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

Maintenance is complete! We got more disk space.
Become a Patron!

/g/ - Technology

View post   

[ Toggle deleted replies ]
File: 77 KB, 1140x760, radeon-vega-frontier2-1140x760.jpg [View same] [iqdb] [saucenao] [google] [report]
60500707 No.60500707 [Reply] [Original] [archived.moe] [rbt]


Wait ©

It's a taboo but...Vega ''1.0'' is DOA because AMD did not expect Nvidia release the 1080ti. Also all >1070 Pascal cards are faster than expected.

HBM was a mistake, customer GPU should never use this technology.

>> No.60500732

i wish they would work on their linux drivers instead

>> No.60500776

>AMD did not expect Nvidia to do exactly what they did last time.

Sounds like AMD is dumb.

>> No.60500779

All still depends on price.
If they manage to make cheapest vega 400$ and still beat at least 100 then it is gonna be pretty gud.

>> No.60500780

I hear the open source drivers are pretty good these days. Never tried an AMD graphics card though, so I can't speak from experience.

>> No.60500793



>> No.60500809
File: 156 KB, 633x758, amdvega.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.60500815

That's a great source.

And refreshing their lineup with higher clocked versions once fabs get better is certainly something new and not something that's been going on for 15 years already.

>> No.60500825
File: 179 KB, 200x200, madoka-7.gif [View same] [iqdb] [saucenao] [google] [report]

>Vega ''1.0'' is DOA
Please stop it. /g/ has been overrun by AYYYMD and INTEL BTFO posts for so long now it isn't even fun to shitpost about anymore.

>> No.60500826

Just Wait™ for eternity.

I'm glad at least Ryzen is pretty neat

>> No.60500829

>2 8-pins



>> No.60500837

They thought Pascal was a rebranded Maxwell without improvement, Meanwhile the 1070 is faster than a OC Titan X (maxwell)

>> No.60500907

>still no """"""""modern"""""""" (read: card that doesn't use 300W under load) AMD card better than a 290x

>> No.60500918
File: 1.07 MB, 1228x1502, 1494487855012.png [View same] [iqdb] [saucenao] [google] [report]


>> No.60500945

i always had horrible experiences with them, starting with cards where the windows drivers had gpu accelerated video decoding, but the linux drivers didn't (for the same card), onboard gpus were only supported in legacy drivers, etc

and the open source drivers were useless a few years ago, that was the reason why everyone says go nvidia or intel if you run linux

>> No.60500948

totally agreed - at this point my mental filter just drops any post asserting feeling or merit unless there are numbers or links

the opensource AMD drivers are quite good now, rx460 using mesa 17.0.4 can run pretty much anything I want in 1080. If I wanted more than 1080 or more than the current linux software selection I would change those attributes of my system.

>> No.60500951
File: 183 KB, 1500x938, flag-sweden-XL.jpg [View same] [iqdb] [saucenao] [google] [report]

AMD is doomed, pic semi related

>> No.60500960

They were correct that far. Pascal is essentially rebranded Maxwell. Only it clocks to almost double the megahurtz.

>> No.60500972

>Tfw Vega 2 is a rebrand of Vega 1, which is a rebrand of a 390x, which is a rebrand of a 290x, which is a rebrand of a 7970

>> No.60500977
File: 48 KB, 600x363, 499d9080bc6b3ddd07ed7ffa941245a2.jpg [View same] [iqdb] [saucenao] [google] [report]

>HBM was a mistake, customer GPU should never use this technology.

You must be retarded, you will sound just like pic related in a few years.

>> No.60501021

>gpu accelerated video decoding
I have to admit I've never understood the need for this. Even my i3 370M laptop from 2010 has no problems with decoding anything I've thrown at it in software.

>> No.60501027

I guess you never heard of a Fury or Nano card

>> No.60501036

>AMD did not expect Nvidia release the 1080ti
come on, everyone saw that coming

>> No.60501051
File: 191 KB, 715x1200, madoka-8.jpg [View same] [iqdb] [saucenao] [google] [report]

>the opensource AMD drivers are quite good now, rx460 using mesa 17.0.4 can run pretty much anything I want in 1080. If I wanted more than 1080 or more than the current linux software selection I would change those attributes of my system.
Sounds nice, senpai. Maybe it's time for me to finally get an AMD card and ditch the nVidia binary blobs.

>> No.60501053
File: 186 KB, 1562x855, AMD-VEGA-10-VEGA20-VEGA-11-NAVI-roadmap.jpg [View same] [iqdb] [saucenao] [google] [report]

How do you niggers not know about Vega 20? It's the Fp64 card

>> No.60501067

So you only accept cards that are complete rewrites and have nothing in common with the previous generation?

>> No.60501074

/g/ is full of shitposting tech illiterate /v/edditors. What did you expect?

>> No.60501075

7970 GHz

>> No.60501082
File: 124 KB, 640x960, op-10.jpg [View same] [iqdb] [saucenao] [google] [report]

Why are you taking OP's baseless accusations seriously?

>> No.60501091

Wccftech is a good source

>> No.60501092

But Maxwell and Pascal are rebrands?

>> No.60501096

modern as in current gen. rx580 is AMD's best current card, and it's barely better than a 290x

>> No.60501106

I guess that means you read the article which mentions zero words about the 1080ti and not just assumed that OP didn't just pull that assertion from his ass?

>> No.60501108

Okay? But its more than enough for 1080p AAA gaming, and its not the $599 or whatever launch price of the 290x

>> No.60501112

RX 580 is a RX 480 with a higher voltage tho

>> No.60501128

>getting so btfo you have to drop the _90
heh amd whatre you doing

>> No.60501133

>good enough for 1080p
fact is, if you want to do 1080p+144hz or 1440p or higher, you HAVE to go Nvidia

my point remains

>> No.60501142


>> No.60501146


>> No.60501147

maxwell isn't, pascal is a refresh

>> No.60501171

>2016 this time
>faster than 980ti, AMD gonna totally BTFO nvidia with polaris
>polaris is decent but nothing ground brekaing in performance

>vega is doa, wait for navi, slower than 1070, not coming until q4

>> No.60501180

Or you can just lower your settings slightly and do just fine with a much cheaper free-sync display

>> No.60501197
File: 7 KB, 570x520, jew.png [View same] [iqdb] [saucenao] [google] [report]

>get a old 480 from a good AIB with decent vrm on the pcb
>update the bios
>save money or sell it as a 580

>> No.60501198

Or just get nvidia and enjoy max settings at 1440p 144hz

>> No.60501201

I use NVIDIA but shit does AMD make some good looking cards. Simple clean designs that don't look like it's targeted for a l33t gaymer. I don't even know what that cube on the corner is for but it's nice.

