Leaked Vega slides showing it beating Titan Xp (the new Titan).

where did you get these from?

>>59919622
He got them from my uncle who works at nintendo

File: amd.png (42KB, 653x726px) Image search: [Google]

42KB, 653x726px

(((leaked)))

>>59919613
ofuq

I wouldn't really be surprised if it's faster than gp102 in certain titles, it's both a bigger chip and a new architecture.

I wonder where these 'leaks' come from.

>Human eye can't see past 60 FPS or 3.5 GB

-Nvidia

>>59919613
> Vulkan
> made by amd
Of course it will beat Nvidia on Vulkan, faggot.

>may 8


Reveal at GTC? Fucking nice

come to think of it 140 fps on 4K is pretty fucking crazy

Fake and gay

Wait, these cards can run Doom at 4k on ultra at 100+fps? I thought they'd struggle to reach 60.

>>59919738
HBM and big die, lads.

Vulkan is best case, but even 20% slower than that is competitive.
Meanwhile getting more perf from the tessellation slider is free performance.

>>59919774
>even 20% slower than that is competitive.
King Pajeet & frinds were supposed to fix shader utilization in Vega so it's probably less. Maybe 5-7% difference at worst.

It'll be interesting to see reviewers not test a single dx12 game in 2 months.
Or using Skylake instead of Ryzen

File: 1483678747438.jpg (138KB, 1200x675px) Image search: [Google]

138KB, 1200x675px

>>59919722
>It's ok when nvidia does it

>>59919799
can't wait for suddenly ROTR,warhammer,BF1 and doom are shit for testing GPUs

you should test only with GTA5, WD2 and some other random game everyone already forgot about

>NEWER GPU IS FASTER THAN OLDER GPU


WOOOOOOOOOOOOOOW

>>59920049
This desu, another massive failure by AMD

>>59920049
>>59920087
oy vey splendid observation brothers, here's 3.5 shekels for your troubles

>>59919613
>VEGA RUNS FASTER ON THINGS OPTIMIZED FOR AMD THAN NVIDIA

>>59920049
How old is the TitanXp?

>>59920129
Both contemporary APIs are "optimized" for AMD rather than Nvidia.

Rly made me ponder

4k is a meme resolutiin ,try 1600x900

>>59919613
I'll believe it when I see it, besides
>Vulkan
>Doom
Very specific case that they would have fine tuned their hardware for considering they have demoed it to death. I love Vulkan as much as the next guy but current library is mostly DX11 and I would like to see its performance with other Vulkan/DX12 titles besides Doom.

>>59920156
Both are good resolutions.

>>59920156
>$700 gpus in 2017 on 2006 resolutions

>>59919613
that's faked, you know

I am an AMD fan and even I can smell bullshit.

>>59919613
>Titan Xp getting 190% the performance of the 1080
>Not something believable like 130-145%
Yeah this is fake.

NEW GPU FASTER THAN OLDER GPU AMAZING AMD

Looks like a GTC slide, if it's not faked of course

Fake as fuck. No one even has the new titan xp to even test it. I can bet my money vega will be at BEST 1080(non ti) level performance while poojet mentions good price/performance

>>59920135
older than nvidia wants you to think (they were selling the same shit as quadro p6000 or something before)

>>59920470
Your money is as good as wasted.

>>59919729
>>59920428
GTC is an nVidia only event. AMD has never participated in it.

File: Screenshot from 16-04-2017 17:07:26.png (167KB, 1920x1016px) Image search: [Google]

167KB, 1920x1016px

>>59920572
Why do you think that can't change?

>>59920608
Because the fucking event is hosted by nVidia.

>>59920470
Since you didn't at least throw in some qualifiers like at 1080p, you're fucked.
Vega will beat even the 1080Ti at 4K so hard.

>>59920049
>Technology falls from the sky

AYY LMAAAOOOOOOO

>>59919613
at this point AMD needs to completely BTFO nvidia if they want people to buy vega

>>59919613
Source or gtfo.

>>59920652
People will still buy NVIDIA.

>>59920667
Gaming cattle enables worst kinds of business practices for this market. Genocide when?

