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

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

/g/ - Technology


View post   

[ Toggle deleted replies ]
File: 265 KB, 960x540, AM4-Chipset-4.jpg [View same] [iqdb] [saucenao] [google] [report]
63669951 No.63669951 [Reply] [Original] [archived.moe] [rbt]

https://wccftech.com/amd-james-prior-interview-vega-11-ryzen-2-am4-vega-supply/

>Rest assured, you will be totally able to plug in the next Zen+ / Zen 2 product in your existing AM4 motherboard.

Well, Intel is dead.

>> No.63670105

>Buy AM4 board, put new Ryzen in board until 2020
>Ryzen+, Ryzen 2 and possibly Ryzen 3

That alone makes this SHIT all over Intel.

>> No.63670166

>>63669951
>will be totally able to plug in the next Zen+ / Zen 2 product in your existing AM4 motherboard.

Man, it's not like we knew this a month BEFORE Ryzen launched in March!

>> No.63670191

>>63670166
there were rumors this changed a month or so ago, some intel guy was saying amd fucked up and they're silicon wasn't compatible

>> No.63671605

yfw, no desktop vega11, just a shitty apu

>> No.63671624

>>63671605
I want my vega11 with 2560sp @ 1800mhz

>> No.63671669
File: 29 KB, 800x473, 1506493837627.png [View same] [iqdb] [saucenao] [google] [report]
63671669

>>63669951
who cares about that shit? new threadripper when???

>> No.63671683

>>63671669
2nd half of 2018

>> No.63671693

>>63671683
fuck I can't wait that long

>> No.63671699

>wccftech
nice meme

>> No.63671700

>>63671693
buy i9 when

>> No.63671710

>>63671700
but that's shit compared to current TR

>> No.63671717

>still no 7W mobile APU for cheap fanless systems
AMD pls

>> No.63671726

>>63671717
im sure you could get away with passively cooling one of the existing APUs if you try to undervolt them

>> No.63671770

When is AMD going to release a decent desktop APU?

>> No.63671771

What percentage of people upgrade their CPU every year? The 1%?

I have an i5 4670K and I still have no reason to upgrade. If you're shelling out $500+ every year for max 5% performance gain, you're simply retarded.

On top of that, the people who can afford a new CPU every year, can also afford to buy a new $60 mobo with that.

"Upgrade paths" are a fucking meme.
Even for me, the Haswell i7 is still so good that it's more than $200 right now, so it's not a cheap upgrade even 4 years later.

Can people stop using this meme argument now?

>> No.63671775

>>63671726
I'm referring to banded kestrel and river hawk. RH is supposed to be on 7nm and have a 4-15 TDP. It could compete with the atom lineup lol

>> No.63671794

>>63671605
>trusting AMD

>> No.63671797
File: 21 KB, 240x498, 1498383784076.jpg [View same] [iqdb] [saucenao] [google] [report]
63671797

what is the chance that the next ThreadRipper will use the same exact socket and be compatible with current TR setups?

>> No.63671822

>>63671797
146%.
SP3 stays the same for everything Zen.

>> No.63671824

In other words Ryzen 2 will be outdated trash without PCIe 4.0 and DDR5.

>> No.63671840

>>63670191
It was just a retard implying that new chipset means no backwards compatibility.

>> No.63671844
File: 109 KB, 2250x2324, 1512249231547.png [View same] [iqdb] [saucenao] [google] [report]
63671844

>>63671824
Not a single intel platform will have enough PCIe lanes to be useful.

>> No.63671946
File: 28 KB, 386x495, JIM.jpg [View same] [iqdb] [saucenao] [google] [report]
63671946

SAY HIS NAME

>> No.63671976

>>63669951

AMD will always be a second-tier brand

and only poorfags care around recycling their mobo when they change CPU. you do that every 5 years anyway so you would want the new mobo for the new features.

>> No.63672013

>>63671946
STALLMAN!

>> No.63672045

>>63671946
CERTIFIED

>> No.63672066

>>63671624
what do you think the 11 means? For once in the long history of graphics accelerators it's a logical naming scheme.

>> No.63672089

how fast will those chips go though?

>> No.63672135

Whenever I get bored of my current PC and plan to renew it, I'm going to start for a new case all the way up to a new GPU/CPU, I won't be reusing 4-5 year old parts. Replacing only some parts it's like putting some makeup on a woman and claim that she's a new woman. No, you don't put makeup on your wife if you get tired of her, you simply dump her and go for another one, a sexbot in this case because women nowadays are all used goods.

>> No.63672189

>>63669951
The closest Intel got to death was during the 32/64 bit leap, when they had the bright idea of abandoning retroactive compatibility. But AMD is getting leverage, and this is just great. Let's hope the same occurs in the memory market.

>> No.63672211

>>63671771
I think AMD said AM4 will be supported at least until 2020, so 3+ more years after the initial 2017 Ryzens. Quite a few people upgrade their CPU roughly every 3-5 years - many of that particular group would be able to benefit from this if they want.
Of course, this selling point progressively degrades as time passes - if you only join the Ryzen platfrom in 2019, you're likely to find you'll need to buy a new mobo (or else a CPU older than current-gen) when you try to upgrade again around 2022-2024.
Also, maybe the latest mobo generation has features you like enough to buy a new mobo anyway - or maybe your old mobo dies a couple months after its 3-year warranty ran out.
In conclusion, there's some potential utility to upgrade paths, but it's limited.

>> No.63672233

>>63672066
Vega 11 and Vega11 are completely different things.

>> No.63672289

>>63672233
gasp

>> No.63672355

>>63672089
250mph

>> No.63672655

>>63671775
It wouldn’t compete, it would decimate. Atom is complete garbage.

>> No.63672724

>>63670105
lmao, that shill argument is so pathetic even though it's obvious to anyone that it's idiotic and beyond silly.

I've had plenty of systems in my time, mostly AMD. The CPU's always been stuck in the motherboard I bought with it throughout it's lifetime. I always upgrade motherboard/CPU/RAM because there's been progress between my upgrades - you couldn't stick DDR1 RAM in a DDR3 board, now could you? My Ryzen 1600X will sit in the motherboard I bought with it until I replace the board and the CPU and the RAM. And even if I were to replace the CPU ... it would be the RAM that's the important part. $200 CPU, $250 for 32 GB RAM and I or anyone would care about the cost of a $100 motherboard?

Regardless, I'm not going to pay another $200 to replace that CPU for a new "next generation" one that's perhaps 5% faster.

>> No.63672800

Ddr5 2018

>> No.63672824

>>63669951
I'm not saying that I don't believe the AMD hype or that they're necessarily going to under-deliver. Often times they've met or exceeded expectations.

But let's be honest. Most of the time they don't.

>> No.63672829

>>63672800
but DDR4 is still a useless meme

>> No.63672832

>Ryzen has worse IPC than Core i7 and i9
>Worse single core performance
>Not compatible with GenuineIntel software

Why the fuck would you waste your time with AMD?

>> No.63672834

>>63672355
you know what I mean wise guy? HERTZ MATTER!

>> No.63672860

>>63672834
>t. NetBurst

>> No.63672901

>>63669951
Yet AMD makes a GPU for Intel.

>> No.63672933
File: 95 KB, 499x700, DPzdU9DVoAczcg8.jpg [View same] [iqdb] [saucenao] [google] [report]
63672933

>>63671770
January or february.

>> No.63672959

>>63672724
>$200 CPU, $250 for 32 GB RAM and I or anyone would care about the cost of a $100 motherboard?
If $100 is nothing, why buy a $250 cpu? You could get a much nicer $350 cpu. Clearly the money acctually matters to you. Reusing ram is great. Ddr3/4 difference is minimal (for Intel anyway)

>perhaps 5% faster.
That was during the 2012-2017 amd monopoly.
This generation was a 50% jump in performance for the same price, assuming enough people buy amd we could keep getting huge bumps in performance.

>> No.63672970

>>63671794
>all the Intel fans believed Coffee Lake will work in Z170/Z270 boards
>now salty

We AMDfags are not such Turmp-voter level cucks as you Intelgfags are, at least.