>> No.60501221

That's 799 for the GPU and 600 for the display

With an RX 580 at 250 and 399 for the pixio 1440p 144hz IPS free-sync display which are more people going to buy?

>> No.60501231

It's a tachometer to measure GPU % utilization and it's useful when you're dealing with a few hundred of these in racks and don't know which one is bricked, that one won't light up.

>> No.60501259

Good looking and functional. Even better.

>> No.60501267

Who cares about poorfags

>> No.60501275

Well apparently everyone who recommends a Pentium, i5 or R5

>> No.60501282


>> No.60501297

You mean 80% of the market?

Sadly they still buy the inferior 1060 vs the 480/580

>> No.60501330

One of these is not like the other.

>> No.60501335

380x is a 285 that isn't cut down. The other three are the same chip.

>> No.60501340

7970 is a different chip

>> No.60501341

> Anonymous 05/20/17(Sat)19:52:58 No.60501297▶
>You mean 80% of the market?
>Sadly they still buy the inferior 1060 vs the 480/580
My friend even bought 1050ti claiming it is faster than 8gb 480
I tried to show him charts,tell him spec difference,he was just like "i do not care,amd is shit"

>> No.60501347

>lower the settings, then you only need a shittier gpu!

>> No.60501363

Well, not like you can have shadows enabled in Nvidia GPUs if you want any usable performance, so you're lowering settings anyway

>> No.60501364

-edit the files and run every game on integrated graphics

>do not play games they are for kids,you do not need gpu just let cpu to render fonts.

>> No.60501367

Can you point out the differences between High,Very High and Ultra? Hell just disable tessellation if its an option and you're good to go

>> No.60501376

Maybe you can not on your 1350x600 thinkpad display.

I and rest of 4k master race can see it very clearly.

>> No.60501388

You clearly should stop shitposting.

>> No.60501394
File: 182 KB, 404x266, 1495288739635.png [View same] [iqdb] [saucenao] [google] [report]


>> No.60501399
File: 42 KB, 653x726, amd.png [View same] [iqdb] [saucenao] [google] [report]


>> No.60501413
File: 116 KB, 536x640, 1429384366718.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.60501414

I can definitely tell the difference between AA and no AA. that alone requires better gpu

>> No.60501457

i have two of those.. i finally don't have driver issues. its now pretty much an obsolete card.
>tfw never buying amd cards again

>> No.60501469

But do you really need X16 vs X4? Or if AA is your worries you'd want to be at 4K where the free-sync displays are half the price of G-sync last I checked.

And isn't there usually a well optimized AA option in a given game that doest impact performance much?

>> No.60501480

I never had a single driver issue on my 7850, and the 7970 is still solid for 1080p gaming

>> No.60501548

if I have the money, I want the best specs

1080p 60fps, or 1440p 60fps medium specs, is all AMD can do

>> No.60501552

>AMD has a product roadmap
Uh, no shit? Is this really the level we're at now?
Nvidia has post-Volta projects in the works, is Volta DOA?

>> No.60501572

of course
because they're developing a future line of products, that means the next line automatically sucks

its /g/ logic

>> No.60501575


w r o n g

>> No.60501588

Glad I do not have the money,it is funny thing about having no money,you can pick from many variants.

>> No.60501608

Ha haaaaaaa.

>> No.60501614

that's not at all what the arcitcle says

>It used to be that every two years we got a full lineup of GeForce and Radeon graphics cards based on an improved architecture
whoever wrote this must be 12 years old.

>> No.60501673

Maybe or maybe you should stop livin' in the past.

>> No.60501788

>woooww, a X80 card is barely better than a older X90 one

>> No.60501794


Let me know when Nvidia stops selling shit tier DX12 cards.

>> No.60501799
File: 18 KB, 576x422, 1493306537112.png [View same] [iqdb] [saucenao] [google] [report]

still waiting

>> No.60501807

spray salt water inside his computer please

>> No.60501814

why do people keep posting this clickbaiting shit.
vega 20 was supposed to be 7nm with 1/2 DP capabilities.
they're probably skipping it in favor of navi coming sooner
I think 14nm+ is what's already used in polaris 20, I don't see a reason why they wouldn't use the same shit with vega. Unless first batch of vega is on 14nm (Frontier Edition) and rx vega will be on 14nm+.

>> No.60501836

that's the thing, I can afford 1440p 144hz freesync, but not gysync. but there's no AMD card that can run it anyway so it doesn't even matter

>> No.60501854

It's not the same 14nm+, that's still LPP what's used in Polaris 20 and Vega, this 14nm+ is LPU most likely

>> No.60501877


It is new when that's what AMD hinges their entire upgrade path on until the next new architecture.

Rather then having an unlocked card from the onset, they lock the voltage down and cripple the clock speeds. Then try to pimp out the new cards that are basically the old cards ungimped as if It's some sort of revelation.

And then Koduri has the balls to call it Fine Wine™ technology.

>> No.60501895

Well the salt he put in my wounds by bullshit that comes from his mouth is already enought.

I was so salty that I wanted to punch him then I was like "well he is wasting money not me"
He also bought old-gen intel instead of new ryzen
again with same "Amd is shit" argument
He would have no kidding 30% faster pc with more cores for same money,even save for SSD because he bought Ultra gaymen motherboard for 140$ when I asked why he needs it while he does not overclock he was like "well it say game so it is good"

>> No.60501909

Before someone claims he is kid or something,he is 26 years old.

>> No.60501910

I think 14nm LPU is a samsung exclusive.
GlobalFoundries only licensed 14nm LPP as far as I can tell from their roadmaps.

>> No.60501939

What the fuck are you talking about you idiot? Your clockspeed limit is the fab's limit, clockspeed naturally increases with improvements to the fab.
AMD's WSA allows them a different supplier, both Glofo and Samsung fabs are pretty compatible, at least for 14nm

>> No.60501945

It's best not to argue with retards. They only bring you down to their level. Never offer up computer advice again.

>> No.60501998

>AMD's WSA allows them a different supplier
sure but both vega and zen are shown with 14nm+
I don't think they'd put off that much of their silicon production outside of global foundries

>> No.60502009


Fucking apologist faggots like you making excuses are hilarious.

>Over at TechPowerUp, forum user TonybonJoby download the BIOS from a Sapphire RX 580 Limited Edition card clocked at 1,411MHz (boost) and flashed his XFX Radeon RX 480 card with it using ATIFlash. He now has a card that is recognized as a Radeon RX 580 with the faster clockspeeds.

