Will Vega be enough to push 120+Hz @1440p? I ask this because

Easily

>>58446263
Vega is built for 4k

1440p is fucking nothing for that GPU.

>>58446263
Yes, but the problem is that UHD@120/144 Hz displays are coming out this year, so 1440p120 will be poorfag tier before you know it.

>>58446477
The highest returns are from above 60 to 100Hz, 144 is just eye candy but has diminishing returns.

My eyes suck anyways, I haven't seen much benefit all the way to 4K for me personally.

>>58446477
>1440p120 will be poorfag tier before you know it.

Good, I need to update my triple monitor setup and I would rather not spend thousands needing to do so.

The question is pointless. Every game is different.

>>58446583
Well I detailed my target game for high FPS is Overwatch.

>>58446263
>Will Vega be enough to push 120+Hz @1440p?
No,

Even my GTX 1080 running at 2,310mhz can't achieve 120fps at 1440p in every game.

Ur gonna have to wait for the 1180 and beyond for that.

>>58446263
4k and 1440p 144hz have similar GPU requirements

But 144hz requires as fast a CPU as you can get

>>58447021
I'm pretty sure that frequency isn't real

But you probably have a CPU bottleneck in that case

>>58447021
>GTX 1080
Vega will be 20% faster than that, lad.

can they hurry the fuck up and release ryzen and vega so I can give them money

I'd like to build a new fucking rig already

>>58447105
Fucking this

>>58447059
>Vega will be 20% faster than that, lad.
[citation]

>>58447149
see
>>58447059

>>58447105
reeeee and this fucking naples shit

got ideas about a dual cpu workstation but i want to see how it compares to xeons

>>58447105
This, for fuck sake.

>>58447059
still not enough for 120fps 1080p

File: 1455592512434.jpg (99KB, 600x600px) Image search: [Google]

99KB, 600x600px

>>58447059
>Vega will be 20% faster than that, lad

Although vega will be a "4k card," with the goal being 4k60 in most games, 1440p is a bit of an oddity in that its generally more demanding than its resolution might suggest.

1440p will largely call for the same pixel dense assets used for 4k as well as the post processing required to make 1080p renders sharper/smoother and angle-correct.

120hz itself is also a tall order regardless of resolution. Cpu heavy games can make achieving that framerate consistently very difficult.

It's not impossible by any means, certainly a game that places the largest load on the gpu and utilizes appropriate assets should have no trouble, but more cpu heavy instances like open world games or those with bad optimization in general will always pose a challenge.

Tl;dr depends on the game yo

>>58446263
even with all the shit vega is currently getting, yes it will.

HOWEVER, the issue lies in

1) not every game can be maxed, or even make it to 144 by design or incompetence.
2) you need a cpu to push it to 144 in every game, many will make it, but even an oc to 5.2 kabby will find it hard to push some games passed 110fps

realistically, the cpu will be more of a bottleneck then the gpu will be.

also, as for 2-3 years, its possible to last that long still staying over 60, but soon lighting is going to be pushed hard as the next front to increase visual quality and it will bring gpus to their knees, so be prepared to run settings down even if you have the best shit possible.

>>58447615
Well I have a 6600K, which seems to overclock nicely and handles just about anything thrown at it thus far.

>>58447672

the cpu will push many games to 144, but will be held back in the 90-110 range for many others.

there is a reason why people say "the difference between 90 and 120/144 is not noticeable" these people literally can not push the frames that hard on a cpu/ram level

oh, and another fun fact, ram has become a bottleneck in many games as of recent.

https://www.youtube.com/watch?v=Er_Fuz54U0Y

depends on the game, some are bigger than others, fallout 4 i believe being the biggest one.

>>58446513
thats what you signed up for tho, retard

>>58446396
2560x1440@144fps is more pixels than 4k@60fps.
Not by a great leap, but still more.
>>58446263
100FPS+? Definitely. 120+? It'll depend on the game.

File: 1483639063516.jpg (993KB, 3840x2160px) Image search: [Google]

993KB, 3840x2160px

>>58447059
That's where your wrong kiddo.
Vega matches a 1070 in doom/Vulkan.
An API designed BY AMD.

>>58448280
Engineering sample.