>> No.63672993
File: 133 KB, 689x766, 48.jpg [View same] [iqdb] [saucenao] [google] [report]
63672993

>>63671824
>this level of damage control
I will remember you when Intel only releases new CPUs on desktop in H1 2019, around the time when AMD puts out Ryzen 3000.

Intel 9000 generation might not even have PCIE4. And DDR5 in 2019? - no, just no.

>> No.63673013

>>63672901
And Intel makes NICs for AMD. I don't see the problem here.

>> No.63673019

>>63672993
>Intel 9000 generation might not even have PCIE4
Probably not, it's coming out next year and Intel will likely want it out sooner rather than later in order to have proper competition against 8C Ryzen.

>> No.63673032

>>63672724
>I
>I
>I
>I
>I

>> No.63673045

>>63672800
2021, realistically.

Intel has a server refresh scheduled for 2018 (Cascade Lake). In 2019, they might launch Ice Lake. That one is still DDR4, too soon for DDR5 that year. And that platform will last 2 years. So Intel servers and HEDT might adopt DDR5 in 2021 or in 2022 if Intel is late.
Desktops won't switch before servers, and so won't notebooks.

>> No.63673077

>>63673019
Gen 9000 is Ice Lake (10nm+). I think it will only land in H1 2019. Late 2018 limited semi-paper launch of a few SKUs if we are very lucky.

>> No.63673107

>>63672993
Who cares, Intel will have SoC the size of a silver dollar that contain an LTE modem gddr4 cache and 128+GB of xpoint flash on board.

AMD isn't even relevant in modern computing outside of antiquated gamer rigs.

>> No.63673127

>>63673107
>AMD isn't even relevant in modern computing outside of antiquated gamer rigs.
Which is why all the big players are hoarding EPYC.

>> No.63673173

>>63673077
Not according to the latest rumors it's not. 9000 is going to be Coffee Lel 'refresh' with 8C CPUs in 2018, probably running on the rumored Z390 mobos. You really think Intel is going to wait over a year while taking it up the ass from Ryzen 2000 and do absolutely nothing, considering how they fucking rushed Skylake-X and Coffee Lel? Fat chance, they're going to have something out to try to compete.

>> No.63673186

>>63671824
It's too soon for DDR5, and as far as I'm aware PCIe 4.0 isn't bringing anything big to the table.

Also, there's no rule against Ryzen 2's chipset supporting it on new mobos but still being capable of running without it on old mobos.

>> No.63673227

>>63673186
Wouldn't they need a different memory controller to run DDR5? I doubt any Ryzen CPU is going to have redundant memory controllers, they'll probably stick to DDR4. I'm not sure how it works with PCIe exactly, that might be plausible since it's backwards compatible anyway, so a PCIe 4.0 capable CPU might be able to run PCIe 3.0 on older mobos.

>> No.63673280

>>63673127
They aren't relevant there either, Intel dominates and Qualcomm is the up and comer.

AMD's only profitable path is marketing "enthusiast" blow dryers with no drivers to miserly morons.

>> No.63673304

>>63673280
>Intel dominates
By that I can only mean "Offers worse performance for more money, except in a few niche applications" because that's what's going on right now.

>> No.63673309

>>63673107
>LTE, modem
Dude, Intel is dead in mobile. They utterly failed to get in that market. The couldn't do it when it was just Qualcomm and Mediatek, much less when the Jobs cult makes their own shit. And they only manage to sell their modems to Apple bceause said cunt is trying to get Qualcomm to go out of business.

>> No.63673340

>>63673173
>latest rumors
Latest rumours are just rumours, aka bullshit.
https://twitter.com/FanlessTech/status/936199410953916416
This guy has been right before, he gets to see Intel roadmaps and said he saw proof all of gen 9000 are Ice Lakes.

>> No.63673442

>>63673227
>so a PCIe 4.0 capable CPU might be able to run PCIe 3.0
That's a sure thing for both Intel and AMD; you've always been able to stick a PCIe 2.0 card in a PCIe 3.0 capable slot with no issues.

If you'd want to buy a CPU that does PCIe 4.0 for a board that can't is another question, you're probably better off just buying a new motherboard.

>> No.63673508

>>63671824
>DDR5
The jews are still raking big monies over ddr4 why stop selling those things?

>> No.63673514

>>63672901
AMD isn't competing in that market, and it's free money for AMD.

>> No.63673635
File: 71 KB, 888x499, 1fx2km.jpg [View same] [iqdb] [saucenao] [google] [report]
63673635

>>63673280

Yeah, fuck baidu and microsoft - who the fuck are they? Nobodies, thats who.

>> No.63673897

>>63672045
SHIT

>> No.63673960
File: 140 KB, 1057x872, DPF3vkZV4AEtTsQ.jpg [View same] [iqdb] [saucenao] [google] [report]
63673960

>>63673442
>If you'd want to buy a CPU that does PCIe 4.0 for a board that can't is another question, you're probably better off just buying a new motherboard.

Why, having PCIe 4.0 is not so important. Won't probably improve GPU performance much, and who needs over 4 GB/s bw from SSD.

And being able to stick new faster CPU into an older PC lets you save a lot of money, plus no need to reinstall/migrate your OS install. That's a great thing.

>> No.63673980

>>63669951
im not defending intel here because their chipset a year keeps goyim in fear policy sucks too, but why would you want to keep the same boards for multiple generations of CPUs? you rarely replace the CPU nowadays unless you are some benchmark whore. RAM and other shit is moving on and theres no point being stuck on old board.

>> No.63673984
File: 2.99 MB, 427x240, Starlight Breaker.webm [View same] [iqdb] [saucenao] [google] [report]
63673984

>>63673897
Starlight Combo Breaker!

>> No.63673995

>>63672901
psssh nothing persenel kid it was only business

>> No.63674038

>>63672800
In 29 days

>> No.63674040

>>63673980
For starters, keeping the same socket means that if your motherboard breaks 5 years from now you'll still be able to find a new replacement without having to replace the CPU as well.

>> No.63674120

>>63673340
Considering Intel's immediate and rushed reaction to Ryzen, I really doubt they'll let a Ryzen refresh sit on the market unchallenged for over a year, but we'll see.

>> No.63674125

>>63674040
good point

>> No.63674161

>>63674120
What can they do against it?
The only limitation for Zen is node, and it's about to go away.

>> No.63674210

>>63674161
Release 8C mainstream CPUs, as previously stated.

>> No.63674223

>>63674210
So, still pricey, a mobo a year, jizz and 16 lanes?
Meh.

>> No.63674231

>>63674120
They could do something like i7-8720K. Not sure they can push the clocks higher though, they already push TDP and bin aggressively. They could make some +100-200MHz low-volume press-edition for the reviews, maybe.

>> No.63674353

>>63674223
I don't get it, what point are you trying to make? Intel most definitely won't try to compete because you don't like their CPUs? Surely you cannot be so naive to believe nobody would buy Intel CPUs, you know full well they'll get a good chunk of sales out of OEM shit, fanboyism, misinformed purchasing decisions and mindshare anyway. If AMD doesn't universally surpass them in gaming, they'll get a chunk of sales from there as well.

Mainstream Ryzen is 16 lane too, I'm not too sure that will change while they're still on AM4 either.

>> No.63674366

>>63674353
>Mainstream Ryzen is 16 lane too
It's 20+4 though.

>> No.63674385

>>63673984
That looks so fucking lazy, fuck CG

>> No.63674508

>>63674366
You mean 16 + 4 + 4, you ain't getting more than 16 lanes from the CPU to the actual PCIe slots.

>> No.63674510

>>63674366
And some SATA and USB 3.0 ports also go straight out of the processor, so they don't steal bandwidth from the pice 3.0 ×4 that both Intel and AMD use to connect the southbridge with other conectivity. For Intel, there is awfull lot of contention there, the AMD platform is way better in that.

>>63674385
You need to watch it in HD. I got to say that I like the TV version of the first season, though. better overall atmosphere.

>> No.63675664

>>63674210
I wonder if that can happen. Octocore Coffee Lake is mostly speculation and I doubt it exists. Ice Lake yeah, but that is H1 2019.

