639993
>Hmmm, which would I rather break
>$150 motherboard (tops)
>$400 CPU

File: 1449551632624.jpg (11KB, 256x256px) Image search: [Google]

11KB, 256x256px

>>56640004
>bending pins

Are you all paid intel shills? you think the intel sockets aren't pins? Jesus christ who gives a shit what the interface looks like

>>56640004
If you've broken pins on a cpu maybe you're too young to be dealing with computer hardware, Timmy.

>>56639710
Pins are superior. No idea why doesn't Intel use pins.

>>56640203
Why?

>>56639400
The only thing about this that annoys me is the cooling solution mount breaking compat with AM3 and co.

>>56640208
First, pins are durable. There is literally no way to break your socket or CPU. Even if you break a fuckin pin, you can solder it on back with little effort.

If you as much as touch any Intel socket or get a little of thermal compound into it your mobo is fucked for good.

Also pins theoretically have better contact.

>>56640237
So why does it matter if the pins are on the socket as opposed to the CPU? You're making no goddamn sense.

>>56640237
>get a little of thermal compound into it your mobo

How potato are AMD users? For fucks sake, how do you even manage that?

>>56640248
>2016
>not using liquid metal

Stay pleb.

>pins on CPU
Thought it was 2016, not 1996

>>56640264
>putting MSI levels of globs on

Gas yourself.

>>56640297
kek

File: received_990221367741593~2.jpg (39KB, 486x498px) Image search: [Google]

39KB, 486x498px

>>56639647
>Intel mobos are still using pins in 2017

File: 1472936156604.png (447KB, 509x422px) Image search: [Google]

447KB, 509x422px

>>56639647
>>56639710
>>56639738
>>56639773
>>56640004


>>56640237
This

>>56640245
He makes sense you just have a fucking lack of understanding.

File: Nice Tips! with Bailey Jay.webm (3MB, 1280x720px) Image search: [Google]

3MB, 1280x720px

>muh pins on cpu muh pins on cpu

Nothing wrong with that bunch of faggot

>>56639825

I know the "pajeet" thing is satire and all, but it stopped being funny a while ago. Now it's just racist. You can stop already.

File: yff.jpg (58KB, 680x646px) Image search: [Google]

58KB, 680x646px

>>56640527

>>56639400
>bent pin
>can actually fix with PGA
>BGA may as well throw out the mobo
Intel did this on purpose you know

>>56640558
I fixed a bent Intel contact thing on my mobo.
I've fixed many bent AM3 pins but only 1 Intel socket

>>56640385
name of girl?

File: 1313122314777.jpg (20KB, 377x396px) Image search: [Google]

20KB, 377x396px

>>56640627
IT SAYS IT RIGHT IN THE FUCKING WEBM YOU FUCKING RETARD

>>56640627
>girl

I have bad news for you...

File: Bailey Jay.png (33KB, 206x126px) Image search: [Google]

33KB, 206x126px

>>56640627

>>56640627
>Girl
>Bailey Jay
You know "she" has a penor right?

>>56640658
>implying thats a bad news

File: tone.png (527KB, 630x587px) Image search: [Google]

527KB, 630x587px

I use a pent 3 chip with pins to brush my cats fur.

Based AMD will let me do the same with current gen hardware <3

>>56640627
Hi! You must be new here.

>>56639738
>Fucking why
>>56639773
>why the pins.

because AMD aren't complete kikes trying to shift RMA cost burdens to mobo suppliers.

it's a fuckload easier to bend a CPU pin back than one on the motherboard, or replace a CPU than replace a motherboard, which is what really matters in enterprise.

time often matters more than saving a couple hundred bucks on one part vs. another in the worst case.

I don't get people complaining about the pins. Does Intel not have them?

Spoonfeed somebody who doesn't own a new CPU.

>>56640694
>all of us here at 4chan have a mental rolodex of shemales and memes

File: THE_RIDE_NEVER_ENDS_2.gif (186KB, 500x436px) Image search: [Google]

186KB, 500x436px

>>56640385
>I was here when bailey jay was linetrap
>I'm still here
>god help me

>>56640732
Intel puts the pins in the socket since LGA775.

Pins don't matter. Zen will probably flop. AMD has given us no reason to trust that they can make a good CPU. They have to prove themselves.

>>56640732
You have to have pins and contact points. AMD uses the traditional way of having pins on CPU while Intel switched to pins on motherboard a few years ago.

>>56640760
And that's a bad thing.

Will this finally give me a reason to switch from my i5 4690?

>>56640248
Killed a prebuilt that way.
Wanted to change thermal paste, but when I pulled off the cooler, the paste was bone dry and was more like power.
New paste, put the cooler back on, didn't work.
The old paste managed to get on the socket and killed the board.

>>56640004
>fuck up pins on LGA
>you will never fix it, motherboard is fucked forever
>fuck up pins on PGA
>cpu pins are much easier to bend back due to being much farther apart

>>56640789
Except it isn't. Bending/breaking pins on mobo is ten times harder to do than on CPU, not to mention thanks to this change Intel CPUs are bulletproof when it comes to buying used ones.