Intel: BTFO
Nvidia: BTFO

2017 confirmed for being the year AMD finally got their shit together again.

>>59919722
AMD gave Nvidia the chance to pick up Vulkan, it's open source. Unlike anything that Nvidia does to try and cripple the competition.

Nvidia have been 'winning' on DOOM since release, as they have the more powerful cards, hopefully just until Vega releases.

>>59920688
>2017 confirmed for being the year AMD finally got their shit together again.
About fucking time.

>>59920697
>AMD gave Nvidia the chance to pick up Vulkan, it's open source.
Almost everything AMD does is open-fucking-source.

File: jpr_q2_2016_amd_vs_nvda_SHARE_575px.png (31KB, 678x474px) Image search: [Google]

31KB, 678x474px

>>59920652
>>59920688
AMD completely BTFO Nvidia from 2008 all the way up into 2014.
People still bought Nvidia.

It's just like when AMD was completely BTFOing Intel from 2000-2008, but people still bought Intel.
AMD marketing sucks ass and Nvidia fanboys will always be delusional.

File: IMG_2564.jpg (102KB, 601x806px) Image search: [Google]

102KB, 601x806px

>>59920688
>Intel: BTFO
LOL no

>>59920742
Ryzen is new hardware, the software needs to catch up. Once games begin to use more than one core, Ryzen will become more and more powerful.

The i7's have no overhead, whereas Ryzen has a shittonne.

>>59920742
>no min framerate
nice try shlomo

File: its afraid.jpg (153KB, 600x339px) Image search: [Google]

153KB, 600x339px

>>59920742
>day 0 benches

>>59920720
Precisely, which makes them the more user and industry friendly company. Which is why I personally prefer them, even if atm they aren't the most powerful thing in the PC market.

File: IMG_2740.jpg (443KB, 2133x1242px) Image search: [Google]

443KB, 2133x1242px

>>59920785
DigitalFoundry benchmark from yesterday, amd's still shit

Got to love having some actual competition in the industry again. Still not buying their filthy pajeet shit though, obviously.

>>59919613
>Doom Vulkan
So when the Titan is using open GL they'll be about even?

Actually better than I thought, it's a shame it doesn't actually destroy the Titan by that much when both cards are using their optimal API.

>>59920803
>using NVIDIA's GPUs
(You), Ramesh. Holy fucking shit you're doing it since Ryzen launch cald down dude.

>>59920775
>10 core 6950x has no ovearhead
>Once games begin to ose more than one core..
>what is 6900k

Sorry bud, amd fucked up big time with ryzen

>>59920797
just fucking this.

File: 1489475219786.jpg (11KB, 259x194px) Image search: [Google]

11KB, 259x194px

>>59919613

>>59920819
Don't get too excited. OP's pic is a photoshop made by a stinky indian. The new titan xp hasn't even been released. Expect 1080 levels of performance out of vega, no more

>>59920742
Get out of here Gaymer trash. No one buys CPUs for gaming anyway.

>>59920844
>The new titan xp
>new
>titan
>xp
Oshit. Hilarious.

File: 1382663872113.png (497KB, 973x2259px) Image search: [Google]

497KB, 973x2259px

Gonna need a new version of this.

>>59919613
>>59919653
I really want Vega to be amazing
But I am the guy in that picture
I have zero faith in leaked shit unless it's French people showing it

>>59920851
Yes, people who upgrade their CPUs most oftenly are gaymen

>>59920742
>thinking gaming is what determines if a cpu is good or not
>using avg fps as an indicator of what the gaming experience is
>not even putting 0,1% or 1%
>no frametimes

good goy

>>59920889
It is, low framerates means the cores of the cpu are shit, and they can't feed the GPU information fast enough

File: pairings.jpg (2MB, 3840x1032px) Image search: [Google]

2MB, 3840x1032px

>>59920803
>dx12 with an nvidia GPU
kill yourself

File: 64652633.jpg (20KB, 500x313px) Image search: [Google]

20KB, 500x313px

>>59919613

>>59920156
The human eye can't see more than 240p.