They launch this shit voltage locked.

>> No.60502101

It has just as much chance of exploding at that speed as it did if it were just OCed.

>> No.60502155

That's just desktop and vega, Naples and Ryzen mobile(that don't get the 14nm+ treatment) will still be at GloFo.

Besides, the gap between 14nm+ and 7nm will be pretty short, the LPU rumors have been going on for a while.

>> No.60502205


It is overclocked, dumb ass, from the factory where they unlock the voltage and increase the clockspeed.

Why the fuck do you think the pre-launch engineering samples had 480s hitting 1500mhz? Why the fuck do you think there is no more overclocking room on the 500s now that they hit 1500mhz clockspeeds?

Fuck, you apologist retards are getting cucked by Radeon. And you're just asking for more.

>> No.60502226

>the 580 draw more power
>improvements to the fab

>> No.60502246

>the gap between 14nm+ and 7nm will be pretty short

Quite bolt claim,considering how huge was 24->14 change

I'd expect +35% boost in performance.

>> No.60502265

Well It is friend after all,he asked me for build advice,i deliver whole build sheet with games he can max,then tells me he asked 1 random person that build bicycles for living [while i do programming and plc building for army]
And he went with bicycle builder's advice that matched his "nvidia,intel=best 4ever" opinion

Like I say tho,It is his money so fuck him.
He also uses 8years old display and say things like "I do not need anything better because it still works"

So he wanted to max his games and get good fps on monitor that can not display any of it.

He is not even poor,middle class.

>> No.60502281

Not a bold claim, it's in the roadmaps, GloFo's 7nm DUV production will be end of 2018, EUV following 2 quarters latter.

>> No.60502290


Just wait for Vega when Volta launches and Navi is six months behind schedule. Again. And suddenly we have higher voltage Vegas being sold with higher clock speeds and no overclocking room, again.

Vega is a disappointment. And Raja knows it.

>> No.60502321


>> No.60502337


Is still shit compared to HBM.

>> No.60502391

oh boy...

>> No.60502432

Nigger, HBM2 can go up to 1TB/s bandwidth, uses 3 times less power, uses 5 times less PCB space, the memory controller is half of a 384bit GDDR* controller and has 3 times faster tFAW .

The only thing GDDR has going for it is that it's cheaper, and that cheaper doesn't mean shit for $600+ GPUs

>> No.60502462

>HBM2 1.4 Gbps
>720 GB/s

>GDDR6 @ 384-bit
>up to 768 GB/s

>> No.60502478

>it's cheaper

Thats his point ''HBM was a mistake, customer GPU should never use this technology.''

>> No.60502499



Faster, more power efficient, less space needed.

Even the budget HBM is superior in every way.

>> No.60502512

>this new tech is too expensive DONT PUT IT IN CONSUMER PRODUCTS

So why did we even get GDDR5, surely it too was once expensive so it should've never been made available.

>> No.60502514

So better technology shouldn't be used because it's more expensive? Do you care about the $30-40 cost difference between HBM2 and GDDR* on a $600+ GPU?

Keep your GDDR to poorfag cards under $400, the rest should use newer and better technologies.

>> No.60502521


HBM3 is how much faster now?

>> No.60502522



>> No.60502528



(in before golden sample sent to reviewers)

>> No.60502543


Bent over faggots get so mad.

>> No.60502553

>let's use a 4TB pcie ssd for my chinese cartoons and porn

>> No.60502581

Mad about what? Only Quadro or Tesla should use HBM, customer grade GPUs does not need it

>> No.60502583

Yes, yes, great hyperbole, enjoy your ISA slot, who needs more i2b bandwidth anyway

>> No.60502595

This guy gets it, I store everything on 2008 era 5400 HDDs, everything else just isn't worth it. Literally don't need anything more. Fucking retards "MUH new tech!!!"

>> No.60502601


>> No.60502607


They overclocked the memory you fucking retard. The silicon is still at 1500mhz max boost.

>> No.60502642

Jesus christ the shitposting is getting unreal.

How the fuck can Vega be a disappointment when it's not even out yet you moron? Where are the independent benchmarks showing its performance?

Jesus at this rate by the time we get to the Computex presentation we'll have idiots claiming that Vega will be a complete failure if it isn't AT LEAST 30% faster than two 1080ti's in SLI and that Volta will be released 3 minutes later and BTFO AMD again LEL.

Can you idiots at least wait until we have some actual DATA on how it performs and how it will be priced relative to that performance before you start shitposting up a storm?

>> No.60502651

Nvidiots getting bent out of shape because they're getting held back with ancient GDDR* revisions. While AMD pioneers the future. Like if you love your GDDRwhatever, you can still get it """new""" in 2019 on your GTX 1330

>> No.60502663



> Samsung has a couple alernative in the works, though. The company unveiled its plans for a mass-market "low cost HBM," a variant of HBM2 with a comparatively lower bandwidth of 200 Gb/s per die, but with a pin speed of approximately 3 Gb/s. (HBM2 tops at 256 Gb/s per die.) Samsung says this memory type will come at a fraction of the price of HBM2, thanks to the removal or reduction of features like ECC, buffer dies, and TSVs (through-silicon vias).


It's like you faggots don't read, and just want confirmation bias.

>> No.60502692


"MUH retro 5400 HDDs!!!"


Do you have CRTs on a lace tablecloth? Box macaroni for dinner again? If you want to live in your time capsule that's fine, but most people just want to do X faster and more efficiently, so that they can fuck off and get to the point

>> No.60502694

>August 23, 2016

We still waiting

>> No.60502705


Because it is six months behind, and it is only at 1080ti levels of performance.


>> No.60502732


Good, glad you can appreciate superior technology.

>> No.60502744

Nvidia does not use HBM on geforce because that's expensive, they are not stupid like Raja.

>> No.60502751

Prove it. Prove to me right now what the performance level of Vega is, with independent benchmarks comparing it to a 1080ti. I'll wait.

>> No.60502757

>wasting money for no reason

Fucking retard

>> No.60502760


They also gimp their memory bus sizes, because they like planned obsolescence.

Your point?

>> No.60502772


You missed the investors meeting?

>> No.60502817

Oh, they did independent benchmarks of Vega in multiple games and compared them to a 1080ti in the same games? I must have missed that during the presentation.

The truth is that no one knows how the highest end gaming Vega card will perform, and we won't know until the NDA is lifted and we get to see a bunch of independent benchmarks.