>> No.63675707

>>63671976
Only poorfags upgrade every 5 years

>> No.63675783

>>63671976
People trying to make people embarrassed of not having enough money are always shills for the companies that try to rip you off.

Shoo, shitter.

>> No.63676330

>>63671824
You could use Phenom II's on Socket AM2+, they were just limited to DDR2. It would likely be the same, I think Zen2 will be split between two different chipset, AM4 and AM4+.

>> No.63676599

>>63676330
You don't necessarily need to switch socket to add PCIE 4.0, see X79 and Z67/Z77 platform on Intel.

>> No.63676683

>>63676599
But then mobo vendors tried to add PCI-e 3.0 to AM3+ and never got it working, FX was stuck with 2.0 forever. I wouldn't bet on PCI-e forward compatibility.

>> No.63676714
File: 14 KB, 938x120, 1503314155502.jpg [View same] [iqdb] [saucenao] [google] [report]
63676714

>> No.63676765

>>63676714

With coffee lake not going to work with the ice lake socket the master jew is straight up fucking consumers and /g/ gobbles it down asking for more. The Intel damage control is explicitly ignoring that AM4 will be supported for years whereas Z390 is turbo dead.

>> No.63677040

>>63676683
>But then mobo vendors tried to add PCI-e 3.0 to AM3+ and never got it working, FX was stuck with 2.0 forever.
Uh, that is not what happened.
AMD just never released a chipset that would feature PCI Express 3.0, that was not motherboards makers thing to "try to implement". PCIe controller was in northrbidge on teh AM3+ chipsets, which in turn was connected to CPU by HyperTransport bus.
And AMD didn't do it because the interim managers in 2011-2012 gutted the desktop platform roadmap. Originally FX succesors were supposed to have more cores and go into FM2/FM2+ socket.

>> No.63677188

>>63672724
what if they still use am4 when you want to eventually upgrade though you could put the money you would spend on a new mobo towards your cpu or gpu

>> No.63677299

>>63673107
>>63673280
>Shill faggot tripfag appears
FUCK OFF ALREADY

>> No.63677307

>>63677188
>you could put the money you would spend on a new mobo towards your cpu or gpu

It's a shill, shills don't care about that, they just want you to threw your money at their employer/fav company.

>> No.63677449

And this is why I bought an overkill motherboard. My Asus Crosshair VI will be one of the first motherboards to get any required BIOS flashed for the new chips. I'll probably be skipping Ryzen 2 though. Just wait for Ryzen 3 and those sweet sweet 4.5+GHz clocks.

Once Ryzen is capable of clearing the 4.4ghz barrier, Intel has no options left. Literally their only strength as it stands right now is the ability to turn the wick way up on clock speed. At the cost of thermals, power consumption, and insane cooling requirements of course.

>> No.63677459

>>63672832
>>Worse single core performance

>> No.63677703

>>63677459
>gaymur

>> No.63677721

>>63672832
>IPC
>not IPS

get it fucking right you reviewer watching faggot

>> No.63677743

I have heard that the AMD platform have had some stability issues.

>> No.63677753

>>63677743
Have you heard of stability issues with Intel platforms?

>> No.63677798

>>63677753
Never, Intel is practically perfect and i trust their ethical marketing practises.

>> No.63677849

>>63677753
>Intel GPU driver
nuff said

although
>multiple buffer overflows and privilege escalation bugs in ring -3 critical infrastructure
is a nice one too

>> No.63677851

>>63677753
At worst Intel will just brick itself or disable some of your hardware, I'll just buy new ones then

>> No.63677896

>>63677753
I'd say the Skylake hyperthreading bug is pretty high up the list.

>> No.63677897

>>63677753
It's kinda sickening how people still flock to Intel "stability" even after the whole Avoton/Rangely shitshow.
Which seems to have been swept under the rug, what a garbage company.

>> No.63677940

>>63677753
No problem.

https://www.reddit.com/r/PFSENSE/comments/7h5ck1/psa_if_your_i350_gbe_nic_is_causing_your_intel/

>buy very expensive "quality guaranteed" NIC
>it disables one of your memory channels

Intel inside.

>> No.63678007

>>63677940
How in the hell does this even happen? Isn't there some kind of reference design memory wiring/trace for all mobo vendors? Unless this is done due to some weird arcane compatibility with some other onboard module.

>> No.63678083

>>63669951
No thunderbolt unfortunately, but seriously, why isn't there a non proprietary PCIe + Displayport external port solution?

>> No.63678125

>>63677798
Nice

>> No.63678188

>>63678083
TB is no longer proprietary, starting H1 2018.
>>63677897
Oh y-yes goyim, bricking Atoms in fuckoff expensive Cisco switches are NOTHING.
Remember six gorillion!

>> No.63678208

>>63672724
>$100 mobo

Sounds like you're the retard.

>> No.63678565

>>63677753

ITT someone who has never read Intel's errata for a given architecture. The list of bugs a given Intel cpu has is as long as my arm (not saying AMD doesn't have its own known bugs but claiming Intel without them is retarded).

>> No.63678596

>>63678565
Careful, you might summon HIM.

>> No.63678633

>>63673960
It's probably meant to be an option for people who have money to spend.
You won't have to upgrade to enjoy DDR5/PCIe4.0, but computer enthusiast Joe with the Threadripper setup can.
Much more ethical than Intel's or Nvidia's shit.

>> No.63678643

>>63672800
>le more DDR meme
>le faster PCIE (that wont get used) meme

you could still use PCIE 1.0 and ddr2 and you would be fine for 99% of things, games included

>> No.63678668
File: 40 KB, 340x104, hmmm.png [View same] [iqdb] [saucenao] [google] [report]
63678668

>>63678083
Thunderbolt is a Firewire tier meme

>> No.63678683 [DELETED] 

>>63673984
Which anime? Haven't been up-to-date lately.

>> No.63678685

>>63677798
what a wise go- I mean guy you are!

>> No.63678696

>>63678683
>not recognizing Nanoha
Newfag.
>>63678668
No, it actually has its use cases due to being a glorified external PCIe x4 link.

>> No.63678716

>>63669951
I worry when AMD speaks out softly like this about their next Zen variant.

Chances are the Zen+ wont have high clocks.

>> No.63678717

>>63671797
Epyc cpus with double the cores/threads already exist. Threadripper is just epyc with 2 dies disabled/removed. It would be quite easy for make them to use the same tr4 socket since the empty spots can simply be filled.

>> No.63678734

>>63678208
Sounds like his builds were mostly intel

>> No.63678744

>>63671824
>Haswell used ddr3/ddr4 just fine
>no one else could possibly do such a thing that has already been done before
How old are you?

>> No.63678747

>>63669951
That's the AMD strategy since 1993.
"you could upgrade to the newest intel slot, or you could use the old intel slot with the new AMD CPU"

>> No.63678754

>>63678716
It's the same company that said they are not going to enter HEDT market.

>> No.63678756

>>63672970
>Turmp-voter level cucks
>Turmp
>hating Trump
lol kill yourself

>> No.63678769
File: 132 KB, 1728x803, 6 years.png [View same] [iqdb] [saucenao] [google] [report]
63678769

>>63677459

>> No.63678820

>>63678769
6 years, and intel only increased the performance in 5% being in the lead the whole time.

Quite a shameful picture.

>> No.63678839

>>63678820
Because they didnt need to AMD has been pathetic in terms of competition.

>> No.63678857

>>63678839
Well congratulations to Intel!
Now their ass is on fire, nodes delayed and roadmaps in shambles.
They fucked up.

>> No.63678881
File: 251 KB, 852x939, 1385420441583.jpg [View same] [iqdb] [saucenao] [google] [report]
63678881

>>63678839
>being such a jew you dont do anything to increase performance because no competition

AMD will have more powerful CPUs that intel by 2020

>> No.63678882
File: 193 KB, 364x397, 1475607897834.png [View same] [iqdb] [saucenao] [google] [report]
63678882

>>63678683
Nanoha

>> No.63678916

>>63678881
Actually, by H1 2019.
7LP will be a huge nail in Intel's coffin.

