Where's the VEGA hype? AMD seems to have finally perfected their fabrication with this launch.
In fact, AMD has done such a good job, it seems that Vega has already burned down their launch venue before even being formally released. How amazing is that.
Fury X fag reporting. Dunno if I'll go ahead with Vega, or wait until the next round. I'm actually pretty happy with my Fury atm
Landon Taylor
The whole team red thing isn't inventive.
All it does is make me think of the communist party or maybe black panthers and piss me off.
Fucking pajeets don't know how to market.
Levi Hernandez
the only thing i want is hbm2 thats what i have been waiting for
Hunter Foster
I've stopped buying into amd hype since zambezi.
Josiah Bell
>Radeon RX 490
kek 1070 and 1080 exist, soon amd gettingcucked by price cut.
Camden Wright
Don't you mean performance cuts?
Nathan Myers
>Where's the VEGA hype
Hopefully nowhere. These hype trains do a better job at hurting AMD than nvidia
Brody Flores
it's gonna be doa dude, who's gonna paid for $400+ when 1070 and 1080 exist ?
Christopher Lewis
Are you selling your 1070/1080 for 400 dollars?
Jordan Cruz
Wew lads, imagine the asspain this is going to cause from all the DX12 benches.
Anthony Clark
ONE HUNDO FPS IN AOTS
Thats ashes of the singularity btw not anime of the season, also people really play this game its totally relevant.
Dylan Roberts
>wait for Navi are we there yet?
Caleb Hughes
daisuki dess x_DD
Joseph Young
Just wait for the next amd shit.
Jaxson Hernandez
nVidia is really losing the anime war.
Daniel Martinez
...
Parker Gray
>amd >hype
i laugh every time. amd hypes their shitty products months in advance, then they turn out to be incredibly disappointing.
meanwhile nvidia announces them a few weeks beforehand, and they perform exactly as promised, or above.
when will amd fucking learn?
Jayden Nelson
does vega compete with the titan xp
Austin Martin
AMD shills their products in advance and they disappoint nvidia releases spontaneously and they're good AMD slowly improves their products post-release nvidia slowly gimps their products post-release
Whether you are an AMD Vindapoo or an Nvidiot I think we can all agree with this.
Juan Butler
communism is the best political system
Jeremiah Watson
WWIII when? This planet needs some regulation.
Brandon Martin
>hyping anything AMD
Jesus christ, you can't be serious. These fuck ups can't do anything right and Vega will be no exception.
Jordan Wood
>amd >hype
After the 480 failure no one cares. No one is going to buy these shit cards.
Nathan Nelson
DIE ALREADY FUCKING AMD! FUCK
Jose Lewis
I'd love it if AMD managed to actually fight back and prevent Nvidia from keeping monopoly in the high end GPU market, but I just don't have any faith in them anymore. Too many fuckups, too many disappointments, I just can't keep my hopes up anymore.
Carter Bell
This news/rumor is going to send AMD over $7 today.
Mason Williams
Said no one ever.
Isaiah Adams
Why are they holding their Vega launch venue at abandoned railway station?
Brayden Wilson
Stocks are irrelivent
Adrian Nelson
That gimping meme happened once and never was even proved. Instead of spouting memes like a fucking retard just monitor new AAA releases and people uploading 980Ti/1070 benches on Internet to see if 980Ti magically starts being 15% behind 1070.
Jack Brown
>irrelivent
OP first line "Where's the VEGA hype? "
The stocks are going up because of the hype. I'm sorry if you're not able to see these connections. It's a primary factor that differentiates us humans from you mice.
Grayson Smith
He is right in the way that every time some stupid shill posts AMD stocks he also completely ignores fact that for every dollar AMD went up Nvidia and Intel went 2 more.
Elijah Davis
No, they're going up because of the earnings report from Nvidia clearly showing an expanding market for GPU's and their applications, which makes AMD look to have a promising future. Nvidia has gone up much more since their release, but there aren't many players in the GPU market so. Institutions aren't buying up AMD because of Vega hype, they don't even know what the fuck Vega is. The market just looks good.
Lincoln Collins
I'm the one who posted the chart. Thank you for a better explanation.
Dominic Watson
If you care to look, their recent rise actually correlates almost perfectly with Nvidia's since the earnings release. Not implying that you would invest in AMD, but people with the amount of money needed to move a stock so drastically aren't going to take such large bets on an unknown product with an unknown launch date when their performance is reviewed for every quarter.
William Mitchell
(you)
Christopher Clark
K A N K E R A N K E R
Ethan Robinson
Chuckled
Joshua King
The important thing is how you feel.
James Morales
>AMD so poor that they have to use an abandoned warehouse to present their flagship card top kek
Angel Sullivan
The 1080 wasn't a big enough jump over the 980 Ti for ~$700, and the Titan X 2.0 will be a 384b
Caleb Edwards
The hard voltage limits on the 1080 really were disappointing for people that had OC'd 980ti's, From some of the results seen with water cooling RX480's getting 1500MHz, Vega might be looking pretty good. Especially if some good DX12 games come out.