The only thing I don't like about PGA is that you have to store the CPU's in anti-static foam, or else the pins can get fucked up.
With LGA you can just store the cpu's in a anti-static bag and not worry.

>>56640827
>>56640004
https://www.youtube.com/watch?v=EdDccsbv5hA

>>56640527
Literally the only time I said it.

>>56640842
>Bending/breaking pins on mobo is ten times harder to do than on CPU
No it's not. You just have to touch it. Because intel pins are paper thin.
To bend AMD pins you have to literally hammer them.

Also if you manage to go full retard and bend an AMD pin you can straighten it out with a knife without even looking.
If you bend pins on your mobo you have to take your whole rig apart and buy a new mobo.

how much does intel pay you for your shilling? is that enough to make a life in your poor indian village?

>>56640909
Why would Intel even need shills don't they have 97% of the CPU market by now?

>>56640994
Because they would like to keep it that way, obviosously.

>>56639400
>Looks like socket 478 from the Pentium 4 era
BRAVO AMD

AM4 is 1,331 pin PGA
SP3 is 4094 pin LGA

This is nothing new.

>>56640627
That's not a girl~

>>56640750
I call it my memory and I keep my memory in my brain. I bet learn those multiplication tables was pretty tough.

>>56640798
No. Zen aims to be a good value, not necessarily have more raw power than Intel. Although iirc Zen will offer more cores for cheaper, which would help a lot in DX12 games, if you care about that.

>>56641217
Summit Ridge is HEDT, not mainstream. Its competing against Haswell-E, and the lower tiers of Broadwell-E.
It isn't going to be priced against intel's mainstream parts.

The 4c/8t Raven Ridge APU is the mainstream platform.

>>56641217
Don't really care about DX12 and most games I play favour fewer cores with higher clocks so I guess I'll just wait for Intel to stack 3 or 4 more generations with shitty 5% performance increases.

I just told my parents to invest in AMD because they're poised to make a comeback (although I did say I'm not confident enough to invest money I'm not comfortable losing). Am I right, or is AMD so fucked that Zen being successful wouldn't increase their stock price by that much?

>>56641287
>giving stock advice to anyone
>when you're asking /g/ for advice

AMD will experience a moderate rally when they start drawing revenue from Zen based parts, but you don't tell old people to invest in a tech rollercoaster unless they watch it often enough to sell at the peak.

>>56640224
if you bought a cooler from noctua they're giving new mounting brackets away for free

>>56642113
NZXT Kraken so yeah...

>>56640203
CRUNCH

>>56639400
>wccftech

>>56642230
http://www.hwsw.hu/hirek/56123/amd-socket-am4-foglalat-platform-bristol-ridge-summit-ridge-raven-ridge-promontory-chipster-alaplap.html

>>56642353
>fucking hungaria
nice

File: do it.jpg (147KB, 1500x1500px) Image search: [Google]

147KB, 1500x1500px

>>56640385
>bangs

>>56639400
>AM4
>fucking 754 clone
KEKSIMUS MAXIMALUS

>2016
>not BGA
>will never experience desoldering a cpu from a mainboard and replacing it
>SMD rework station rarely gets used
>i have no friends
>no job
>no girlfriend

:+(

>>56640854
I've stored an old shitty laptop pentium PGA in a random box filled with shitty old DDR1 and other assorted shit and it couldn't care less, 1 pin is bent so slightly it would fix itself in the socket.

Meanwhile, with LGA, breathe on that fucking shit too hard and enjoy getting a new motherboard.

>>56639400
>ITT: le fucking pins
What has me worried more is that the mount is made of fucking plastic, how is this allowed? How can anyone think this is better than a metal bracket?

File: hardware1.jpg (263KB, 1024x768px) Image search: [Google]

263KB, 1024x768px

>>56639400
>1331 pins
why not just make it 1337zorz??

>>56643983
You are aware every other CPU mount/socket is plastic, right?

>>56640237
Those are absolutely irrelevant.

>>56641246
Performance and Performance/Watt kinda go hand in hand.
Somewhat worried that they won't make it on the Laptop CPU/APU side of things, and then they'll be left only competing in a small section of the enthusiast market.

File: 1470548083133.jpg (17KB, 480x592px) Image search: [Google]

17KB, 480x592px

>>56643983
What are they supposed to me made of, idiot? Fucking metal? What the fuck is your point?

>this is neo-/g/
>inb4 pretending

>>56641246
Everything uses AM4, so It's the same set of motherboards.

>>56644209
>Somewhat worried that they won't make it on the Laptop CPU/APU side of things

Mobile Raven Ridge scales from 35w to 4w.

>>56643983
are you too underage to remember the times when most CPU sockets were like that?

>physical connections
AMD is AMDoomed

>>56639647
When will they finally use bluetooth and integrate a battery so I don't have to mess around with pins.

>>56640203
Intel uses pins. The pins are on the mainboard.

just wait for zen 5 guys

>>56640554
Did he shit inside the toilet paper roll? It's brown inside.