Until then, it would be nice if people stopped trying to frontrun the launch with an avalanche of shitposting.

>> No.60502825
File: 63 KB, 500x500, 1459445564945.jpg [View same] [iqdb] [saucenao] [google] [report]

>muh hbmeme
Get your GP100 with HBM and go away

>> No.60502830
File: 56 KB, 1276x720, _.jpg [View same] [iqdb] [saucenao] [google] [report]


>no reason

>> No.60502856

>I feel so bad for AMD. It put all the work into HBM and launched Fiji, a product that really didn't benefit from extra memory bandwidth at all.

>Nvidia and Samsung are now running with the investment and AMD still have nothing to show for it other than a whole bunch of invested man-hours that didn't provide an ROI.


>> No.60502867

what a qt

>> No.60502874

she is cute, what her name

>> No.60502889


>I don't know how to extrapolate data because I am scrambling to prove a point

The post.

Six month head start, and it doesn't look promising for them to merely catch up.

We can't even claim immature drivers at this point with Vega taking up 80% of the driver team. So you aren't going to see huge gains in performance six months later.

When Volta launches, they will increase the voltage, overclock the chip from the factory and sell it as a whole new product.

And try to convince you it's "Fine Wine". When all I see is Raja pissing in my face thinking I am stupid.

>> No.60502934


As is standard with AMD. Do you think this is the first time they have done something like this?

Why the fuck do you think they have so many damn fan boys?

>> No.60503155

>it doesn't look promising
Source: My ass

>When Volta launches, they will increase the voltage, overclock the chip from the factory and sell it as a whole new product.
Source: My ass

>And try to convince you it's "Fine Wine". When all I see is Raja pissing in my face thinking I am stupid.
Source: My ass

>> No.60503203

>Sadly they still buy the inferior 1060 vs the 480/580

I did this, and couldn't be happier.

>> No.60503653


If you missed the investors meeting, it's not my fault.

Just like they did with the 290 to 390. 285 to 380. And the 400 series being revranded as the 500. How many times do you need to be played for a sucker?

How often does Raja mention fine wine at Capsaicin events? How often does he mention it at the share holders meeting? He has no idea why it exists, other wise he wouldn't be artificially locking down your GPUs voltage and limiting your clock speeds at a time when AMD desperately needs performance wins.

The only thing you should be talking about coming out of your ass is my fucking foot.

>> No.60503728

>Source: My ass

>> No.60503744
File: 19 KB, 261x273, IMG_5991.jpg [View same] [iqdb] [saucenao] [google] [report]

Oh look, another post based entirely on assumptions sourced from your ass, with zero actual evidence.

Here's your (You).

>> No.60503805


Run that damage control. Let Raja cuck you harder.

>> No.60503856

Where did I assert or imply that Vega would be good?

I have no idea how Vega will perform relative to Nvidia's current offerings: It might be terrible, it might be great, but we won't know until there is some real independent evidence.

>> No.60504016


Does his taint taste like curry?

>> No.60504029
File: 16 KB, 250x680, 1256594250772.jpg [View same] [iqdb] [saucenao] [google] [report]

It's going to be barely above 1070 with a similar price. screencap this

>> No.60504643

7970 had 2048cu IIRC

290X had 2560 (or 2816 I forget)

Now those were similar but 290X was GCN 2.0 or 1.1 I forget. While similar, not 100% identical to HD 7000

390X is a cut down 290X

You missed Fury 1 which was the full GCN design of 4096cu but it had HBM so it was somewhat new.

Vega is a different cu altogether from GCN. It might have components, but it's biggest architectural change for Radeon since the...


>> No.60504797

Just wait, Vege in Q2.
Oh wait, it is only pro vega. Well, maybe Q3. Or Q4, who knows.
Anyway, just wait!

It can't be later than consumer volta, can it?

>> No.60504832


This is about 90% wrong

>> No.60504849

It is going to be above 1070 with price that one would expect
Screencap this.

>> No.60504862

So when is consumer Vega coming out? Wondering whether to get a 580 or wait for it if like 300 to 400 option

>> No.60504901


Get a second hand 290x for half the price nub

>> No.60504922

GTX 1070 barely got over 5000 and a Fury X managed 5000 at Firestrike Ultra

We're talking OC here

So what makes some of you guys think a Vega would do only the same as a GTX 1070, or rather Fury X from 2 years ago?

Oooh OK so you were talking out of your asses? Okaaay

>> No.60505049

Me like talking out of my ass. Vega sucks buy 1080ti.

>> No.60505302


Ten more days it will be unveiled. Release date, price, and preliminary AMD sources benches.

>> No.60505337

Their reluctance to talk about vega. Likely due to memory issues

>> No.60505396

Asses are like mouths, but it doesn't sound like it when I'd agree

1080 Ti managed 8000 at Firestrike Ultra. That's 60% from Fury X and 1070. Still costs $700 though.

Based on Tflops alone Vega should be 45% faster than Fury X, (bandwidth is relatively unchanged) but it's more complicated than just that.

>> No.60505720



>> No.60506713

>customer GPU should never use this technology.

been saying this for 3 years. NIce to see /g/ finally catch on. Retards.

>> No.60506778

well technically speaking, pascal is literally the same as maxwell, except the process node shrunk, they moved to FinFet, and they drastically boosted clock speed. No architectural changes were made.

>> No.60507114


You would be wrong, but I'm not surprised. HBCC is pretty new.

>> No.60507435

Amd expected Volta to be another fermi, the only way amd would win if nvidia fail.

>> No.60507476

Why is it in Sweden colors?

>> No.60507516
File: 158 KB, 1920x1080, maxresdefault.jpg [View same] [iqdb] [saucenao] [google] [report]

I heard you've been talking shit about AMD?

>> No.60508552

lol.... they are, hugely

>> No.60508603

topkek nvidia lowers the settings for you at the driver level, enjoy ur noshadows lol

at least with AMD you get the full graphics rather than cutdown crap to artificially boost fps in benchmarks

>> No.60508612

You're a fucking retard holy shit

>> No.60508627

>nvidiafags this desperate

Or maybe an idiot who doesn't understand how process revisions work, but probably a desperate nvidia shill.

>> No.60508693

>upgradable via the bios
>more powerdraw
>Or maybe an idiot who doesn't understand how process (((revisions))) work

480 = 580

>> No.60508754

Pascal is a rebranded Maxwell and Volta is a rebranded Pascal though Maxwell was just Kepler with tiling rasterizers and Kepler was just a cut down Fermi.