>>59920881
I am pretty sure people upgrade their CPU when it can't keep up anymore with average tasks or they want more productive performance. Both of which Ryzen does much better than Jewtel(TM) Stutterlake (TM) for the price.

>>59920928
>amd 480
Fuck off with your low end shit tier hardware you slime

File: op-9.jpg (175KB, 649x621px) Image search: [Google]

175KB, 649x621px

>>59919613
In the odd chance that this isn't fake, the real interesting part is actually the last line, since it implies the release date.

But since it is fake, OP is a fag.

>>59920928
GLORIOUS NVIDIA DRIVERS I ASSURE YOU YOU CAN'T SEE PAST 3.5 GIGABYTES.

>>59920907
Then how does it explain ryzen CPUs being at less than 50% utilization in most game.

>AMD sandbags
>Nvidia creates fake "leak" to generate false hype
>People think the leak is real and are extremely disappointed when the card comes and doesn't perform to "leaked" expectations
>AMD reputation ruined
All according to plan.

>>59920961
Shitty novideo drivers.

>>59920961
Because it's horseshit, not unpotimized, but shit at a physical hardware level

>>59920826
Actually, given that Intel will go towards 6c/12t and 8c/16t processors with Skylake-X (Kabylake-X) its practically certain more software will be optimized for multi core cpus.

If they can't manage to pricematch Ryzen, no point in buying it though.

>>59919613
i saw a leak that showed vega having 1 000 000 fps, nvidia is finished!

>>59920975
AMDtards jerk off all day about fake leak.

File: aHR0cDovL21lZGlhLmJlc3RvZm1pY3JvLmNvbS9RL1MvNjU4MDM2L29yaWdpbmFsL2Rvb20tNGstZnBzLnBuZw==.jpg (31KB, 600x450px) Image search: [Google]

31KB, 600x450px

>>59919613
>TitanXP is 10% faster than a 1080ti

Seems pretty consistent

>>59920844
>what is P6000

>>59920985
OR Game devs didn't bother coding their games to utilize more than 4 threads because of Jewtel(TM) paying them $$$.

>>59920985
Why are you making shit up on the Internet?

CPU utilization numbers don't even show any hardware-related numbers, it just shows the amount of time that the OS schedules a thread to run on a CPU. If a process shows less than 100% CPU usage, that means that it doesn't want to run more than it does. Such as by blocking or by not having enough runnable threads to cover all CPUs in the system.

>>59920803
>from yesterday
you mean from 2 weeks back?

>>59921135
Or they coded them to evade numerous Intel's errata.

>>59921038
Is that Vulkan? Those results seem too good for Vulkan.
But if true, that's actually more impressive results than I expected for the 1080Ti in Doom Vulkan 4K.
If so this is still a 20% improvement still over the 1080Ti which seems too high.

>>59920826
>what is 6900k
best gaming CPU on the market if money is not concern?

>>59921173
>a fucking housefire
Nah.

>>59921156
Different CPU and memory can affect results, Doom loves a lot of threads

File: 1435966903275.jpg (19KB, 366x380px) Image search: [Google]

19KB, 366x380px

>>59920907
can't you read its written AVERAGE

>>59921193
>if money is not concern

see? you can buy a new house

File: firefox_2017-04-02_15-29-47.png (55KB, 764x954px) Image search: [Google]

55KB, 764x954px

>>59921173
>best gaming CPU on the market if money is not concern?
Only because Nvidia drivers are proven to be gimping Ryzen, friendo.
When Vega comes, that'll change.

>>59921203
I should've said "was best gaming cpu"

i'd say 1600x is the best CPU right now.

File: 1486665602657.webm (3MB, 640x344px) Image search: [Google]

3MB, 640x344px

>>59919613
>leaked Fiji slides show the Fury X beating the Titan X
And look what happened there
AMDrones never learn now, do they?

>>59919613
You know how GCN needs async compute (aka async shaders in DX12) to make the most of the hardware but Nvidia architectures don't because of a more efficient cache and front-end (and rasterizer while we're at it)?