File: 1457565164589.jpg (627KB, 1500x1101px) Image search: [Google]

627KB, 1500x1101px

>>56639647
>intel uses pins on mobo
>amd uses pins on cpu
>somehow one solution is oh-so-fucking-superior

i bet it's because when you feel like using your cpu as shuriken, intel has better chance of survival

>>56639400
>no thunderbolt 3 integrated in the chipset
AMD is doomed in the laptop segment

>>56641021
why would they need shills for that?
amd is doing a great job at securing intels marketshare

>>56644718
Intel users are gaymen.

>>56639400
>shitconn

>>56640862
holy shit

ive built computers smashed drunk and never bent a pin. knock on wood.

>>56641287
imo it's too late because most holders already calculated zen's success in their decision to buy so it will mostly fall really hard are rise really soft

>>56644008
hummmmm

>>56640862
ha, I had to bend about 30 - 40 pins back to straight after fucking up an AMD processor, I used a bamboo skewer and thankfully none broke. CPU has been running pretty much constantly for 2 years now without issues

>>56639400
Single-core benchmarks when?

>>56644609
Cardboard

>like intel
>see cheap old cpu
>motherboard are overpriced as fuck because ruining one pin destroys everything

>>56639400
Seriously? Pins on the cpu?

Fuck, that is some grade A stupidity.

>>56647091
>t. intel shill

>>56647091
do tell, why is it stupid?

>>56647170
because it was used before socket 775 times in everything and then intel abandoned it

if intel abandoned it, then it had to be stupid

>>56647091
>>56647170
>>56647198
Have any of you read the thread?
Intel moved the pins to the motherboard because if you do fuck up the pins it's cheaper to replace the board than the cpu in a lot of cases. Pins on the cpu are easier to fix though. They basically cancel eachother out. It doesn't matter unless you're a clumsy motherfucker.
I personally prefer lga because pga chips with old thermal paste tend to get ripped out when you remove the heatsink.

>>56647269
Have even installed both types of cpus? The old intel pined ones were fucking awful because the design needs the pins to be so long they become weak. Moving them to the motherboard allowed them to be shorter and stronger and far harder to damage.

>>56640004
Intel motherboard for 150 dollars? Are you shitting me? Maybe if you bought the most bargain bin mobo. The average mid level intel mobo is 250 bucks.

Anyway it's all got a warranty.

>Outdated, obsolete PGA instead of LGA

AYYMD HOUSEFIRES

AYYMD HOUSEFIRES EVERYWHERE

>>56647269
>it's cheaper to replace the board than the cpu in a lot of cases

It's way less convienient. It's maybe going to save you 50 bucks, but you will have to take everything apart, take apart your cable management and shit.
Any sane person would rather replace the CPU.

>CPU is $300
>Motherboard is $100 because it's AYYMDPOORFAGs tier
>better to replace the CPU

AYYMDPOORFAGS LOGIC, EVERYONE

>>56647309
>Pins on the cpu are easier to fix though

>>56647313
Bargain motherboards are under $100. Average boards are around $150 and there is rarely a reason to spend more. $250 boards are either entry level server or gamer trash.

>>56647356
>$250 boards are either entry level server or gamer trash.
$250+ board is a must with any K series CPU.

I'm rolling on a $350 board right now.
With Intel your motherboard has to cost MORE than your CPU to get some decent OC.

>>56647390
>he fell for the OC meme

>>56647269
it seems like you don't understand: you're in thinly veiled amd vs nvidia thread

you either like amd and shit about superiority of their sockets, or you like intel and you shit about superiority of their sockets. you can't just hold the middle ground. pin logic doesn't matter and it's obvious, everyone knows that

>>56639400
>AMD Zen is coming, we swear!
Zen has been "around the corner in Q(1-4) 201X!" for something like 2 years. I refuse to believe that it will actually be released until I'm able to physically buy one.
That being said, if it's released and actually is a meaningful upgrade over my 8320 (read: better than the Z-series Icelake i7's or whatever generation intel is on by release), I'd strongly consider making the jump.

File: A1oyayd3ZLL._SL1500_.jpg (562KB, 1500x1491px) Image search: [Google]

562KB, 1500x1491px

>>56647390
My z-97 board was $139.99.
>4 ram slots
>great vrm cooling
>3 pci-e 16x slots
>sata express ports
>m.2 slot
Maybe you suck at shopping.

>/v/ complaining about PGA
ayy

>>56647390
>Paying $300 premium for an OC board
>Paying $80 premium for a good cooler
>Paying $50 premium for a K CPU

Why don't you instead invest in something useful like a large SSD?

>>56639400
https://www.youtube.com/watch?v=B9Aoaxwouu0

File: 961221_1__62055-1[1].jpg (114KB, 800x800px) Image search: [Google]

114KB, 800x800px

>>56647390
125 € when I bought it last april.

File: Picard Pointing.png (667KB, 1402x804px) Image search: [Google]

667KB, 1402x804px

Lots of very bored tards on /g/ today. Honestly who gives a fuck if it's LGA or PGA, no-one on this forum is going to break pins unless they are underage.

>>56647541
Yes, adults have the money to buy prebuilts.