>> No.60508766

the ROTR demo with 2GB of HBM2 already pits it above the 1080

>> No.60508803
File: 60 KB, 501x750, 84a04121219bd274ea24ab6628d221a9.jpg [View same] [iqdb] [saucenao] [google] [report]

AMD has "professionals" making Binary Blob proprietary drivers for their graphics cards

and a bunch of hippies come along and make free software drivers that are now faster and superior:


The GNU/Linux MESA driver stack is now _faster_ in most games as of MESA 17.2.

Windows users are BTFO because they don't have free software drivers for AMD cards. All we need is slightly more games in Steam and more Windows ransomware like WannaCry and the nations will be forced to accept the GNU World Order.

>> No.60508817

and then no money will be circulating because nobody will be getting any work done because Linux is a broken mess for people who have too much spare time to spend fixing it

>> No.60508818



Raja Kodury is fucking awful at his job.

>> No.60508844

raja's job description has no mention of linux anywhere

>> No.60508850

Linux is used as a daily driver for of the biggest and highest-productivity companies in the world.

>> No.60508864

ahahahahahaha sorry but mac os is not linux and bsd is the only real server OS out there

>> No.60508886


>> No.60508895

> t. bait

mac os is a timebomb that keeps getting shittier and more locked down with every version. Next they are going to ram down their own junk file system just to fuck people over/

>> No.60508914

yeah let's just compare how many people use mac os at microsoft, apple and google compared to ubuntu or some other heap of shit linux garbage

>> No.60508935

Lots of people at Google use Linux. Pretty much all their core engineers. They even have their own fork of Ubuntu they call Goobuntu, because of how widespread it is used. Facebook and Oracle are also big users of desktop Linux, along with Amazon and several other major companies.

Apple uses mac. Microsoft uses mac. Big surprise!

>> No.60508948

>Big surprise!
yeah it is pretty unsurprising that big, respectable companies with engineers who aren't just apple/ms rejects as employees use posix compliant operating system

>> No.60508955


I hope this is b8.

>> No.60508987

As usual it's going to be slower than Nvidia in 8/10 games, end up costing more globally, lack driver support for indie titles and will have twice the power consumption. This has been amd for the last 5 years. Iirc the last time amd was cheaper was the 7950.

Competing with volta in ~6/8 months will be another mission altogether too. I didn't even expect volta to be unveiled till late 2018.

>> No.60509019

Volta is literally Pascal with added ASIC cores which won't be in the gaming GPUs anyway.

>> No.60509050


Where are the Pascal cards at 2GHz? If you expect a manufacturer to only sell cards at the maximum clocks that can ever be achieved, then why not the other one, you fucking shill?

>> No.60509062

>artificially locking down your GPUs voltage and limiting your clock speeds
This is literally what NVIDIA do and why there is the misconception that their cards overclock better.

>> No.60509069

volta is a rebrand too, dont forget

just more shit packed in it

>> No.60509137




>> No.60509147

sorry what is this meant to prove and why should i watch through this morons video? he doesn't demonstrate any knowledge of the hardware on a technical level

>> No.60509181

>nvidia lowers the settings for you at the driver level, enjoy ur noshadows

>> No.60509274


>Waaaaaaah! It's not fair! How can you use basic information and call me out on my

>> No.60509276

>basic information
>does nothing to disprove anything i said

>> No.60509279

Not him but Nvidia doesn't have to sell a rebrand at higher clocks because of gpu boost. Most, if not all 1060's, 1070's, 1080's boost to at least 1900 mhz assuming they're not throttling in any way. Some custom models even reach over 2ghz with only like 100mhz factory oc and gpu boost. We already know pretty much every 480 could oc to 1350 mhz even with the stock cooler and power set up. Instead of doing stupid things like sponsoring the cancer that is Bethesda they should be investing in relevant gpu tecgnologies such as gpu boost. If a 480 is competing with a 1060 or even a bit slower with only 1266 mhz when the 1060 is boosting to 1700 mhz then imagine what a 480 would do if it had it's own gpu boost to 1300+ mhz. They have had 5 years to develop tech like this yet they still decide to release overclocked rebrands instead which causes the price to inflate once again. The level of incompetency is tragic. The rx 580 is £30 (~$40) more than a 6gb 1060 now.

I say this as someone without bias. I had a 390 and now 1070.

>> No.60509481

>source: my posts on 4chan
I've been saying this for a while. Vega refresh will come in early 2018 to compete with Volta.

yep. Nvidia is very secretive and it works to their advantage.

You have no clue what you're talking about.

>> No.60509670

>the opensource AMD drivers are quite good now, rx460 using mesa 17.0.4 can run pretty much anything I want in 1080
It really depends.

In some games, the Fury X matches the 1080 with best drivers of each.
In others, the Fury X does slightly worse than a 1060.

>> No.60510019

Not anon you replyed to but look at beachmarks of watch dogs

>> No.60511096

Have you ever used a 290x, 390x, or fury? Forget the 480/580, Hawaii is much better at higher resolutions anyway. Those cards can do 1440p60 ultra in most new games (obviously trash performers like PUBG, Wildlands, W3 are exempt) quite easily and can be bought for at or less than the price of a new RX 470.

>> No.60511110

Do you live in Turkey or South America?

>> No.60511135

>They launch this shit voltage locked

>what is Kepler, Maxwell, and Pascal
Hell, Kepler had a hardware-locked maximum voltage. My 680 clocked up to 1293/7000 with a modded BIOS, if I could have gone higher than that I would have likely seen 1350 at least before artifacts start to show

>> No.60511167

Wow. You really embarrassed yourself with that post.

>> No.60511177

>implying they'll have 384-bit memory I/O when even the 1080 ti has less and the 1080 is only 256-bit
>implying nvidia wont meme about minor power savings (cost savings really) and instead equip the next x80 card with 256-bit gddr6, so up to 512GB/s going by that metric

Just because they can do something, doesn't mean they will

>> No.60511208

The kinds of people to say this stuff are not planning on buying Vega or anything AMD, regardless of price, performance, or release date.

We barely know anything about Vega as it is, so wait until Computex before trying to pass off anything you say here as not being shitposting.

>> No.60511427


>company's R&D already working on improvements to product close to release
>somehow this is a surprise to /g/

Next you're gonna tell me Nvidia is waiting to release the whole Pascal line to start working on Volta, right?

>> No.60511457

I don't know why in the fuck you're trusting GloFo's in-house litho development claims but let me tell you you're a moron for doing so.

>> No.60511516

I love ignorant statements like yours. The way people just BTFO themselves while calling someone else a moron.