>>58446477
>so 1440p120 will be poorfag tier before you know it.

i hope you're right. i'm going to get a new monitor sometime this fall probably and I hate having to decide between 1440p OR 120+Hz on my budget

File: ayyymd.jpg (439KB, 1920x1080px) Image search: [Google]

439KB, 1920x1080px

>>58448280
>37fps

>>58446396
They said vega was for 4K just like they said the 470 is for 1440p

>>58448544
Breddy gud maymay but the inappropriate use of the xzibit macros is hideous

File: 1422805571908.jpg (117KB, 768x576px) Image search: [Google]

117KB, 768x576px

>>58448544

It is typical of Nvidia fanboys to be impatient. Neither RyZen nor Vega are out and you are spouting bullshit memes.

It has been proven time and time again that with AMD hardware (at least with the last generations of GPUs) you get more and more returns with every patch and they just get better and better.
But Nvidia-children literally have no willpower and always need instant gratification even if it comes with gimped drivers half a year later.

You people are like morons that shoot themselves in the foot and then wonder why you can't walk anymore.

>>58449396

by the way, for the consequences of the inability to exert your willpower to delay gratification, please refer to the Stanford Marshmallow Experiment - it might explain why Nvidiots are basement-dwelling retards.

>>58446263
Look op, I've been using 120hz for years. The refresh rate doesn't matter as much as the pixel response time - that means motion blur. 120hz is nice and fluid and all but with some panels it is a blurry mess once you start swiping quickly. That means you can't hit shit with characters like genji/mccree/pharah requiring flicks. Tracking characters like tracer are fine however.

What you want is to buy a g-sync panel with ULMB support.

t. solo q 3700

>>58447672
I bet you can't get 60 fps stable at 1080p playing tera online, a game from 2011

>>58446396
Vega does 60fps at 4K WITH highly optimised code. It's the min for 1440 100hz gaming

When will we be able to virtualize the windows aspects of gaming from linux and get decent framerates?

>>58446641
Overwatch, much like other Blizzard games, has shit GPU usage efficiency and tends to scale horribly with a GPU (Play WoW or Overwatch with a 1060 and Titan and enjoy being within 10 FPS difference), they scale astronimcally better with CPU (WoW went up 20 FPS for me upgrading from a 2500k to a 4970k while retaining my same 980ti) How this retarded shit happens is known only to Blizzards Pajeet game engine team

>>58448572

yeah at low settings.

my sapphire fury that overclocked cant handle constant 120fps in BF1 Ultra at 1440p

>>58446477
I hope 1440p 120 becomes poorfag level. Then I could finally get good 1440p 144hz IPS monitors at reasonable prices.

For a 27" screen on my desk 4k seems pointless anyways.

File: 1478718255653.jpg (208KB, 728x518px) Image search: [Google]

208KB, 728x518px

I don't think so. Not even the Pascal Titan X can do it and Vega isn't that much faster than a 1080. This generation is bullshit with the only decent cards being the RX 480 and GTX 1060 where the latter is already getting fucked.

>>58452335
RX 480 is the best in perf/$ this gen (excluding yurofags with punishingly expensive electricity), but there's not really some grand bargain this time around.

Vega and/or Volta may end up decent enough, but I suspect we'll only see great performance if a lot of games move towards fp32/fp16 hybrid shader precision.

>>58448280

Lmao how long were you going through frame by frame to catch that dip lad

File: AMD business model.png (62KB, 1008x854px) Image search: [Google]

62KB, 1008x854px

>>58449396
>Neither RyZen nor Vega are out and you are spouting bullshit memes.

>>58446477
The real "difference" is once you start hitting framerates over 90-95. This is why the VR market is trying to hit that fps, and Steam has commented that 95 is the optimal refresh to feel "presence" in a VR simulation. The difference between 1440p with AA and 4k is not a life changing experience, but going from 60 to 90 from my experience (and from the gaming industry) is very noticeable.

>>58453075
The only thing magical about the 90-95 fps region is that's around the point where most people won't get flicker-induced eyestrain and headaches from the strobed displays needed for high motion clarity.

Some people still have issues (old fluorescent bulb ballasts still bothered a lot of people at 120 Hz flashing), and it seems to become a greater issue in wider FOVs, so there's a strong argument to getting content and display chains up to the 200-240 Hz/fps region for that alone.

>>58446263
Vega will probably suck. Or not.