File: urnot.jpg (9KB, 480x360px) Image search: [Google]

9KB, 480x360px

>>56647440
>3 pci-e 16x slots
nice try but you've been PCIe cucked those are clearly x8 electrical slots. How sad he thinks he actually has 3x PCIe x16 but instead he's just a loser.

>>56647895
Close, but not quite. I can run one at full x16, or run the top two at x8 and x8. The bottom one is run off the chipset controller. Useless for a gpu, but fine for other things. You can't get multiple 16 lanes without an extreme edition processor and chipset, and the intel tax is insane.

>>56647771
gullible fools have the money to buy prebuilts unless its actually a good deal unlike 99.9% of prebuilts

>>56647925
>I can run one at full x16, or run the top two at x8 and x8
Even on a dirtcheap AMD FX CPU you can run two graphics card at x16 (PCIe 2.0).

How can Intel get away with this shit?

>>56648035
Because with PCIe 3.0, the difference between 8x and 16x is fuck all.

>>56639400
That looks like the same old ancient shit to me.

>>56640004
>being retarded enough to bend cpu pins

>>56640761

From what I understand with 14nm Samsung. The cpu will do just fine at 3.0 GHz and will compete with Intel in that GHz level.

BUT Intel cpus can clock to 4.8ghz.and zen probably won't pass 3.4ghz.

Same story as with Polaris. Can't clock because Samsung/glofo

>>56648835
GCN is not a high clocking arch, it never would be no matter what process its fabbed on. 14nm LPP provides an increase in fmax over 28nm HPP, the fact that 100% of all Polaris 10 cards has to hit at least 1266mhz is incredibly impressive for the arch.

>>56648946
>GCN is not a high clocking arch

the GCN architecture doesnt have a single thing to do with clock speed.

>>56640385
faggot

>>56647313
bargain motherboards are well under $75. if you're going to shill don't do it with misinformation.

>>56644008
I thought it was 1497 pins

File: okay scooter..png (250KB, 1205x576px) Image search: [Google]

250KB, 1205x576px

>>56647390
>$250+ board is a must with any K series CPU.

yeah okay.

>>56649176
Yes, it does.
Tech illiterate retard.

It has an 8 stage pipeline.

>>56640753
feels bad man

linetrap was so cute, everyone immediately fell in love when he flashed the camera

if only people would lurkmore.com

File: jay.webm (810KB, 480x480px) Image search: [Google]

810KB, 480x480px

>>56649834
>>56640753
''Bailey Jay'' is cuter, ''Linetrap'' was for real homos: no ass/tits.

Why does /g/ fall for AMD's lies so often? I know people desperately want Intel to have some rela competition but holy shit how many more times are you going to get burned?

I personally prefer having the pins on the motherboard. You don't really handle the motherboard socket much with your hands, but you do handle the CPU. I've bent the pins on my FX-8350, but I was able to bend them back.

It's just a preference though. Now that I know how easy it is to bend the pins I'm just extra careful with the CPU. If the price and performance on the new Zen chips are a good improvement, I might pick one up.

>>56649721

hi again Newfriend.
still nothing to do with clock speeds tho.

if AMD had gone for TSMC and had a different production that 14nm LP/P then they would have had a completly different and probably higher clock speed.

sadly because they are locked too glofo, they are stuck with shit speeds. Thats one of the main reasons they want to get out of the contract.

>>56643957
I had an ancient pentium 1 lying around in a box with other stuff, and while one row of pins were bent, it was easy to straighten them out. straightening pins on a mobo ranges from hard to impossible. I managed to fix a 1366 board with a bent pin, but it required a sharp tool and a very steady hand.

File: 1472657564810.jpg (13KB, 600x589px) Image search: [Google]

13KB, 600x589px

>everyone jumping ship from x86
>Intel already pushed it to the limit with Haswell and declares Moores Law dead
>AMD goes out to the world and says, "look, we can compete in a fucking dying market!"

AMD is without a doubt the most pathetic, yet hilarious company in tech. Its like the slow kid who finally figures out how to do Algebra when everyone's already learning Calculus.

>>56650697
>everyone jumping ship from x86
There's not really any alternatives as of now, except maybe POWER, but nothing has really gained ground yet.

>>56650697
>one company has no competition so they decide to put some bullshit in order to justify their 6xxxs being just as powerful as previous generations
>hurr amd is dum why are they even trying

also
>implying x86_64 is going anywhere

>>56647771
Adults don't play games tho

>>56650731
AMD doesnt innovate, their only goal is to be 2nd place. And they finally catch up when the race has been over for 2 years.

Im just wondering why anyone finds any of this impressive or worth supporting. Its pretty clear Intel hit a brick wall in performance and "da Joos" dont have anything to do with it.

>>56650890
>amd doesnt innovate
>pentium 4 times don't count
>dual core don't count
>x64 extensions don't count

yeah, cool

>>56650890
>AMD doesnt innovate
Then why were they ahead of Intel in the early-mid 2000s? Bulldozer was also pretty innovative, although it turned out to be poop. Their IGPs are also way better than Intel's still.