Gabriel Miller
A 20% overclock on water is nothing to write home about. Polaris ended up just OK, but was still very disappointing after the excessive hype.
I really hope Vega is good, but I'll believe it when I see it this time.
John Perez
>nexgen
Luke Barnes
Yes, AMD has licensed the second gen of Z-RAM, they might be using it.
Asher Evans
The patent depicting GCN with variable size CU and double(or more) scalar units is Navi. Vega is 4th gen GCN equivalent of Fiji.
Aaron Baker
That factory looks like its been defunct for years typical for AMD
I bet Nvidias facilities are pristine
Jonathan Parker
>with variable size CU The fuck is that? Why would CUs be a different size and why the hell would that be a patent?
Colton Harris
>AMD >Compete with Nvidia they cant.
Christian Mitchell
without spelling errors, you could dismiss every slide by AMD as fake.
Michael Ross
...
Logan Bailey
What nvidia facilities? Neither company has in house foundries.
Owen Wright
>why Lower energy per op. In their 4X16 SIMD setup not all of the lanes are used all the time. Having them draw power when only half or less of the CU is doing something justs wastes power. AMD's evolution of the GCN arch will feature CUs that are varying size. Some will be 4X16, others 2X8, 4X8, or any variation therein. They're increasing scalar throughput and allowing them to handle smaller ops as well.
>why would they patent this!? Because thats how patents work you fucking imbecile. You patent new proprietary technologies and implementations of technologies.
Liam Powell
>That gimping meme happened once and never was even proved.
The gimping meme relates to nvidia dropping the ball on the 7xx line, and not optimizing fuck all for it after the 9xx cards were out. The 780 was head to head with the 290, then a few years later it ended up way behind.
Nvidia gimping means that they drop optimizations the second the next card is out, and this is pretty damn well documented.
Brandon Ross
So it's literally copying Nvidia again... typical AMD
Jaxson Morgan
Nvidia has nothing remotely similar. Shitposting is a bannable offense.
Carson Smith
Haha, the drone is mad as fuck
Luke Murphy
>>with variable size CU >The fuck is that? Why would CUs be a different size and why the hell would that be a patent?
The fundamental challenge with general software on an SIMT-style processor is that different pixel threads can take different branches of execution depending on their input data.
The earliest and simplest strategies were to simply mask off the instructions in branches not taken for a given thread and to execute every path, which was obviously very wasteful.
The name of the game now is to dynamically re-pack threads with the same execution flow while using as little power and logic space to do so. Variable-width execution units are one way of avoiding ALU waste and potentially to make flow rescheduling simpler.
Jason Sullivan
Hoping it gives me a reason to upgrade from my 290X, as the Fury cards didn't.
Jackson Morris
I don't give the argument any credence because it was obviously due to game engine changes and the cards capability and not poor drivers like you imply.
John Moore
So at this point Vega at least won't be
Logan Price
oh i thought it was a misspelling
Julian Lopez
>halfway competitive in legacy engine design
Then you have guys like me who don't give a fuck about older games. I just need it to play DX12 BF1 good.
Jace Scott
Satori is such a little dirty slut.
Benjamin Nguyen
BF1 already runs max out above 60FPS using DX11 at 1920x1080 with a GTX970.
Oliver Allen
You fags are expecting too much from amd or just companies in general. Competing doesn't mean surpassing. They only need to match the performance, then charge a lower price.
Lucas Baker
the 970 is a good card, but being what it is you should not be promoting it. Tell me to get a 980 or something.
Luis Smith
Why would you get a 970 "3.5" if the 480 exist. 900 series was literally declared legacy by nvidia.
>Nextgen memory Why the fuck wouldn't they just stick with HBM2? It's already such a huge improvement on GDDR5, why would they spend even more money on the R&D to adopt something else so quickly? What's the point?
Only logic I can fathom is that whatever the "nextgen" memory is it'll be cheaper to produce than HBM, but I doubt that.
Thomas Nelson
What's the next step to improve bandwidth and throughput over HBM2?
Gabriel Harris
AMDfags just need to get the fuck off of this board. Tired of their obvious shilling
Brandon Robinson
Why do you assume it means VRAM when a GPU can have 3 different kinds of memory?
Anthony Clark
>Exactly as promised 3.5
Gavin Murphy
HBM is still conventional DRAM. DRAM is not the fastest thing on the block, and it is not the lowest power per clock memory around. Each PHY is 1024 bits wide, and the memory attached to it can only clock so fast. You end up hitting a bandwidth wall pretty damn quick because you can't squeeze magic out of DRAM slices. Say HBM2 has a memory clock of 1000mhz, double that of HBM1, and it does so with the same drive voltage. Outstanding, but now what? You're not going to double it again. You need something that has room to scale in the future.