>>58448296
just like Poolaris and Furry?

>>58453317
Poolaris is great.

File: 1483629357673.png (181KB, 293x397px) Image search: [Google]

181KB, 293x397px

>>58453329
for (you)

>>58453336
>not for (poo)
One job anon.

>>58446263
>@1440p
meme rez, please stop ruining the market by wasting manufacturers resources on this instead of focusing on 4k on IPS with Freesync.

File: silicon lottery.jpg (446KB, 3840x2160px) Image search: [Google]

446KB, 3840x2160px

>>58453317
>mfw this is the RX 570

>>58453415
>please stop wasting resources on higher refreshes please rape my face
Fuck off man, high refresh is always a better thing unless all you do is watch anime or movies. That's when you use a TV hooked up to your PC and call it a day. Speaking of that, why are TVs so shit? Where are my lower latency TVs with freesync?

>>58450678
>Vega does 60fps at 4K WITH highly optimised code.
and with
>throttled cooling
>ES clock speeds
>unoptimized Fiji drivers

it's simple logic. Vega will perform better out in the field than it did at CES.

the drawing area of 1440p is roughly 45% that of 4K, so it's safe to say the difference in performance between the two resolutions should be massive. not double, but still.

I do believe OP's target is possible for Vega, but probably not in games like TW3 or GTA V or anything that will come out in the next few years.

>>58453075
I agree higher framerates are very beneficial in FPS, in fact I noticed a difference in serious sam 3 going from 60->90->120 (that game has excellent multi gpu support, by the way. Like >200% scaling which makes no god damn sense), but other games less so.
Racing games, yes in first person mode, but It sarts to blend in once the camera pulls back to third person. Games like GTA and Mad Max (fuck you I loved that game) demand 60fps stable, but beyond is lost to me.
Third person games lacking fast paced driving sequences benefit the least from higher framerates in my opinion. TW3 1080p60 on a 1ms tn didn't feel especially fluid compared to 4k30 on a big dumb ips tv to me. I mean i could tell the difference easily but it didn't feel super special so I stuck to 4k cause muh pixels.

I imagine that's why so many games are going third person open world: to cover up the laughably inferior framerates of shitty consoles and their shitty pc ports.

Haven't tried VR yet but it's been making my dick hard just thinking about it since the kikestarter.

File: worship toilet poo indians.jpg (103KB, 597x799px) Image search: [Google]

103KB, 597x799px

>>58453365

File: 1483644949084.jpg (38KB, 600x453px) Image search: [Google]

38KB, 600x453px

Who cares, I just want AMD to merge their code on the Linux kernel and hopefully get better performance. I want to get rid of this 970 already, fuck.

File: AMD-VEGA-10-VEGA20-VEGA-11-NAVI-roadmap.jpg (222KB, 1562x855px) Image search: [Google]

222KB, 1562x855px

>>58453417
that (dip to 37 fps or whatever) is clearly a driver/shader hiccup from not having some assets loaded when they needed to be.

Polaris is gonna get phased out pretty quick by low-end Vega, and the only question is whether it sticks with GDDR5(X) or has a single HBM2 module.

If the High Bandwidth Cache Controller noise is somewhat real, Vega (10, at least) might be using the HBM2 modules as L3 cache with main system RAM as a backing store, and Vega 11 could skate by with only 4GB HBM2 to save costs.

>>58447300
underrated post

>>58453446
>mfw went to the movies recently after weeks of glorious 60fps gayman on low latency monitor
>camera pans looked like choppy shit, almost like screen tearing
>The lack of fluidity looks slow and unnatural
>can distinctly tell when lips are out of sync with dubbed-over dialogue
>feels like I'm watching mannequins
>tfw I've gone full circle and 24fps feels unnatural and the "soap opera effect" lends realism to the scene

Fuck you Peter Jackson for killing 48fps with your shitty interpolated digitized mess of a Hobbit remake nobody asked for.

Kekeke, Vega will be released around May/June.

Have fun waiting

>>58453446
>playing online games in general
pleb

>>58453629
I watched that one guy's video about Vega, you know the one, and I have to admit he brought up some valid points.
Vega reminds me of the original Titan, that is a floating point monster capable of the kind of parallel computation CPU's have found themselves laughably lacking in over the past couple years thanks to market stagnation that can also play games quite well.