>>56650908
he was probably in diapers at the time

>>56650908
>10 years ago is the same as today
Other companies look for new things and innovate, AMD is trying to show off that it can be 2nd place in a dying market, Thats all Zen is. At most itll be a cheaper, shittier i7 6900k.

>>56650979
>he thinks x86 is going away anytime soon
oh wait, you're actually serious. let me laugh even harder

>>56640203
because the intel engineers who designed the sockets were incompetent fucks that allowed the destruction of many CPUs.

>>56640627
>girl
ayyyyyyyy

>>56651034
uses for x86
>Chad can build his sick GAYYYMER PC
>Rajeet can get a desktop using his AMD shill money to shill for AMD more
>Chink Incorporated can keep using their housefire tier Windows servers to host their malware websites

Just about everything else relevant is being phased out by ARM

>>56640671

Found the feminine penis sucker.

>>56651173
>Windows servers
Seriously, who the hell uses this shit? For what purpose?!

>>56651222
Thats apparently what AMD is going for with Zen
And as I said, Chinks and Pajeets

>>56651173
you are deluded if you think that windows, gaymes and commercial software, which are all multi-billion dollar markets, are going to be moving away from x86 soon

>>56651277
>And as I said, Chinks and Pajeets
But why Windows?

>>56650979
>>10 years ago is the same as today
tell us, faglord, how does intel "innovate" these days?
most, if not all AMD CPUs support amd-v, iommu, opencl, sse4.2, avx, ...
in terms of "innovation", adding SIMD extensions is the same shit as usual: add more registers, widen them, add some instructions

>>56640827
I never fuck up pins on either LGA or PGA, but PGA retention mechanism is shit compared to LGA.

NEW TECHNOLOGY OUT SOON!
AMD RELEASING NEW CPU/SOCKET
LETS DISCUSS!

Replies from /g/:

HURRR PINS
LOL PINS
PINS
2016 PINS
WHY PINSHURRR PINS
LOL PINS
PINS
2016 PINS
WHY PINSHURRR PINS
LOL PINS
PINS
2016 PINS
WHY PINSHURRR PINS
LOL PINS
PINS
2016 PINS
WHY PINSHURRR PINS
LOL PINS
PINS
2016 PINS
WHY PINSHURRR PINS
LOL PINS
PINS
2016 PINS
WHY PINSHURRR PINS
LOL PINS
PINS
2016 PINS
WHY PINSHURRR PINS
LOL PINS
PINS
2016 PINS
WHY PINS

This is all of /g/ in one thread

>>56651459
>AMD
who cares?

>>56651467
Because a Monopoly in any industry is very very bad for the consumer, the fact you don't understand this is sad

>>56640721
>which is what really matters in enterprise.

AMD has been using LGA for enterprise products for ~10 years now m8.

>>56651459
PIN DUDE LMAO

Meanwhile...
>>56651791

>>56650192
You know nothing and spew straight bullshit.

There was absutely no meaningful difference between the performance, heat waste, and battery life of Samsung or tsmc produced apple A9 SoCs.

What youre suggesting is idiotic, contradicts evidence, and shows that you in fact know nothing about IC design.
Yes, the short GCN pipeline, hardware scheduler, and the decision to couple all parts on their GPU to a single clock driver are the reasons why AMD parts dont clock like Nvidia parts.
It has nothing to do with the fab process and it's also why they draw more power than Nvidia.

>>56651459
you misspelled /v/

File: 1456013208626.png (81KB, 624x628px) Image search: [Google]

81KB, 624x628px

>>56640627
>>56640650
>>56640658
>>56640665
>>56640667
>>56640694
>>56641158
>>56651149

Best part about Zen is that I probably wont need a new Mobo when Zen+ comes out

AMD arent socket jews like Intel

>>56639814
I'm dying, I don't know why that was so funny but it was.

>>56639400
>not LGA

This is a damn joke.

Intel pins are Ultra Durable.

>>56640237
>Even if you break a fuckin pin, you can solder it on back with little effort.

It that pin density, good fucking luck soldering back a broken pin. Especially if it gets lost - then you'll need a substitute that is the exact same size (length and girth).

Pins on the motherboard are objectively a better choice because they are much less likely to get damaged from accidents.

>If you as much as touch any Intel socket or get a little of thermal compound into it your mobo is fucked for good.

If you have thermal compound getting into the socket then you are using way, way, way too much. That thing is meant to fill out the uneven parts between the cooler and the heat spreader (because it transfers heat better than air), not act as an auxiliary layer between the two. Given the force with which the cooler sits on the CPU, which will flatten out the compound as much as possible, even a small pea in the middle can be too much.

>>56653753
>>>/v/

>>56640677
fuckin onion eater

>>56647269
I am reminded of the last time I removed an old single core amd cpu heatsink to replace the thermal paste. It ripped the cpu out with it. Didn't damage any pins though, so I was lucky in that regard.

>>56649721
do pipeline lengths even matter with SIMT?

the entire point is that you just juggle a fuckton of threads in parallel to hide various form of stalls, and merge threads that rejoin the same instruction flows after branches, etc.