NCU will fix that and basically bring it onto par with more modern Nvidia GPUs. We're going to see a 20-30% jump in "IPC" over GCN4 in most cases. Vega's performance will come down to what AMD can get the frequency too.

>>59921256
I heard Maxwell has Async computing but its virtual/CPU heavy.

Aren't Pascal filtered this out. I saw gains in Vulkan with my GTX 1070 (Like 20-40 fps more)

>>59921256
is it's 1080ti approximate level for $600 i'm already going to buy it, I really don't need world records
just after how my 770 died from performance loss I really don't want to go nvidia

>>59921326
kepler had hardware scheduler maxwell and pascal don't, it's handled by driver
that's how they got to low TDP easily

>>59921368
No. Fermi was the last to have a hardware scheduler. Which is why it did not get DX12 driver support: it would beat Keplar in it.

>>59919613
if it were true they would have realeased this long time ago

>>59919749
1080 can run it at 60fps, 1080ti is 35% faster so around 100fps, while titan xp is another 15% faster so it kinda make sense, but i dont see vega being 140fps, il believe it, when we get legit reviewer getting this numbers (tpu for example)

>>59921409
>he thinks a company can get that much performance out of a chip as soon as it gets taped out

Lol no

>>59921199
So what? That means it's a metric of how many frames the CPU can produce per unit of time. Not the only performance metric, perhaps, but certainly an important one, especially if you intend to measure execution throughput.

>>59920729
amd wasnt btfo'ing nobody in those years, ppl are claiming amd ryzen is beating intel, but for gaming intel i5 is still getting more fps for less money, same thing with nvidia, amd cucks keep claiming 4850 was destroying nvidia, but 260gtx kept up with it and it similarly priced.

>>59921457
You can have 200 fps in AVG if your frametime is 40ms here and there you WILL suffer having a good gameplay

You can also have 150 fps in AVG and still end up having 5 fps min such as the i5 7600K in GTA 5 benchs -> ruins the game expirience

0,1% are extremely important indicators of what the game will actually look like since you know around what your frames can drop to and through what repetition

>>59921246
But it is faster in a few very specific tests

>>59921490
>You can have 200 fps in AVG if your frametime is 40ms here and there you WILL suffer having a good gameplay
Yes, but so what? All I'm saying is that average FPS is a very relevant metric for CPU performance. Subjective experiences completely aside, a CPU that can produce more frames in a given time performs better than one that produces less.

>>59921447
>lol sure, thats why amd keep delaying it, cuz they are refining drivers and shit, which magically boost its performance, even though, teraflops shows performance performance of manufacturers cars pretty consistentally. As in amd cards ussualy have higher teraflops than nvidia, but perform similarly, like 480 performed like 1060. while like 40% more teraflop.

> he thinks 12tf vega will perform faster than 12tf titan
You must be amdrone or complete retard

>>59921538
im kinda drunk so ingnore my errors in grammar, im also not american

>>59921490
gta5 is bugged when you get over 150fps, it bugs out and have huge spikes it been recorded by gamernexus, where they tested older i5's, which dont have this problem do to not being able to achieve this high fps

>>59921409
Writing drivers is tricky. Writing drivers for major architecture change is even trickier.

>>59921432
>tpu for example
nice bait.

File: 1490423862598.png (37KB, 450x500px) Image search: [Google]

37KB, 450x500px

>>59921475
>but for gaming intel i5 is still getting more fps for less money
>MORE FPS
>except when it keeps locking up for 100ms+ at a time and stuttering way under 60fps
>BUT THE MAX FPS IS SO HIGH!!!! bUY NOW IDIOT
uhuh

>amd cucks keep claiming 4850 was destroying nvidia
Hmm...

>>59921522
>All I'm saying is that average FPS is a very relevant metric for CPU performance.
Well you're wrong. It's not.
1% minimums are the most important.
Doesn't matter if you get 150fps if you keep fucking dropping under 60. A smooth 90+ is better.

>>59921326
>Aren't Pascal filtered this out.

No, Pascal is literally dieshrink Maxwell.

>>59921678
>Well you're wrong. It's not.
Did you even read my post? I'm not discussing subjective gaming experiences, but actual CPU performance.