>> No.63678924

>>63670166
We actually don't. Supporting AM4 does not mean supporting older chipsets. Intel has supported LGA1151 yet Coffee Lake/chipset is incompatible with Kaby Lake/Skylake. LGA1151 itself is still supported. AMD could do the same and banish X370 forever. They still are technically supporting AM4 in this case though.

>>63671976
Yes, new features like RGB lighting on the mobo. What's next, UV/IR lighting on the mobo? Microwaves when?
>>63672724
Some people buy the C6H/C6E/X370F Strix for a reason. Also, they understand something called selectively selling your old hardware. Alternatively, they can get Ryzen 2 and a cheap mobo and swap the CPUs out.

>> No.63678941

>>63678881
I heard the same thing about bulldozer I hope you're right so that we get some processors that actually warrant an upgrade.

>> No.63678973

>>63678941
Bulldozer was a bad core on a decent node.
Zen is a great core on a bad node.
Do the math.

>> No.63678996

>>63678756
>not hating Trump

>voters got cucked on healthcare, cucked on taxes, cucked on environment, everybody will pay so that corporations and rich sons can get richer
>but they like it and convinced themselves that being NTRed is great because of liberal tears
I don't know but that sounds like a thing to hate. Not NTR fan myself and Americans are now sadly the gravest IRL case of cuckage I have seen this century.

>> No.63679009

wtf i love amd now

>> No.63679026

>>63678996
>I know nothing about current geopolitics: The Post

>> No.63679033

>>63678924
>Intel has supported LGA1151 yet Coffee Lake/chipset is incompatible with Kaby Lake/Skylake.
But that was done on purpose.
Now in that video n OP, the AMD dude actually said that current boards will accept new CPUs in 2019-2020.
Of course, stuff can happen that will fuck it up, but it seems they aim for compatibility with AM4. As an AM4 board owner, yay. I think I'll appreciate this when I'll drop in some cheap 8-12core bought used in 5-6 years. Will be my longest-lived machine I'm sure, even if I'll be somewhat limited with just 2 DIMM slots (mini-ITX).

>> No.63679067

>>63679026
>geopolitics
>ruining healthcare, tax increase for 99% of people pushed through by distracting people with >muh liberals
https://en.wikipedia.org/wiki/Geopolitics

I don't want to troll, but you literally did
>I don't know what "geopolitics" means, the post

>> No.63679114

>>63679067
>domestic politics are not influenced by international powers and their respective economies
>healthcare is ruined
>obamacare was good and wasn't bankrupting us by literally robbing fanny may and freddie mack
>tax increase for 99% of people
>implying that's even what's happening
>muh liberals
Quit watching CNN and watch fucking CSPAN or something, fucking christ. The fake news is so rampant I can't believe you believe the garbage you're spewing.

>> No.63679130

>>63672724
>I always upgrade motherboard/CPU/RAM because there's been progress between my upgrades - you couldn't stick DDR1 RAM in a DDR3 board, now could you?
Anything past DDR4 is redundant.
DDR5 will have even more negligible micro-advancements than 4 has.

And what else is there new and interesting?
USB 3.0, USB-C, even fucking M.2 is already put into boards nowadays.

>> No.63679153

>>63679114
nice denial you have there. I guess it must be really comfy

>> No.63679167

>>63679114
>>obamacare was good and wasn't bankrupting us
Exactly.

>> No.63679169

>>63679153
ur a faget

>> No.63679185

>>63679026
>geopolitics
You are extremely uneducated if you have no idea what that word means.

>> No.63679218

>all this samefagging

>> No.63679297

>>63679169
you know, in cases of a deceptive campaign like these, you usually call the enemy names based on what you do or are guilty of yourself
trump groupies called the blue voters "libcucks", you are calling me faggot
I hope you catch my drift

>> No.63679310

>all this reddit
Do you guys not know that reddit loves AMD shit? Go shill on your own website pajeets. 4chan is an Intel website.

>> No.63679404
File: 29 KB, 364x597, post dragon hugging a pillow.jpg [View same] [iqdb] [saucenao] [google] [report]
63679404

>>63679310
right, shill

>>63678973
I don't think 14LPP at GloFo is particularly bad. At worst it's not ideal. You can't call it bad if you are calling 32nm SOI decent.

32nm had horrible yields and killed Llano, cost AMD lots of money, calss action suit from shareholders even. Delayed BD. At its end it had good clocking potential, but it was at the time when AMD needed power efficiency.

Even now with Zen, the enthusiast give 14LPP shit for not allowing clocks over 4GHz, but they fail to see that it gives AMD power efficiency that is on par with Intel and that is what allowed Zen to make a splash and Epyc to crack servers. If Ryzen had worse power consumption than Intel alternatives, it would get bashed in reviews and waved off.

>> No.63679433

>>63671946
JIM KELLER THE SHIT WRECKER

>> No.63679449

>>63679404
14LPP is not a bad node, but we're talking about client products here where absolute performance for your favourite peenus measuring contest if everything.
12LP will fix that.

>> No.63679475

>>63679449
>12LP will fix that
Better wait how it pans out and not go ahead of ourselves. I don't want to have too big expectations and then be disappointed.

So I only expect 4,3-4,4 GHz XFRs, for now. Improved memory controller as a bonus.

>> No.63679480

>>63679449
Love me some peenus measuring contests honestly.

Aaaah reminds me of back in the days when my 7970 did 1375/1880 or something on water and i shat on 680s in firestrike and 3d11.

Cot DAM those were the days senpai.

>> No.63679501

How often do people upgrade just their CPUs exactly? I've never understood the benefit behind this.

>> No.63679509

>>63679475
14LPP would also fix it, if SR wasn't a server die.
They went with 78/9T for Summit Ridge because it's a server first product.
Expect AMD to go 84/10.5T for Pinnacle Ridge, it's client only.

>> No.63679542

>>63679509
That's just speculation/conjecture though, not confirmed?

>> No.63679577

>>63679542
Confirmed, go check Samsung's spec sheets for 14LPP.
They offer 3 options: 78CPP with 9T libs for high density, 84CPP with 9T libs for high performance, 84 CPP with 10.5T libs for ultra high performance.
Because SR (ZP-B2) stays as their server die of choice, PR will go for nuclear clocks.
Expect good old 125W SKU to be back.

>> No.63679597

Prior even confirmed that PR is client-only.

>> No.63679613

https://pc.watch.impress.co.jp/img/pcw/docs/733/713/html/9.jpg.html
Also one should expect additional performance optimizations, given that it's GloFo's own node.

>> No.63679692

>>63671771
I have an i5-4440 bought in 2014 and modelling stuff in Fusion or any other CAD program is starting to become a chore. I don't know if it's just because software gets bloated overtime or if the transistors inside the chip actually start losing electrons or whatever. I'll probably wait one or two generations before upgrading.

>> No.63679697

Given that AMD is separating client and datacenter products again, I'd expect Ryzen 3000 series to be fabbed on 7HPC with some 14-15 metal layers to make even six gorillion look like a joke.
Goddamit AMD, it's nice to see you back.

>> No.63679727

>>63671976
>and only poorfags care around recycling their mobo when they change CPU.

>USB controller is just fine
>SATA interface is just fine, it even has one of those tiny slots for SSD's on the board
>PCI-e works just fine, don't even use all the slots and could put a PCI-e ssd in one if one that is cheap enough is released
>memory just works
>wifi works just fine
>gotta toss everything in the garbage bin because there's a new CPU

No, thanks

>> No.63679741

>>63679727
Exactly.

>> No.63679815

>>63678188
>TB is no longer proprietary, starting H1 2018.
It's a cost-free license, not open-standard.

>> No.63679891

>>63679697
>I'd expect Ryzen 3000 series to be fabbed on 7HPC with some 14-15 metal layers to make even six gorillion look like a joke

Brainlet here. Can you explain this for a brainet like me? What is 7HPC and how do these metal layers matter and what exactly are they? Thanks in advance.

Just really want to learn all of this.

>> No.63679918

>>63679815
No longer proprietary you say, but does it actually run on non-Intel platform?
Doesn't seem so, at least not with all the functionality needed (like display output and so on).