>>56639647
>>56639728
>>56639738
>>56639773
>>56639825
>>56640036
>>56640286
So many retards without much experience.

>>56640721
>>56640827
Listen to these guys, kiddos. There are good methods to fix a pin that gets bend. Plus you never bend a pin when you know how to remove CPU heatsinks (release mechanism but hold in place, then turn the heatsink a bit until it gives in to "tear it off").

It is nigh impossible to fix pins on LGA mobo if you screw up. Kill.

>>56644019
But how does everything short itself if it is not METAL like proper GAMER HARDWARE or iPhone phones?

>>56647472
That board is dope.

File: dd446ff3771bb10b47e37c5ca2cde9e8.jpg (156KB, 720x481px) Image search: [Google]

156KB, 720x481px

How powerful is this new Zen CPU?

>>56649995
i want baily to fuck me in the ass

>>56655408
AMD claims 6900k, reality will probabl be a shittier i5

>>56655408
even if AMD are not lying and it gets +40% over FX series

thats just less than haswell performance so it better be fucking cheap

>>56655769
Haswell performance for normie applications like gaming on emulators that still cant use more than 1 core

Broadwell E performance for video rendering or anything that actually requires heavy processor performance

>>56640354
>soldered onto mobo
>laptops only
Looks like you sure corrected the record!

>>56649263
>AsRock

>>56655769
>thats just less than haswell performance so it better be fucking cheap

It's Haswell performance but with 8+ cores. So basically the equivalent of Intels $1000 extreme edition CPUs.

>>56640750
Everyone used to know about linetrap.

>>56656358
yeah I think people forget AMD means 40% ipc while still having moar cores

>>56647449
/v/ read the wiki page on lga now they're computer masters

File: 1458588717092.png (30KB, 493x402px) Image search: [Google]

30KB, 493x402px

>>56640627

File: 1420349002507.webm (3MB, 1280x720px) Image search: [Google]

3MB, 1280x720px

>>56640248

>>56644718
>>somehow one solution is oh-so-fucking-superior

Yes, one is embedded onto a plane that is surrounded by elevated parts, it takes a concerted effort to fuck pins on the motherboard.

There's a reason the company with actual market share moved away from pins on CPU, and the idiots who can't sell a product to save their lives haven't.

>>56639400

AMD makes such shit products that compete with years old Intel CPUs to the point that even the fucking sockets for AMD CPUs look a decade out of date.

File: clean.jpg (42KB, 640x640px) Image search: [Google]

42KB, 640x640px

>>56653003
>he took the meta bait

>>56658683
>He took the new meta bate

>>56658702
>he took the meta meta meta bait

>>56644735
thunderbolt belongs to Intel
Thats like saying why doesnt AMD support Gameworks

File: image.jpg (37KB, 478x478px) Image search: [Google]

37KB, 478x478px

>>56651527
People support AMD only as a token competitor
Fuckin kek

You should be saying, "I like them because they make good products", but of course this is impossible because everything from AMD is dogshit. So you say, "Im glad theyre here as a competitor cuz monopolies r bad!".

Its like saying your child's awful anime drawing is a masterpiece and hanging it on the fridge. You dont want to hurt their feelings, just as you dont want AMD to go under like the cancer to the tech industry they are.

Not smoking 30 cigs and poppin a xanax before building a computer to ensure you do not fuck up the pins.

Get out you fucking casuals

>Kindergarten teacher gives class an assignment to color a drawing of a caterpillar
>One kid does it at lightspeed all while staying inside the lines
>One kid shits himself and smears it all over his paper
>Teacher gives the shitter a reward because at least he is providing competition against the child who's extremely good at coloring.
>Doesnt send child to special ed to prevent a "monopoly" of children who are good at coloring

AMD.

>>56639566
>btfo a now 4-year-old CPU
>maybe by May 2017

Also keep in mind:
>first come SoC
>then come server
>then comes OEM release
>THEN...then...assuming they don't go bankrupt again....a year later comes packaged desktop CPU's
Don't expect to be able to buy a consumer tier Zen CPU without buying a prebuilt until 2019.

By which point Intel will be 2 generations ahead...again.

>>56659372
>look at me make shit up: the post
Summit Ridge and Opterons come out before anything else.

>building new pc to replace 5 yr old setup
>want to wait for zen/kabylake
>we dun support older than win10

that means intel gets my money with skylake. already bought my motherboard. too late turn back now. if only AMD would support 7. i could have waited. in fact i'm waiting vega/volta before buying a GPU. just slowly collecting parts on sale atm.

>>56659383
And yet here we are looking at a functioning SoC...

And not a Summit Ridge.

With no mention of when Summit Ridge will launch since December of last year (when they promised "late 2016"...which we're in).

>>56659424
>I still have literally no idea what I'm talking about

Raven Ridge APUs aren't coming out until the second half of 2017.
Summit Ridge and Opterons launch first.
You're just pulling things out of your ass.

>>56659014
bad bait.

so apparently microsoft will only support Zen on spyware 10. does this mean that the pajeets have won and windows 7 won't run properly on it?