The reason he's optimistic about GloFo's 7nm is because it's really IBM's and not an inhouse development, retard.

>> No.60511577

As someone mentioned, 7nm glofo will be used for IBMs power chips, and if you know anything you know those clock to the moon.

>> No.60511881


Unlike most of the toys used by anon POWER is all business all the time and it clocks (partly) because anyone buying such a beast doesn't care about power draw - they want ultimate computational performance. It does help that IBM designs POWER to be immensely "wide" (hmm, that delicious 4-way SMT - iirc POWER9 will feature some 8-way SMT options). It is also why bulldozer clocked to the moon as well - it was essentially running off the back of IBM's process and 5ghz is nowhere near what the process was capable of topping out at. Issue is again, cooling a 300w chip on the desktop isn't going to happen.

>> No.60512318

And what the hell do you think AMDGPU-PRO builds on top of?
Correct, amdgpu and Mesa. AMD themselves officially contribute to both of those projects.

>> No.60512844



I am all in prepared to buy Vega. Provided it doesn't fucking disappoint like it always does since Raja took over.

Vega will be another massive disappointment.

The tears are going to be glorious.

>> No.60512889

7nm IBM will be used for Navi
2x6000 RPM San Ace fans will clock bulldozer to 5500mhz with a heatsink.

>> No.60512937

"working on"

Just like they've been "working on" for the last 2 decades?

AMD is done, play for some competent company to buy them off.

>> No.60512953


>> No.60512957

You mean AMD has a product roadmap? WOOOOOOOOOOOOOOOOOOOOOOOOW.

>> No.60512984

Remember, Raja joined AMD after GCN had already gone through its first release, so years after it entered the engineering stages.

Vega could surely have been mostly his idea, if not entirely. However, due to a limited R&D budget and diminishing graphics market share, he was likely forced to make do with what they already had most immediately available.

Designing a brand new architecture, especially from the ground up, is a massive undertaking and extremely costly. This is why Nvidia has not yet done a full redesign since pretty much Fermi, and even that contains a ton of tech from Tesla. Incremental optimizations and die shrinks (cutting consumer FP64 heavily from Fermi to Kepler to reduce power consumption and heat, increasing CUDA core count and clock speed with each architecture update, Delta Colour Compression, etc.) are both much cheaper and more reliable. If it ain't broke, why fix it?

Vega is absolutely carrying a ton of GCN with it, because not everything about GCN has been bad. I don't think they will patch up every single gap present in GCN when they release NCU, but that will hopefully be the goal with Navi. Navi is Rajas chance to truly make his mark on RTG and the GPU market as a whole, without carrying the scars of mistakes of the engineers that came before him.

>> No.60513013

>just wait
Stop dreaming poojet. Amd has always and will always be the low end supplier for both cpus and gpus. They don't have something against intel's most powerful cpus and they don't have something nearly as powerful as nvidia's most powerful gpus. Amd is an intel and nvidia cheap ripoff after some time

>> No.60513035

>Amd has always and will always be the low end supplier for both cpus and gpus.
1.7 porkshoulders woodscrews fuck off Huang.

>> No.60513048

Look at their best gpu, 580, it draws more power than a 1080 and barely holds up to a low end 1060

>> No.60513056

>it draws more power
No shit, it has no tiling or compression of any kind.

>> No.60513094

It's also running on 14nm instead of 16nm

>> No.60513106

It's "14nm" and "16nm". Feature sizes are pretty even.

>> No.60513263


>2x6000 RPM San Ace fans

They will cool just about anything. For laughs I bolted a 4900rpm delta fan to be the first fan on a silver arrow (the space for the middle fan isn't wide enough to allow for a 38mm thick fan) and despite sounding like an airport it shaved a good 10c off load temperatures. If my motherboard could've handled it my 8 core chip would've been good for (I guesstimate) 5.1ghz.

>> No.60513345
File: 1.11 MB, 1505x4363, mark.jpg [View same] [iqdb] [saucenao] [google] [report]

>Stop dreaming poojet. Amd has always and will always be the low end supplier for both cpus and gpus.
Gas chambers weren't enough.

>> No.60513776

x86 without AMD is pretty much dead. Even if someone buys them out and starts making x86 CPUs themselves. Which isn't likely.

>> No.60513813

A N O T H E R . M A S S I V E . D I S A P P O I N T M E N T

>> No.60513839

And.... ? the future is ARM anyways.

>> No.60513854

You mean RISC-V?

>> No.60514809


>> No.60514984

VIA** sorry made a typo.

>> No.60515775
File: 2.74 MB, 476x341, 69f9979993d63791c84fdee415e942c505249c02056f0330732fa7ec80881829.gif [View same] [iqdb] [saucenao] [google] [report]


>> No.60515785
File: 138 KB, 653x726, amd.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.60515880


Fury X with all the bottle necks removed.

>> No.60516003
File: 152 KB, 548x629, Dawn of War III 1400p gpu.png [View same] [iqdb] [saucenao] [google] [report]


Does this.

>> No.60516135
File: 1.69 MB, 320x240, wheresmymic.gif [View same] [iqdb] [saucenao] [google] [report]

1.7 porkshoulders woodscrews fuck off Huang.

>> No.60516181


Now add 500mhz to the clockspeed, larger bandwidth and faster clocks to the memory, improved tesellation, geometry, and primitive shaders and rasterization.

Along with infinity fabric communication with Ryzen.

And what do you get?

>> No.60516296


...only as fast as a 1070? I mean that is the current belief held on places like /g/ and OCN.

>> No.60516326

So what about that game footage they showed, then? Fake news? Was actually a 1080 ti under the hood? Only runs good in Doom on Vulkan?

>> No.60516353


Maybe in power consumption is it comparable to a 1070.

>> No.60516378

How come Nvidia's rebrands significantly improve performance while AMD's seems more like a decent overclock? The 580 is pretty underwhelming even if it does have good price to performance. Meanwhile the 1070 matches the 980Ti. Saying "hurr Nvidia does rebrands too" is extremely reductionist considering they're still giving you a reason to upgrade, while the 580 is not a good upgrade from a 480.

>> No.60516394
File: 52 KB, 1024x576, 1494974208295m.jpg [View same] [iqdb] [saucenao] [google] [report]


You mean this game footage?