>> No.63680004

>>63679918
I'm quoting someone else m8. Intel said they'd make Thunderbolt royalty-free to device and component manufacturers. They still control the license, it's just free. Considering Nvidia has a similar model for basic physX and gameworks I suspect AMD won't be able to use thunderbolt in the same way they can use any of Nvidia's shit. They should've continued working on XGP desu

>> No.63680075

>>63680004
I liked Dock Port. Simpler, cheaper. Too bad nobody jumped on it.

>> No.63680146

>>63679501
If you have a great warranty like Microcenter's, it's worth it.
Can literally upgrade your CPU every year (or two) for under $100. (Usually $30-70)

>> No.63680168

>>63680146
You can just sell your old CPU on reddit and do the same thing.

>> No.63680175

>>63680168
I got my Ryzen 1700X for around $270 near black friday.
Whenever Zen+ comes out, that's $270 I have in credit.
Can't do that with Reddit or Craigslist or Ebay.

>> No.63680190

>>63680175
Explain

>> No.63680196

So how doable does 4.5-4.6ghz on ryzen+ sound on a custom loop? Based on current info and reasonable speculation of course.

>> No.63680212

>>63680196
Threadrippers are already hitting 4.2GHz in some cases, so I expect 4.6 to be possible with the higher end SKUs on regular Ryzen

>> No.63680217

>>63680190
I paid $30 for the warranty at Microcenter.
Q1 2018 when Zen+ comes out, I can say it doesn't work, even though it works fine, they give me credit. (same amount I paid at the time)
I then use this in-store credit for the Zen+.

>> No.63680237

Too late, I already bought an 8700k

>> No.63680250

Thank you based AMD

>> No.63680425

>>63678769
That i5 isn't running at 3.3ghz though

>> No.63680542

>>63679297
tbqh your politics are a dead end
>invite millions of foreign voters in to bolster your numbers against dur rethuglicans
>they aren't like you in any way shape or form, just vote with you to get more of their family and friends in
>all of a sudden the majority of the electorate does not care about healthcare
>all of a sudden the majority of the electorate does not care about the environment
>all of a sudden the majority of the electorate does not care about taxes because they're used to rampant corruption and elite rule
>all of a sudden the majority of the electorate does not care about religious separation because they are used to theocracy
>all of a sudden the majority of the electorate does not care about anything except minority privileges and what color the person who just got elected is
I've watched this for decades after attending university in your country. You western liberals are simultaneously the most intelligent and yet the most incredibly stupid and short-sighted people in the world. You did not simply cut off your nose to spite your face, you drove a dagger into your own guts and continue to twist it in a fit of rage.
>You never realized that you were being manipulated by vast international capitalist interests in order to submerge your own democratic power.
I could almost pity you, but for what you did to the working class that looked up to your guidance, nothing but hatred could be appropriate.

>> No.63680560

>>63680542
desu your post does not belong on g

>> No.63680613

>>63680560
>anon bad mouths gonald grumpf and his two scoops
>gets btfo by anons who know what they're talking about
>you come in and say this shouldn't be on /g/
The whole thing shouldn't be on /g/, but you singling out this anon is just trash. Kill yourself.

>> No.63680640

>>63680217
Come on man, don't abuse their warranty. That's how you get bad customer service

>> No.63680647

>>63680613
this was the only pol related post i read, im not going to go up and look for every post that is pol shit and mention it just so you wont be butthurt about me missing a bunch of other retards talking about unrelated shit.

You fucking kill yourself you dumb mutt.

>> No.63680652

>>63680640
The guy who told me about the warranty and sold it to me literally told me to do this.

>> No.63680660

>>63680613
>gets btfo by anons who know what they're talking about
>know what they're talking about
uh when did that happen?
It would sure be nice if things became real when you imagined them, right?

>> No.63680678
File: 103 KB, 1280x720, the_rolling_girls-04-banko-western_tokyo-green_bows-bandage-yelling-in_your_face.jpg [View same] [iqdb] [saucenao] [google] [report]
63680678

>>63680647
No! Kill yourself right now!

>> No.63680696

>>63680542
>nothing but hatred could be appropriate
So you would fuck your situation up even more because you are angry? That makes no sense if all you get in return is that the people you voted in take even more advantage of you.
Being satisfied that you made some imaginary enemy they showed you angry doesn't really feed you or anything.
It just made somebody laugh at your expense all the way to the bank.

>> No.63680736

>>63672933
Source?

>> No.63680772

>>63671710
Buy EBIN

>> No.63680781

>>63678769
I love this screenshot, IGNORE THE 2600k WITH A 124 AT THE BOTTOM GOYS!

>> No.63680918

>>63680696
>So you would fuck your situation up even more because you are angry?
This is what makes you so contemptible. You have no self-awareness. You don't even care that your very own vote is being washed away by millions of competitors. To you, your own asshole and the political talking points that fill it are all that exists. You are an empty shell, simulacrum of the soulless capitalism that controls your nation. All the while you pretend to be the antithesis of capitalist excess. You are truly a wretched and ugly creature.

>> No.63680945

>>63680918
>this entire post
w-woah there buddy, calm down
you don't have to kill him that hard

>> No.63680962

>>63672829
Ryzen has higher IPC according to Agner Fog, and i7 has the same IPC as i9 dumb /v/edittor
gb2youtube

>> No.63680978

>>63680652
If he told you to bend over and take it bareback, would you?

>> No.63681001
File: 781 KB, 1440x834, 1512046698332.png [View same] [iqdb] [saucenao] [google] [report]
63681001

>>63680978
Of course. Duh.

>> No.63681038

>>63680918
Is that some pasta?

>> No.63681049

>>63680736
Should be Nanoha's boatlights daughteru from Nanoha.

>> No.63681085

>>63672724
>5%
Of course you wouldn't.
Except we already know it's going to be much better than that.

Zen is immature and there are many things that can be fixed - but even without any internal changes, even on 0% IPC improvement - 14nm to 12nm is 10% higher clocks and glofo is saying 7nm will be 5ghz baseline.

>> No.63681096

One thing I've realized about same socket/chipset is that if you want to upgrade (probably 2-3 gens) then you can upgrade the CPU first while keeping the same mobo, and if you want more mobo features you can get that whenever you want such as in cases when prices are lower (DRAM). With a full socket change, you have to upgrade everything together or else you're fucked with the upgrade. You also get a backup mobo in case things go wrong once you swap to a new one. I've only bought Intel and had a CPU and mobo fry on me so this is actually important IMO.

>> No.63681113

>>63679577
I expected higher clocks but I didn't know that 12LP will have higher perf libraries since I thought it was based mostly off 14LPP

>> No.63681205

>>63677753
>avoton hardbricks
>cougar point SATA ports dying

And in both cases, requiring a respin of silicon and RMAs of affected hardware.

>> No.63681312

>>63681113
Was 14nm LPP limited to only 7.5T? If AMD could only use that for their caches then moving to 9 or 10.5 for the SRAM alone would net a large performance increase, that's without touching any of the other critical paths that were also 'gimped' by 7.5T

>> No.63681404

People don't seem to understand that if a node is viable for RF/Networking it's in most cases a high frequency node, your NICs are pretty fucking useless if you can't clock them over the moon.

>> No.63681460

>>63670191
*their
Nigger.

>> No.63681610
File: 724 KB, 1920x1081, 1080p vega 64 cpu comparison.png [View same] [iqdb] [saucenao] [google] [report]
63681610

>>63673280
Good god, you're still posting your dumb shit
Either wake up to the data or get off /g/, retard

>> No.63681659

>>63673280
>Qualcomm is the up and comer
I've heard it all now. Some truly retarded shit but this takes the cake

>> No.63682556

>>63671946
B A S E D
A
S
E
D

>> No.63682601

>>63681312
GloFo has only ever fabbed 78/9T one.

>> No.63683238

>>63680217
Your a crook, you know that.
I hope one of your components fail the day after warranty.