>>56658652
If you are too dumb to store CPU in a box so pins don't bend, it's a solution.
Too bad the market is mostly made of dumb people (as is the world).

>>56659422
You need to read more. Everything is supported by everything, just not fully (i.e. new meme features).

>>56659422
>>56659798
As >>56660168 said, W7 can still work in skylake, thing is you need the USB drivers because they changed the host controller from EHCI to xHCI and microshit never made a new ISO with those in, think of it as what happened with XP and the sata drivers that at first needed to be installed with a damn floppy disk...
The other changes are the use of newer power states but again see it like with XP that to open solitaire it made the cpu use all cores at max frequency even if it was for 1ms, it would just mean it will manage it like shit between states and have more overall heat and power consumption (still this would be a real issue with notebooks). And Directmeme12, that at least for now it means nothing and by the time it make an OS exclusion (like Dx10/11 games without 9.0c engine/codepath were to XP) either your hardware would be old or who knows, maybe all pajeets died and microshit decided to ditch 10 because reasons.

And same should apply in Zen, unless the assholes never release motherboard drivers for
7/8.1 and you get stuck without USB and god know what else as you would depend exclusively on the windows default drivers compatibility if there's a new hardware like a new version of the lan controller, audio controller and such that don't work by default.
Still, seeing they aim their product to the cheap marketshare they would probably deliver said drivers at least for the first chipsets, and maybe as a market strategy to eat up the "upgrade/new/first build but don't want the w10 shit" customers.

>>56655855
>no source

All we know so far is that it slightly beats Broadwell-E clock-for-clock in a multithreaded test on 6c/12t. Considering it's likely AMD's SMT implementation is worse than Intel's (who've been honing it for 10+ years), this could actually mean Zen's single-thread IPC is around Skylake level.

We don't really know at this point. Skylake's IPC is only 5% (if that) better than Broadwell-E on average so either way Zen should be competitive. Let's just hope there isn't a huge spanner in the works like "it won't clock past 3.5 GHz".

>>56659424
Summit Ridge is Q1 2017 m8, AMD's most recent slides said this.

I don't think AMD are scared of Kaby "marginal gains" Lake.

>>56639400
>pins not on socket

what is this, the year 1993?

>>56655769
FX series is Piledriver. AMD claims it's 40% above Excavator.
That one leaked gaming benchmark shows Zen to have at least 80% higher IPC than Piledriver.

>>56660515
>All we know so far is that it slightly beats Broadwell-E clock-for-clock in a multithreaded test on 6c/12t.
No it was 8 cores and 16 threads. Broadwell-E i7 6900k vs a Summit Ridge ES chip, both fixed at 3ghz.
At this test showed that in the specific case of Blender, AMD's new arch does well.

AMD have claimed that Zen has the same power per clock as Excavator, and worst case scenario that means 3.2ghz is 10w per core. With 8 cores and 15w for uncore that gives you a 95w part clocked at 3.2ghz. Thats the worst case scenario for the top binned Summit Ridge at stock.
If they were comparing it to Excavator in Bristol Ridge with its tweaked AVFS then clocks would be higher still.
If the uncore draws less power that also provides more power margin for higher frequency.
How the chips handle turbo speeds is still entirely up in the air as well. They could have 4 threads on one CCX clock up to 4ghz.

>>56639400
>dat cheap plastic socket

once this hits 90c, like opening a browser, shit gonna melt

>hole in middle of socket

Welcome to 2003

File: L_Socket_FM2-top.jpg (96KB, 596x661px) Image search: [Google]

96KB, 596x661px

>>56660664
FM1, FM2, FM2+, and socket AM1 all have the same "hole"
Not that any of you tech illiterate children understand this.

>>56660695
Meaning AMD has not xhanged their tech at all since 2003

File: 2011v3.jpg (1MB, 2969x2705px) Image search: [Google]

1MB, 2969x2705px

>>56660713
Yep. You're a tech illiterate child.

>>56660586
It's the same plastic that intel's sockets are made from.

>>56660730
Thats not a hole though that is so those boxes can communitcate with the cou wirelessly

>>56660713
Intel has never NOT had the hole in the middle of the socket on ANY of their non-mobile sockets, PGA or LGA.

And intel uses pins still for their mobile sockets.

>>56660736
I bet you actually believe that too.

>>56660746
Well what else would the little boxes be for?

>>56640827
>fixing things
>not just buying replacements

are you poor?

>>56660765
they hold spare socket pins in case you break some.

>>56660765
For exposing tech illiterate retards who can't identify a surface mount capacitor. Like you.

Well we know that Intel are at least half-worried about Zen because they are finally planning to bring out a hex-core mainstream CPU based on Kaby Lake called Coffee Lake. It's also come about because 10 nm is harder than they thought and Cannonlake will ship later than expected.

>>56660765
they're surface mount capacitors, dumbass.

>>56660780
>capacitor

More 2003 technology all CPUs suck

>>56660783
>inb4 Intel shills saying competition is bad

>>56660791
>2003

that's like 60s technology.