>> No.60516433
File: 8 KB, 276x182, images (5).jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.60516446

But "AMD is shit" is a valid argument. AMDfags only care about one number, price to performance. Meanwhile both Intel and Nvidia both rape AMD in terms of software support and extra features. Sure AMD is getting better but they've had such a bad reputation in the past it makes people hesitant to buy it. Look at every Steam game forum, it's full of 200 post threads complaining about some AMD specific issue. You get that occasionally with Intel or Nvidia but it's not nearly as common. /g/ has to realize that price to performance isn't a selling point for a lot of people, they want something that "just werks," which is part of the reason why Intel and Nvidia have near monopolies in the business market.

>> No.60516477

What number do you care about?
Isn't it the most important fucking number in any consumer product?

>> No.60516520

Most people don't want software glitches or games crashing just because they're using AMD. AMD's biggest problem is their reputation, AMD products are seen as cheap and inferior even if they have good performance.

>> No.60516539
File: 75 KB, 500x448, LL.jpg [View same] [iqdb] [saucenao] [google] [report]

I'd rather have this

>> No.60516614

> AMD products are seen as cheap and inferior even if they have good performance.

Maybe in Intelshillingland, but here on planet Earth where consumer checks out products before buying them, we see Intel as overpriced garbage.

>> No.60516679


Not to any one who has used their products since Crimson launched.

>> No.60517527

If they want to complete, they'd need to release a 1070 contender for less than $350, which I highly doubt they can accomplish.

>> No.60517551


They released that chip years ago - its called the fury x.

>> No.60517566

Fury X competes with the 970, but the 1070 blows it way the fuck out.

>> No.60517592

Seek medication.

>> No.60517622
File: 43 KB, 500x1130, perfrel_2560_1440.png [View same] [iqdb] [saucenao] [google] [report]


I'm sorry AMD cucks.

>> No.60517653

people knew fury x was powerful but it was also power hungry and hard to find

>> No.60517658

So there's a 12% difference after a year? How is that blowing the fuck away anything? The 970 is so far down it's not even on the radar.

>> No.60517681

You can find 1070s everywhere for $300-$350, while Fury X is at the same price point and near impossible to find without looking through second hand hardware.

>> No.60517701

Obviously because AMD sent a EoL notice for Fury X to make way for Vega
They're not even produced anymore.

>> No.60517706

Gas yourself.

>> No.60517718

And the way it's looking Vega won't be out till fall, will just barely match the 1070 and sold at $400/$500. Even with G-Sync tax, it's dead on arrival.

>> No.60517726

Nice try.

>> No.60517731

Source: My ass

>> No.60517740

I think it's a bit odd that people on /g/ have been shilling the Fury X pretty hard lately, when it's so difficult to find. Pretty much every thread has some guy saying the Fury X is the best card ever despite the fact that it and the 390X flopped so hard.

>> No.60517753

>Muh feelings, my subjective opinions about brands
This board Isa fucking dumpster fire.

>> No.60517754
File: 1.18 MB, 1920x1080, vega announce.png [View same] [iqdb] [saucenao] [google] [report]

>But it still has more memory and TFLOPS?

>> No.60517755

Nobody believes your retarded nonsense, even Fury X at 1600MHz would be close to a 1080ti, much less something with 50% more transistors like Vega.

>> No.60517757

>The rumours also put a $499 price tag on the middle-order RX Vega Eclipse and the RX Vega Core at $399.

Why do you think they keep delaying Vega?


>> No.60517781

it's the AMD street shitting shills. they are really obnoxious. I was considering AMD but now I'm going to build Intel pc rig to spite them because they shit on it so much

>> No.60517783

Great source, started from a reddit thread.

>> No.60517785
File: 395 KB, 2400x2000, the final proof.jpg [View same] [iqdb] [saucenao] [google] [report]

>mfw intel shills kill themselves when ryzen+vega perform better for $800 less

>> No.60517804

K, I'll just come back in September to collect your AMD tears.

You know something is wrong when a GPU architecture as been delayed over a year. Nvidia won.

>> No.60517823


>> No.60517826
File: 9 KB, 231x218, images-1[1].jpg [View same] [iqdb] [saucenao] [google] [report]

>tfw AMD still can't stop tripping over their own dicks
>tfw 2017 will close w/o the 10 series being challenged
>tfw AMD cucks will be hard on the 'jus u wait' meme for another decade

>> No.60517839

I'm pretty sure you'll be too busy damage controlling in a few weeks rather than collecting any tears about reddit pricing rumors

>> No.60517843

Except Vega is likely going to be a flop. It's been delayed too long, the 1070 and 1080 are a year old now. The price isn't low enough and yields of HBM2 are too low. Wait until Vega 2 I guess, there's no point even trying to get Vega 1 since it will be sold out so fast and lord knows when it will get restocked.

>> No.60517855

Idgaf, I'll buy vega day one

>> No.60517858

whatever nerd when are the stocks going back to what I bought them for

>> No.60517863

I want a good cheaper card than a 1070 and avoiding G-Sync tax, but it's clearly not happening. Ryzen may have become a success but Vega is measuring up to be a flop.

>> No.60517881

So you want a 1080 for $300 ? Not happening, AMD isn't charity.

>> No.60517882

>Vega is measuring up to be a flop.
How so?

>> No.60517883
File: 250 KB, 274x403, hestia.png [View same] [iqdb] [saucenao] [google] [report]

The amdrones' anguish and despair are almost palpable.

>> No.60517934

We know it's you, you retarded avatarfag retard, when those ES 1200MHz samples leaked you were so enthusiastic to shitpost, now that it's known it can go over 1600MHz(which was known for months due to enterprise cards which you conveniently ignored) you can only scream about rumored price with no actual source.

To summarize, you're feeling very frustrated, even moreso after the recent Whitehaven leak

>> No.60517989
File: 500 KB, 954x1022, 1461295285157.png [View same] [iqdb] [saucenao] [google] [report]

Chip makers are always working on the next revision before the first one hits the market. This is standard practice retard. Nvidia was working on Volta while Maxwell was still on shelves.

>> No.60518011

AMD isn't Nvidia, they don't keep quiet about new products. Both Ryzen and Polaris received heavy marketing and attention before launch, meanwhile there's almost no info on consumer Vega. Hell, a casual PC gamer probably wouldn't even realize Vega was a thing, there's just been no mainstream marketing. AMD refuses to talk about consumer Vega at all. This means either
>it's a stinker
>or it's not coming out Q2\
The longer AMD waits the closer Volta gets and the more Vega hype dies. If it is an issue with HBM2 it's not their fault obviously, but things really aren't going very well.

>> No.60518051

>the closer Volta gets
Wow bigger Pascal.
>If it is an issue with HBM2 it's not their fault obviously
There are NO fucking issues with fucking HBM2.