>> No.63683633

>>63678083
we already know the new chipset will have thunderbolt support...

not to mention almost all the x399 mobos have the rumored pins for the thunderbolt already on board

>> No.63683659

>>63679449
14lpp is a safe node for a testbed which ryzen is
7nm ibm black magic is were it all ends anyway

>> No.63683761

>>63680781
Thats a non oced 2600k oced it should be on par if not better than the 2500k.

>> No.63684553

>>63679310
reddit loves LGBTQ too

>> No.63685402

new geode when

>> No.63685448

>>63679130
kek I remember when people were saying ddr4 was a meme and 3 was all you needed

>> No.63685511

>>63669951
vega reference production halted. if someone wasn't sure it's a flop, you can be sure now.

>> No.63685530

>>63671771
>What percentage of people upgrade their CPU every year? The 1%?
everyone did that when it was possible, nobody wants to buy new motherboard with cpu for no reason

>> No.63685546

>>63685511
How is that related to AM4 and where's AMDs press-release on that anyway?

>> No.63685572
File: 519 KB, 876x844, increasingly_nervous_gman.png [View same] [iqdb] [saucenao] [google] [report]
63685572

>>63669951
>>63670105
>>63670166
>>63671669
>>63671683
>>63671693

Give it up, guys.

>> No.63685594

>>63671771
>Every year
Literally nobody does this, have you ever thought that maybe people buying cpus are doing it because they're old computer is weak and shitty?
I've had my pc for nearly 7 and it's starting to show its age.

>> No.63685720

>>63685546
>Last but not least, Prior also confirmed that will be halting the production of RX Vega 56 and 64 reference models.

Interview.

>> No.63685739

>>63671669
>tfw people bought threadripper because it was a top tier but now it's 2020 and they switched to threadripper 3 so you buy one of gen 1 threadrippers and use it with care because you want to protect it

it's like taking care of old clasic used cars anons
feels good

>> No.63685752
File: 38 KB, 645x729, Many-brein.png [View same] [iqdb] [saucenao] [google] [report]
63685752

>>63671771
>gee i wonder how many people will buy it
>it's not like new young consumers are entering the market each year

pic rel is you

>> No.63685784

>Intel finally gets some competition
>Forces them to knee jerk react with early Coffee Lake launch and 'moar cores'
>Affordable multi-core CPU's becoming available for everyone
>Both AMD and Intel fanboys benefit

And this is a bad thing how?

>> No.63685811

>>63685784
This. AMD brought us a huge jump in multithread performance/$ and ignited competition, whats not to like?

>> No.63685822

>>63685811
Kaby Lake buyers being salty as fuck probably. They got cucked hard.

>> No.63685825

>>63673984
D-did the cute blonde die?

>> No.63685833

>>63685811
>whats not to like?

why are you asking such ridiculous question goym ??

>> No.63685883

>>63685825
She's not cute

>> No.63685900

>>63685572

>with coffee lake out and pascal AMD is sure to be bankrupt for reals this time guise

>> No.63685912

>>63685822

EoL inside of a year - truly the greatest of ally.

>> No.63685955

>>63678996
>>>/pol
go on, amuse me

>> No.63686293

>>63672135
>Being this autistic

>> No.63686348

>>63679114
Ha, Republicans watching CSPAN. They only get like every third word of the broadcast, Republican voters are double-digit IQ or lower for the most part.

>> No.63686553

What about Ryzen Mobile being DOA?
The chip is good but if all notebooks have this shameful battery life is pointless.
https://techreport.com/blog/32893/here-a-first-look-at-the-battery-life-of-hp-ryzen-powered-envy-x360
https://techreport.com/blog/32904/how-much-does-screen-size-matter-in-comparing-ryzen-mobile-and-kaby-lake-r-battery-life
https://techreport.com/news/32912/amd-confirms-that-the-hp-envy-x360-uses-mobile-xfr

>> No.63686571

>>63685448
they are right

>> No.63686572

>>63686553
The battery life is about the same as x360 with KBL-R.

>> No.63686615

>>63686553
That wouldn't be AMDs fault though
>>63685448
There is not much of a difference. Still haven't bothered with ddr4 and I have no problems in anything. I'll probably jump to ddr5

>> No.63686639

>>63671946
Pat gelsinger

>> No.63686655

>>63671771
>Buy Ryzen + mobo on launch
>can upgrade my CPU again at the end of AM4, thus extending the longevity on my current system by another 3-4 years, easily, without having to replace the entire system (only CPU, no mobo/RAM replacements)
Gee, that was fucking hard.
Spend 500$ first year and then another 200$ after 4 years, system lasts 8 years total.
Alternatively, buy a completely new system every 4-5 years, 500$ every time. Wow, where did the moneys go!

>> No.63686679

>>63686639
I miss him (and Otellini).

>> No.63686695
File: 12 KB, 283x320, rabbi.jpg [View same] [iqdb] [saucenao] [google] [report]
63686695

>>63686655

>Wow, where did the moneys go!

>> No.63686721
File: 143 KB, 730x428, 1428766231174.jpg [View same] [iqdb] [saucenao] [google] [report]
63686721

>>63678208
Yeah lol, for just $100 you wouldn't get a real gaming motherboard, amirite lol.

>> No.63686803

>>63681085
>14nm to 12nm is 10% higher clocks
No, the slides said the 12 nm process was 10% better performance than "industry standard 16 nm processes", it didn't compare it to GloFo's 14 nm.

>> No.63686833
File: 244 KB, 1633x939, Global-Foundaries-12LP.jpg [View same] [iqdb] [saucenao] [google] [report]
63686833

>>63686803
>10% better performance than "industry standard 16 nm processes"

>> No.63686845

>>63671771
Can only agree. I'm still on my i5 2400.

>>63679692
>if the transistors inside the chip actually start losing electrons
Sure hope you're just pretending to be retarded.

However, you may actually want to try replacing the thermal paste. I did so recently in the above mentioned computer, and it actually made a very noticeable difference in that the chip had started throttling now and then before I replaced it.

>> No.63686846

>>63686833
That's at the same power.
Who cares, what matters is how high they will push fmax.

>> No.63686856

>>63686833
>over industrry 16nm FinFET solutions
Are you really just mad because I didn't quote it exactly word-for-word? It was obviously still correct.

>> No.63686891

>>63686856
wasn't meant as a imsult, you autistic fuck. just the source of your damn claim, now suck a dick

>> No.63686931

>>63681049
On the claim, not the pic

>> No.63687011

>>63686891
You first

>> No.63687025

>>63686845
I've got the same board as that other guy. How do I know if its throttling cause I've had it for 3 years now

>> No.63687232

>>63687025
In my case, I was just looking at the CPU frequency and temperatures.

>> No.63687279

>>63673107
>>63673280
/g/ is the worst when it comes to taking the bait

>> No.63687286

>>63671946
JEW KILLER

>> No.63687341

>>63678643
>t. Athlon 64 X2

>> No.63687351
File: 206 KB, 999x1072, not-doa-just-dead.jpg [View same] [iqdb] [saucenao] [google] [report]
63687351

>>63686553
>What about Ryzen Mobile being DOA?
From what I can see Ryzen mobile isn't dead on arrival, just .. dead.

Browsing the laptop selection in local stores based on last arrived I see a lot of laptops with old re-cycled 28nm cores - like the A12-9720P. These are the very old AMD Excavator cores re-cycled with DDR4 support and no other improvements.

There's not a single AMD Ryzen Mobile laptop in this country, not that I can see or find anyway. It's all Intel+NVidia and now these "brand new" pure garbage AMD APUs.

I don't really get what AMD and the OEMs are trying to do here, perhaps "regular people" will buy these cheaper laptops but they would have to be rather uninformed. This Excavator technology from 2011 isn't remotely close to competitive with Intel's latest offerings. ..then again, they've also put out a bunch of last-generation chips like the AMD A10-9700 for desktop AM4's too.

It's just ridiculous. Though I could actually see a place for the desktop parts if RAM was way cheaper, a $60 CPU and a $60 motherboard and you've got a working basic computer.. would make sense if you could get enough RAM to make it usable for $60. But you can't.