>>56660783
Watch CannonLake roll out like Broadwell CoreM parts did. Horrid yields, SKUs a total mess of ridiculous segmentation since they couldn't get enough chips to bin with the same clocks.
If intel really does release Coffee Lake in 2018 as a 14nm part then they'll be on 10nm until 2021~

It'll be interesting to see intel's 10nm bulk Trigate vs IBM's 7nm SOI FinFET.

>>56660814
The 60s are a myth perpetuated bu the government so they can justify banning a natural plant

>>56660824
it all makes sense now!

>>56660824
lmao

>>56639400
>INTEL BTFO SOON
does anyone still believe this

>>56655364
mobos are cheaper than cpus though, so it's good that it's on the mobo

>>56639497
fpbp

>>56654374
Fathers PC is overheat. I come over with tools
>Unplug fan and try gentle twist and pull on hsf
>Thermal paste is like glue, CPU comes straight up with hsf
Haha oops!
>Gently lever CPU off heatsink with screwdriver
>Paste releases with a pop, CPU soars then bounces across table
Whoops! This almost never happens..
>Clean up, boots up fine, 30° cooler, thanks AMD
Professional at work

>>56639400
>A new socket!

OMFG, A NEW SOCKET THATS AWESOME, THANK GOODNESS THEY ADDED A SPACE IN THE MIDDLE, NO WAY WILL THEY EVER RELEASE AN EVEN NEWER ONE AND FILL THAT SPACE AND THEN WE WILL HAVE ANOTHER NEW SOCKET. GOD I LOVE NEW SOCKETS!

WE NEED TO DO THIS WITH EVERYTHING, EVERY YEAR CHANGE YOUR HOUSES SOCKETS SO YOU HAVE TO REPLACE THEM ALL WHEN YOU BUY A NEW TOASTER.

>>56661334
Can I subscribe to your conspiracy blog?

Cool. Now show me the prices.

>>56663399
All AMD has officially stated is that ASP would be higher than previous FX parts.
That puts it at $200+. Personally I'd bet on the top end Summit Ridge selling for $500, and $250 for the top end Raven Ridge APU.

>>56663453
compared to the $1000 chip its competing against it better be gud

>>56640189
This, I've had crap thermal paste turn to glue, and rip AMD cpus out of their sockets with the retention arm down, and I have never bent a pin.

What are you guys doing to bend pins, dropping them for third story buildings?

>>56640909
how about just dont stick your fucking fingers into the damn socket.

>>56639400
Why do AMD CPUs get so hot?
Intel CPUs turn off when they overheat. AMD CPUs catch on fire.

https://www.youtube.com/watch?v=hxSqCdT7xPY

>>56663682
Those are some ancient chips.

File: am4_mounting-840x760.jpg (90KB, 840x760px) Image search: [Google]

90KB, 840x760px

>>56663709
I saw the article. AMD still needs a Huuuuge radiator and fan.

Just make fucking ARM emulate x86.

>>56663760
Wraith cooler will do the job if you don't plan to overclock.

>>56664247
The Wraith cooler actually comes within spitting distance of the Hyper 212 in terms of performance if the Anand article is any accurate.
Downside is that due to its size, even with a 92mm fan its a bit of a noisy bastard.
For a stock cooler though thats really good.
http://www.anandtech.com/show/10500/stock-cooler-roundup-intel-amd-vs-evo-212

>>56663760
theres no fan on that radiator though
if you have no fan of course you need a huge radiator

I did notice they had to have extra fans on the motherboard during the Zen demo, though

Im hoping I can keep my AM4 mobo when Zen+ is out

save like 200-300 dollars right there

>>56664307
I mean I can deal with the noise. I'm used to loud fans and they don't really bother me. However Intel stock coolers are becoming more and more garbage. Thankfully they don't include them with K parts anymore because they are just trash. Good for the average Joe but not for many people.

>>56664417
>>56664307
Also let's hope that AMD does something similar with Zen like they did with the FX 9590 and ship it with an AIO. That was freaking cool.

>>56644773
That's just what you shills want us to think.

File: 1471396486391.png (182KB, 578x691px) Image search: [Google]

182KB, 578x691px

>>56664446
AMD is the only company to give us liquid cooling for free and its ONLY because their products are fucking housefires

>>56664417
Seriously, stock Intel cooler on my 4790k couldn't keep it below 100°C at stock clocks when doing anything intensive.

>>56664417
Intel stock coolers have been trash since Ivy Bridge I think when the quality really started to nosedive even as thermal density of the chip went up. The only decent stock coolers intel has made recently were for the LGA 2011 chips, and even then thats another $30 and it still gets shit on by a good tower cooler.

>>56664446
Depends on what they do with the top 8c Zen SKUs in terms of power and clocks. Keep in mind that the AIO coolers for the 9590 was required due to the whopping 220W TDP. the 8C SKUs are expected to be sub-125W.

Now, if AMD got really ballsy and decided to release a modified Opteron variant as a super-high end (12C most likely using chips that failed the cut for Optys) unlocked option, then I could see the point of a AIO cooler. I doubt it will happen though, that's a really niche market.

Is Bristol Ridge being sold to OEMs only, or just a month or two of exclusivity?