>> No.60518054

AMD has shown more leaks of Vega then they did of Zen, which was pretty much that single blender demo at 3.0GHz.
While Vega they showed two games, Doom and Sniper Elite, the latter one on the enterprise card.

>> No.60518074

Buying ATI was a mistake, AMD should just focus on CPU tech again, maybe restart their ARM division. Sell ATI/Radeon tech to Samsung or whoever.

>> No.60518090
File: 217 KB, 848x480, 1262317192659.jpg [View same] [iqdb] [saucenao] [google] [report]

we confirmed for paper launch boys

>> No.60518096

>There are NO fucking issues with fucking HBM2.
Amdfags are so funny

>> No.60518098
File: 42 KB, 499x636, 1488054924552.jpg [View same] [iqdb] [saucenao] [google] [report]

>There are NO fucking issues with fucking HBM2

>> No.60518104

Fucking hell AMD uses fucking 8-hi stacks for Vega FE, there are NO fucking issues with HBM2 production whatsoever.

>> No.60518153

A V A I L A B I L I T Y of HBM2

>> No.60518163

SK Hynix has MORE than enough wafer capacity to produce enough fucking HBM2 to supply Vega cards.

>> No.60518174

So why we still waiting™?

>> No.60518188

You're being delusional dude, even Raja himself has complained about HBM2 availability. It's the sole reason Vega is taking so long, and it's likely supply will be short on launch. AMD fucked up, they should have had GDDR5X to fall back on like Nvidia.

>> No.60518189

Because writing drivers for new arch is hard as shit. Also AMD doesn't want another 290 launch.

>> No.60518208

Launch drivers more than likely.

That HBM2 rumor is silly and completely ignored manufacturing scale, Hynix produces some 1.4 million wafers per month, you can fit well over 600 HBM2 dies on a single wafer.

And then using still unannounced 8GB stacks means they have more than enough HBM2 availability.

Raja mentioned nothing like that, if SK Hynix can only produce some 32000 DRAM stacks in a few months it would literally go out of business.

>> No.60518210

Buulshit, they have issues with the HBM production

>> No.60518234

>That HBM2 rumor is silly and completely ignored manufacturing scale, Hynix produces some 1.4 million wafers per month, you can fit well over 600 HBM2 dies on a single wafer


>> No.60518235

>Hynix produces some 1.4 million wafers per month, you can fit well over 600 HBM2 dies on a single wafer

>> No.60518262
File: 74 KB, 600x264, bulletin20160106Fig01.png [View same] [iqdb] [saucenao] [google] [report]


>> No.60518287

So only when the drivers will be decent, AMD will start the production of the chips... yeahhh rrrrright...very logical lads...

This is very stupid.

>> No.60518301


>> No.60518306

They already started production, and they're releasing datacenter cards first because drivers there need stability and accuracy more than fixing shader bugs in engines like gaming cards.

>> No.60518324

And? You asked for a source about Hynix wafer production per month, you got it.

>> No.60518332

Stop lie on internet

>> No.60518344

I see you've run out of steam, it's okay.
Can't really expect much from gamer kiddies.

>> No.60518347


>> No.60518350


>> No.60518362

Source: My ass

>> No.60518367

>Measuring up to be
Source: My ass

>> No.60518373

My ass is a valid source.

>> No.60518381

>This means either
Source: My ass

>> No.60518405


Oh jesus christ, not again this meme

>> No.60518406

What is the price? They are ready?

Stop pls

>> No.60518442


This card is not even ready



>> No.60518453

You're right, it's not likely, it's guaranteed. Everyone already has a 1070 or 1080, the 1080 outsold the fucking RX 480. There's zero demand for Vega outside of AMD subreddits like /g/.

>> No.60518470

>AMD marketshare went up 8% in a year with only Polaris
>1080 outsold the 480

Source; My Ass

>> No.60518490

Nice made up bullshit there, sempai.

>> No.60518497

>we waiting (tm) the drivers
>nothing to do with HBMeme manufacturing

>> No.60518498

I bet his source is steam, only fucking Nvidiots would participate there.

>> No.60518507

You hate good drivers? Good goyim.

>> No.60518511

Yeah, they should launch with 20% less performance which they'll get back in 2 months, completely ruin their perception in the public eye because nobody cares about anything but day 1 benchmarks.
It didn't happen to Fury, it didn't happen to Hawaii, it didn't happen to Polaris, lets do the same mistake again!

>> No.60518522

Where you bought your game?

Steam is a botnet

>> No.60518526
File: 216 KB, 842x1321, STEAMGPUS.png [View same] [iqdb] [saucenao] [google] [report]

Polaris was a flop. After Vega fails I wouldn't be surprised if AMD shareholders force a sale of ATI, they just can't gain any ground against Nvidia.

>> No.60518535

I don't buy digital games.

>> No.60518554

>Polaris was a flop

The increase in marketshare and their graphics division revenue seem to disagree, but hey, steam opt-in survey, very reliable.

>> No.60518555


They can do both, the production AND the drivers. We wait because HBM2 is turd

>> No.60518572

We are waiting because it's the first fucking major change to GCN you moron and they need at least fucking decent drivers to launch it.

>> No.60518583
File: 52 KB, 568x612, amd2.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.60518584
File: 2.40 MB, 3930x2700, CD-ROM.jpg [View same] [iqdb] [saucenao] [google] [report]

>I don't buy digital games

Of course

>> No.60518597

Even for Nvidia it takes months after launch to polish drivers, AMD's GPU division is smaller and they can't afford to launch another half-assed driver launch.

You clearly have no clue how IC manufacturing works, go sit in a corner.

>> No.60518616


But why not start the production and when the drivers will be ready release all the cards then make money?

They have issues with the HBM, this is the only reason

>> No.60518627

Because they have already started the fucking production?
Weird, huh?

>> No.60518647

Too late

>> No.60518662

Production started a while ago, it takes time to ramp up a launch, usually 2 months earlier than sale.

>> No.60518663


>> No.60518759

APUs when?

>> No.60518767
File: 3.66 MB, 480x270, wait.gif [View same] [iqdb] [saucenao] [google] [report]

inb4 404

>> No.60518775

Q3-Q4 or laptops (SP5), Q1 2018 for AM4.

>> No.60518784

Now, ps4 xbone :3

>> No.60519047

what? the surface pro 5 is going with amd?

>> No.60519059

What? No.

>> No.60519325

Is it just silly to use an AMD CPU and an Nvidia GPU?

>> No.60519653


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