HBM is only a first attempt at standardizing and popularizing wideIO memory. Its not a future proof standard. GPUs will be utilizing 1TB/s of bandwidth in no time, and they will be limited in no time. If they went with the approach of GDDR and just added more PHY then they'd be making negative improvement on bandwidth/watt again.
Ryder Stewart
If this thing manages to slot in between the 1070 and 1080 what I would do, knowing that everyone is fucking with retail price, is peg it at $359 MSRP so it hits shelves at 370-400
Also good luck finding a 1070 for non gouged msrp.
Isaac Hernandez
Another Fury X fag here. Custom looped mine, so unless Vega is something mindblowing, I'm with you.
Christian Rogers
It will be a hybrid memory solution for workstation and compute applications, similar to 3dxpoint. Micron has already made prototypes for a multilayer hybrid memory that combines DRAM and flash memory into a single chip, and the product may be to market by the time of Navi. In fact, AMD has already expressed their interest in hybrid memory solutions by making a GPU with dual M2 slots. It does wonders for machine learning and working with very large datasets.
Luke Bennett
but it was already came water cooled desu :^)
Jack Allen
The cooler is shit compared to my loop. 720mm x 26mm length/thickness. Maximum temps dropped from 62c to 41c.
Charles Walker
Forgot pic.
Alexander Cook
Nice loop. I don't get all the people who dissemination Fury X. It's a great card desu, perfect for 1440p
Michael Powell
>Dissemination Diss on*
Carter Torres
Exactly how I feel about it. I'm using a 2560x1440~144hz Freesync monitor and it does me wonderful. The thing that really helps though is I'm not autistic and expect EVERY setting to be the ABSOLUTE MAXIMUM. I'm happy turning down settings which either don't do much/anything when you're actually playing and not looking for it, or cost too much resources like gimpworks shit (godrays in Fo4... eugh.) and so I get a very satisfactory experience out of it.
Is it top dog in every game? No, but I knew it likely wouldn't be when I bought it. Am I upset that it isn't the OCers dream like Lisa Su said it would be? A little. Custom loop cooling definitely helped it, but even then I can only do 1150 - 1200 at absolute most.
If Vega can provide a worthwhile upgrade, I'll consider it, but I doubt it will be. Navi is where I expect the real improvements and viable upgrade path to come from.
Nathaniel Roberts
Also, thanks. I've definitely got changes I want to make to it, but not just yet. It hasn't been too long since I installed it in the first place and can't give in to the monkey.
Ryder Diaz
you don't need a high school diploma to work as an engineer at AMD, they are pretty much taking anything at this point
Hudson Brown
i would rather have something reliable than a paperweight
Ethan Roberts
>Getting hyped for Vega after Poolaris You only have yourselves to blame when it turns out have a 250W power envelope and still not outperform the 1080.
The last good GPU AMD designed was Hawaii, and even that was flawed
Charles Myers
>it just werks lol Are Nvidiots the new Macfags? Why does Sup Forums actually support this shit?
Benjamin Anderson
Polaris 10 is technically impressive, it was just designed to be a cheap mid range GPU.
Fiji shits all over Hawaii.
Vega will shit all over Fiji.
Grayson Flores
Aren't you the same faggot saying two months ago that this was gonna be Polaris's secret sauce?
This may or may not ever see the light of day, but it will sure as fuck take more than two years. Fundamentally new SIMT architectures are brutally hard to implement and verify.
Ethan Anderson
>Polaris 10 is technically impressive,
it's just another rehash of gcn, same shitty 8 stage pipeline and terrible perf/watt and OC headroom.
Isaac Taylor
i mean if i was poor, i would go with amd if i wasn't, amd is out of the question
Asher Taylor
Sup Forums seriously needs to do something about the Nvidia shilling, its like how Macfags invaded back then but nobody's collectively telling them to fuck off.
Levi Rodriguez
That feel when still rocking 7970 That feel when bf1 released soon That feel when probably unplayable without upgrade
Caleb Garcia
We try user, but it goes nowhere. There's too many of them that believe it to be true.
Hudson Gomez
where are you getting the idea that there are nvidia shills on Sup Forums? nvidia's products are currently indisputably superior to AMD's across the board, if anyone's shilling here they're shilling the indian poo abortion that is the RX 480.
Logan Robinson
Its been going around various forums for a long time, people speculating on it being Polaris was only a guess. Navi promises scalability, and that is scalable GCN.
Its 4th gen GCN. Each iteration in the family is distinctly different from one another. The pipeline length by itself means literally nothing, stop regurgitating things you don't understand. Polaris based GPUs have more OC headroom on average than any prior GCN cards. Perf/watt is only limited by pixel throughput. The Polaris 11 and 10 dies were designed to be cheap to produce.
Stop being a shitposting fanboy and try to actually learn something. This isn't Sup Forums.