His best point is in the appeal of hitting the professional and enterprise markets which make the consumer market look like peanuts. Zen's focus on high threadcounts from the bottom up supports this as well to a degree.

I really want both to be winners but I fear they might be winners in the market that doesn't apply to me.

>>58453585
This is hilarious, but it has nothing to do with Raja or ayymd, just like any /int/ meme.

>>58453916
I remember buying my gpu last year and /g/entoman told me to wait for polaris which would beat the 980ti and come out in mai (2016). now they're singing the same tune for vega.
truly AMD is the best bang for your buck because you never get to actually buy anything. just close your eyes and dream of the wonders they will bring...soon.

>>58454524
What did you buy?

>>58454524
>I remember buying my gpu last year and /g/entoman told me to wait for polaris which would beat the 980ti and come out in mai (2016). now they're singing the same tune for vega.
first off you should have waited because even the 1070 beats the 980TI iin 9 out of 10 games. second people didn't know that AMD wasn't going to release high end cards which would have lowered costs for all cards that were already out. if you can't see the value in that well then you're just plain retarded.

>>58453417
>1475MHz
this is a hefty overclock that guzzles power like your mom guzzles cum.

most AIB cards should be halfway between that and reference, perf wise.

>>58450878
someone who likes the word "scale" so much should probably look up the word "bottleneck"

File: overclocked stats.jpg (737KB, 3840x2160px) Image search: [Google]

737KB, 3840x2160px

>>58454824
Hence the silicon lottery, can you even read? Also, it's one of the god tier binned cards that only have a 95W TDP. Even with the hefty overclock, it consumes just about 135W of power. This is probably what we're going to see on the next series of cards, refreshed Polaris with better binning.

>>58453203
Try going from 90-100 and then back down to playing at 60. It's more noticeable (and important) for VR but the same smoothness translates to playing on a monitor.

>>58455357
that's true, but:

Locked 120+ Hz w/ strobing >>>>>>> locked 120+ Hz w/o strobing > 90-120+ Hz adaptive-sync > everything else.

Increasing fps does help smoothness incrementally, but it's nothing compared to strobing on a sufficiently high Hz display.
With the advent of DP 1.3/1.4 displays, I'd rather AMD and Nvidia focus on dynamic detail scaling with locked high fps instead of variable display sync.

>>58455292
Dude Polaris is designed for the highest yields possible. The stock clocks are basically on the edge of reason. The 480 itself is basically the lottery winner chip, 470 is like what you'd normally expect to get and 460 are the binned ones.
The high-OC capable 480's are unicorn-tier.

Not to discredit Polaris. It's actually pretty genius and proof of the production efficiency with how many 480 dies they can pull out of a wafer.
I imagine margins are pretty good too, despite being budget range.

>>58455823
Hey genius, you're right about the rarity of high-clocking Polaris chips, but RX 460 is a separate chip (Polaris 11 = 128b) from 470/480 (Polaris 10 = 256b).

>>58455292
That was only reading the die power, not the wattage of the entire card. I like his channel but he should have been more clear about what it was actually measuring. Realistically the whole card could have been pulling between 170-200w for that run.

>>58446263
>but not sure if AMD is a viable option for a GPU that will last 2-3 years at 1440p.

If you expect to hold a card for 3 or more years, AMD is the only option.

>>58457281
I'd expect NCU to be OK for years, but I'd be more nervous about picking up a Polaris card now and expecting it to get 3+ years of decent support.

>>58455292
You're the single retard who keeps repeating this despite being corrected dozens of times.
The 95w Polaris based parts that AMD is offering are not miraculously higher binned, they simply are not sustaining the same clocks. They have a lower power target. Power target is an adjustable metric that you can freely change in the Wattman software.

Of the RX 480's 150w power target 110w is taken up by the ASIC, and 40w is taken by the GDDR5.
For this GPU to have a stable 95w pull at the same clocks would mean a 55w reduction in ASIC power draw.
A 50% reduction in power, when 14nm LPP parts already have ungodly low leakage.
This is not happening.
Ever.
The average RX 480 can have power target reduced 10% to 20% with zero to 2% performance impact.
Thats the power saving you can get from a decent die, anything even remotely higher than that would be astronomically rare.

A 55w 1266mhz Ellesmere die is not possible. Stop regurgitating this.