>> No.63687362
File: 13 KB, 399x400, nupidstigger.jpg [View same] [iqdb] [saucenao] [google] [report]
63687362

>>63672832
>GenuineIntel software

>> No.63687399
File: 344 KB, 2343x2486, yunocchi.png [View same] [iqdb] [saucenao] [google] [report]
63687399

Bought Ryzen 1600 and B350 dirt cheap on black friday, knowing that AM4 socket is viable until 2020.

Gonna plug the 10nm 16 core Ryzen there when EOL approaches.

>> No.63687467
File: 1.39 MB, 862x1291, lmao.png [View same] [iqdb] [saucenao] [google] [report]
63687467

>you definitly need a new mobo, goyim. trust your merchant!

>> No.63687502

It would be nice if AMD had thunderbolt.

>> No.63687768

>>63678820
>6 years, and intel only increased the performance in 5% being in the lead the whole time.
Meanwhile, the graph shows 10% improvement just between the 6700K and 7740X. Comparing the 7740X to the 2600K, which would match your claim of 6 years, the 7740X is 58% faster. If you would please, then, exaplin where you're gettint 5% from.

>> No.63687791

>>63687502
What do you actually use Thunderbolt for, outside of eGPUs?

>> No.63688010

>>63671824
you say that like those things will be of any use

>> No.63688026
File: 19 KB, 217x232, 1462161235421.jpg [View same] [iqdb] [saucenao] [google] [report]
63688026

>>63672832
>>Not compatible with GenuineIntel software

>> No.63688027

>>63686721
I used to buy mobos for $100. Always had strange hardware issues with my PCs. My 6th mobo I decided to splurge and got a highly rated one for $250. All those annoying issues and problems disappeared.

You get what you pay for.

>> No.63688054

>>63688027
I also always buy motherboards for around $100, still do, and never had an issue.

>> No.63688055

>>63687768

Not that anon but I imagine if you normalise clocks that gives 6%.

>> No.63688058

>>63687768
IPC =! Clock rates
Overclocking the same architecture doesn't make it a new architecture.

>> No.63688080

>>63688054
I said the same thing back in 2012. Never gonna cheap out on a mobo again.

>> No.63688109

>>63669951
To be true I'm more likely to buy Intel now as IME apparently has a disable bit. Trustzone doesn't and is also a larger attack vector because it is also in mobile chips

>> No.63688148

>>63688055
>>63688058
The 2600K gets 36.5 ST points per GHz without turbo or 32.6 with turbo, whereas the 7740X gets 45.6 without turbo or 43.6 with turbo. That's still between 25% and 33% better depending on which one you count.

>> No.63688223
File: 409 KB, 1209x1050, the pussie destroyer.jpg [View same] [iqdb] [saucenao] [google] [report]
63688223

>>63671946

>> No.63688270
File: 172 KB, 576x1024, 8012616222_a59ec94002_b.jpg [View same] [iqdb] [saucenao] [google] [report]
63688270

>>63688223
>tfw no amd gf

>> No.63688407

>>63686931
Sorry, didn't expect that. The usual anon here doesn't care about sources.
http://www.gdm.or.jp/crew/2017/1119/243271/2
https://benchlife.info/amd-cpu-gpu-plan-for-2018-include-navi-pinnalce-raven-ridge-12032017/

>> No.63688415

>>63670166
yes and no.

we knew the socket would still be used, however we didn't know if amd would force compatibility as a thing.

we still don't know if zen+ and zen2 will be fully taken advantage of on current motherboards, especially considering zen2 may be going for 6-8 core ccx opposed to 4 if amd's slides are anything to go by.

>> No.63688455

>>63688415
>6-8 core ccx
Very unlikely. The 4-core limit of a CCX is probably set by the scalability limit of the crossbar or whatever local fabric design it is that they're using inside the CCX.

>> No.63688576

>>63672970
Voting for Trump is good though

>> No.63688595

>>63688455
amds roadmap has 48 core cpus, so either they are making a 48 and a 64 core epic with 8 dies, or they are dropping in another 2-4 cores a ccx

I think the shift to 7nm and increase the core count to 6, along with increasing the cache by a factor of 2-4 sounds reasonable for their roadmaps.

>> No.63688611

>>63688595
>either they are making a 48 and a 64 core epic with 8 dies, or they are dropping in another 2-4 cores a ccx
Or they're adding another CCX per chip.

>> No.63688617

>>63688576
we got trump or hillary, I buy amd but voted trump because I wanted to jam my thumb in someones eye, and so far, trump is doing everything I wanted him to, I will take the good with the bad and so far, laughing at progressives is far outweighing the internet/tax shit for me.

>> No.63688625

>>63688611
possibly, I honestly see the 2 more per ccx to be a potentially better option then another 4 core segment to hop between.

>> No.63688636

>>63671771
You niggers upgrade every 6 months and you fucking know it.

>> No.63688637

too bad most am4 motherboards are complete garbage

>> No.63688672

>>63671771
I was in same position with a 4670k didnt NEED to upgrade, I just wanted a few new features so I upgraded for the fuck of it. Would have just Z97 board if the ones left over wern't shit and inflated in price (I was on z87). Ah well.

Had no reason to desu.

>> No.63688710

>>63688625
Of course that would be better if possible, I just don't think the local fabric scales.

I think the reason for the CCX is that they're able to use a simple fabric (like a shared bus or a crossbar) between the cores in it, and there seems to be an almost universal limit on the market of four cores sharing such a simple fabric. That four-core limit holds true for everything from ARM, to AMD's cat cores, to Silvermont cores, and seems to hold somewhat true also for the K8 chips and Core Quad chips (up until Intel implemented their ring bus, a more advanced fabric). It might also be the reason why there don't seem to be Bulldozer-derived designs with more than 4 modules.

>> No.63688833

>>63675664
Why wouldn't it be possible? Worst-case scenario they have to increase TDP. That wouldn't be a problem either, but they'll probably still use their mayojizz, resulting in temperature monster. Intel have already made 8C ringbus CPUs though and they worked fine, why would 8C Coffee Lake be such an impossibility?

>> No.63689321

>>63688617
I would literally die if I was at the "mercy" of USA's healthcare, under Trump. Fucking lucky to live in EU.

>> No.63689343
File: 139 KB, 1092x614, Intel-roadmapa-2018-coffee-lake-cascade-lake-X-mydrivers-videocardz.jpg [View same] [iqdb] [saucenao] [google] [report]
63689343

>>63688833
It's not impossibility, it would just need lower clocks to stay in the 95W TDP.

And mainly, it would need a new chip being designed way back, which it seems they didn't do >>63686807 so no octocore in 2018.

>> No.63689382

>>63689343


...but /g/ told me Intel were sitting on warehouses full of The Next Best Thing™ ready to counter AMD!

>> No.63689396

>>63689321
well lets see here.
I had an infection
Pretty much resigned myself to death when any way to treat it would cost to damn much to be worth while
Then I found out draining said infection could be done fairly cheap, so went and did it, and I found out about my states free health care.

basicly, I get the doctor I want
I get all my shit done for free
My medicine is covered by the state
And I'm completely reimbursed for everything that was out of pocket till I got on the health care

i'm living in a state that has one of the worst governors america ever had (at least worst real governor, not a joke candidate) and trump is on office.

Honestly, I fail to see how the eu's health care is any better than mine, in fact I can only see how it's worse because without money, I am able to get your equal, but with money I am able to pay for better, whereas the eu will force patients to die rather than try to get treatment, or delay treatment to the point people do die in the case of cancer.

Neithers fucking perfect but it's really hard to say one is better than the other.

>> No.63689421

>>63687791
External storage, external monitors, external USB hubs, external anything really, that's the beauty of it.

>> No.63689428

>>63677449
in same boat
people saying you wont upgrade is stupid
its a first gen, the first years can see massive gains, ill sell my 1700 second hand and get a real nice upgrade for cheap
feels good

>> No.63689473

>>63689396
>whereas the eu will force patients to die rather than try to get treatment
What bullshit is this dude? And I'm not even in the western rich part, Im in the poor part that was ruined by soviets.

In my case it was cancer BTW, neurosurgery that took 10 hours, two months in hospital, rehab needed. Free, including 8 MRI scans the first year and semi-annual monitoring scan ever since.

I was quite young when it happened and would probably be running around without insurance if I lived in USA, at the time when it got diagnosed. Bonus points: it was getting critical, the surgery was needed immediately.

I guess I would have the option to get some fundraiser on the internet and trying to borrow money from friends, but damn, just the MRI scans cost so much, not to begin with the hospital care and the actual surgery.

>> No.63689521

>>63689396

> whereas the eu will force patients to die rather than try to get treatment, or delay treatment to the point people do die in the case of cancer.

So how many times a day do you check fox news?

>> No.63689525

>>63689473
Like I said, my state covered literally everything.

also, as for force you to die, what was the kid recently who the parents wanted to take to america for treatment but they eu refused to allow it and forced the kid to die? I mean sure it was a low survival chance no matter what, but after all the court cases and shit, the chance was gone completely.

>> No.63689545

>>63689525

You do realise healthcare regulations aren't mandated by the EU right? Each country has their own systems and policies. You are a retarded burger who has no idea what the outside world is like.

>> No.63689574

>>63689545
yes, the retarded burger doesn't know the health care perfectly for a region of the world they will never live in, just like the eurocuck only knows the horror stories from america.

we both got our problems.

>> No.63689576

>>63689525
Post links, sounds like you are mixing up things in any case.
Also, since this is about healthcare and payments, those are generally covered in your state only.

>> No.63689672

>>63689576
http://www.cnn.com/2017/07/28/health/charlie-gard-death/index.html
Funny how this was link one for a google search of 'kid forced to die eu'
My understanding was the kid had a death sentence in europe, but in america there was a small chance of survival and the state forced the kid to remain in the hospital, by the time anything was through court, that very small hope of a survival window was shut.

The person with the brain cancer was mostly right, in europe if you need the care now, you get it now, but let's say you have a slower cancer or have more time, you are comparatively more fucked then if you are in america. Personally, and this is a method people use, if you want to live and have something life threatening, assault a cop, you go to jail and they are forced to keep you alive, I also don't think you are billed at the end. there are several higher profile cases with cancer where people purposefully go to jail to get treated. Its shitty, but you are never left to die in america there are always options, it usually comes down to people wont take them/don't know about them. my case, I could have had health care for the last 12 years, but had no idea there were programs for me.

>> No.63689771

>>63686553
The Ryzen Envy X360 has similar battery life to the last gen intel one. Except the Ryzen Envy is configured for 25W (mobiel XFR is only enabled at 25W). And that's with worse components like wifi and screen.

But Ryzen mobile does seem to have a problem with video playback using a shitton of power, even more than 100W dedicated GPUs do for basic video playback

>> No.63689790

>>63689672
>The case became controversial because the medical team and parents disagreed about whether experimental treatment was in the best interests of the child
https://en.wikipedia.org/wiki/Charlie_Gard_case
Based on wikipedia, it seems this was not about heathcare insurrance/payments at all. More like the experimental care attempt on the infant was not permitted due to some medical regulations.

>> No.63690098

>>63689790
Im jumpinh in here
But its Still retarded. We'll have to let the baby suffer and die because muh regulations.

>> No.63690160

>>63690098
I think the problem was probably that somebody else thought, "no, this will make him suffer more" (not saying they were right, but they at least didn'T realize they are wrong). The problem when you have a clash of good intentions.

>> No.63690176

>>63690160
>the road to hell is paved with good intentions

>> No.63690615

>zen 2 releases
>still worse than my 8400 in games
I will laugh

>> No.63690710

>>63690615
>>still worse than my 8400 in games
amd is already at or above intel in ipc, the only thing holding intel up is they can clock higher

zen 2 is 7nm on a process that is planning to hit a optimal base of 5ghz, remember, the current process is 3ghz for its optimal base

unless intel hits 5ghz on 10nm without any issue at all, its trouble for intel, oh, and fun fact, even intel is throwing away the first 2 iterations of 10nm will be worse than current 14nm.

zen+ is up in the air, but it looks to be a custom process that aims to address the clock speed issue on desktop to make it more inline with current norms with 4ghz bases.

>> No.63690764

>>63690710

>planning to hit a optimal base of 5ghz

More commonly known as Kike Killing Klocks. If zen+ or zen 2 hit 4.5ghz base on the 1800x replacement Intel's lead will basically vanish.

>> No.63690800

>>63690710
What exactly dictates how much a CPU can overclock? What is it with intels that let them reach higher clock speeds than ryzen? I would love to learn.

>> No.63690834

>>63690764
honestly, I don't expect much from zen+ as it looks like its entirely a process shift, not interesting as an upgrade at least

zen2 however, that is what has me interested, because they won't be able to fuck with the chip much from the server version, so we will get a 12-16 core cpu for mainstream, just that it will have the same pcie and ram capabilities, all while the cpu clocks 5ghz or ballpark this on air.

this will likely be my first cpu upgrade I ever do, as my other systems have all been full system upgrades.

>>63690800
>What exactly dictates how much a CPU can overclock? What is it with intels that let them reach higher clock speeds than ryzen? I would love to learn.

If I remember right, it was larger fins, there is something that can be done in design that allows for higher clocks or using less voltage, but its majority done through process anymore.

>> No.63690836

>>63690800
Process and pipeline length.

Intel has, at the moment a massive advantage in the process, the pipeline length between AMD and Intel should be similar.

>> No.63690862

>>63690836
>>63690834
I assume ryzen+ will address process right?

>> No.63690868

>>63690862
Seems so.

>> No.63690896

>>63690862
ryzen+ seems to be a custom process that is geared toward desktop only, as if they made a server sku of it would they sell any server parts with people knowing better is literally a month or two away?

anyway, it seems to just be something that will enable marginally higher clocks
intel beating clocks, probably not, but closing that gap enough to toss in the air if you want a 230$ 6 core 12 thread or a 300$ 6 core, fuck yea.

>> No.63690917

>>63690896
Does 4.4-4.5GHz sound unreasonable with a custom loop based on what we know?

>> No.63690937

>>63690896
>custom process that is geared toward desktop only,
12nm LP is also made for Automotive and RF, but it could be a custom one for AMD as well, since RF also requires high clocks.

>> No.63690967

>>63690917
based on what we know 4.4 or 4.5 if the voltage does not need to be cranked passed 1.4 would be doable on air.
fuck, the shitty versions of the air cooler they have can push the 1600 to 3.9~ and that wasn't a heat issue, that was a cpu stability and not wanting to push it to 1.4 volts issue.
a custom loop should shit on the air coolers amd gives you.

That said, I have a 1700 and noctua d15, so I can't give you first hand, just that I have never seen the fans spin up even when I prime 95'ed it,

>> No.63691339

I wanna upgrade CPU but ram prices are killing my tech boner. Would the jump from a 4690k to 4790k be worth it? New prices are kinda asinine for an older CPU.

>> No.63691370

>>63691339
No.

>> No.63691469

>>63691339
Look whether you have a chance of selling your DDR3 for decent money and then buying a DDR4 kit used also for decent money. probably the only way, but it will still be expensive.

>> No.63691975

>>63691339
I have some lightly used (spent like one week before upgrading Black Friday) 8GB(2x4GB) [email protected] I'll sell to you.

>> No.63692109

>>63691975
That's great anon but I'm broke atm.

>> No.63692197

>>63690160
>the road to hell is paved with good intentions
>>63691975
Why would you buy 8gb?
Do you know its the current year. Why sell it after a week?

>> No.63692266

>>63692197
Because I don't really need it and could do some more Christmas shopping.
I found 8GB adequate for my purposes before I got greedier on VM usage.

>> No.63692674

>>63692197
Fuck. People sell their used RAM at least 40% off. Given how easy it is to replace I think I'll just do that for the rest of my builds.

>> No.63692816

>>63689396
>voted trump
>uses state-funded welfare for healthcare

Why am I not surprised?

>> No.63692960

>>63692816
Did I have another option? I wanted bernie and I was not voting a war mongering lich who promised to provoke ww3 or at the very least another cold war into office.

>>
Name (leave empty)
Comment (leave empty)
Name
E-mail
Subject
Comment
Password [?]Password used for file deletion.
Captcha
Action