>>59920608
Hey my dad works for analog devices.

File: perfrel.gif (45KB, 400x683px) Image search: [Google]

45KB, 400x683px

>>59921678
>find one game 4850 performance is better claims its all end truth
>actual over performance 260gtx beats even 4870
>amdesperate drone

>>59919613
do we really need to fall for the overhype meme so retarded shills can post le crying amd meme shitposts every day?

vega will be competitive. brand loyalty is autism and doesn't belong here.

>>59921688
just like polaris is literally die shrink 390 with slighty lower amount of cores, but high clock speeds, giving it exactly the same amount of performance.
>good job amd...

>>59920975
better for me, just want a $350 1070 performance card

>>59921862
Actually it's half-Fiji shrinked to 14nm.

>>59921882
Why does it perform so close to Fiji then if it's half of it?

>>59921918
It's tweaked Fiji. The original Fiji was very unbalanced. Fuckton of shaders, not enough ROPs.

>>59921882
same shit anyway, really most of the architectures dont differ to much other than some optimization, improvements, die shrinks. Take any company and its literally the same. AMD have been releasing same shit architecture cpu's for years and so was intel (expect they where significally faster). Even on phones its same story, other than die shrinks and mhz gains, actual IPC gains on snapdragons are minimal for years now.

>>59920928
>dx12 at all
enjoy the 12 games. it won't ever be the be all end all of video gaming, not as long as it's hard locked to one os.

AMD never change

nVidia will still reign top.

>>59921966
>not as long as it's hard locked to one os.
Oh gosh, i wonder if there's another low-level API out there.

>>59921918
Because Fiji was 1000-1050Mhz and now RX580 is nearing 1500Mhz.
And it handles tessellation a lot better.

>>59921947
>same shit anyway
Except not.
The 280X was a 7970, yes.
The 390X was a 290X, yes.
Polaris is not Hawaii. It is more similar to Fiji, like someone else pointed out.

>>59921966
Only 2 of those are DX12, dumb fuck.

File: firefox_2017-04-16_12-57-04.png (93KB, 1720x772px) Image search: [Google]

93KB, 1720x772px

>>59921966
These are just the Vulkan ones.
MS is pushing DX12 super hard now, too.

Nvidia's strategy of bribing devs not to implement it coming to an end.

>>59921986
Fiji is optimized hawai with more cores, its performance gains with amount of cores is consistent with R9 390 compared to fury x. Also its literally have same clock speeds. If you removed hbm it literally would be just bigger hawai die

>>59919722
stupidest comment ever

>>59922005
nvidia doesn't have to bribe anyone not to use it, the market doesn't dictate implementation blanketly because of windows 10's 50% gaming market. there's a whole other 50% that will continue to get dx11 legacy support.

and those users will not have to deal with the atrocities of windows 10 cucking them, while getting great frame rates.

>>59921862
>just like polaris is literally die shrink 390 with slighty lower amount of cores, but high clock speeds, giving it exactly the same amount of performance.

Not really no. Polaris has half the render outputs. The clock increase isn't big enough to get you the same performance as the 390 just with that.

>>59920049
like 1 week old, faggot

>>59922059
>there's a whole other 50% that will continue to get dx11 legacy support.
The others will simply use Vulkan. It's a win-win situatuin.

>>59920803
https://www.youtube.com/watch?v=RZS2XHcQdqA&feature=youtu.be
This kills the 7700k

>>59922067
480 have 2300cores clocked at 1300, 390 have have 2800 clocked at 1000.
They also literally have same amount of teraflops at around 6.00.
Numbers are very consistent.
Polaris is just a die shrink with mhz boost, just like Pascal was die shrink of maxwell with mhz boost as well. Vega should be new architecture, also Volta will be new architecture as well.

Literally the same performance from math perspective and from actual performance numbers. I dont even know wtf you talking about.....

>>59922145
That doesn't take into account the ROPs you twat, which are the single most important piece of hardware when you talk about gaming, and it's a fact the 480 has HALF of the 390

>>59922145
core count doesn't mean shit if you don't have the ROPs to actually draw stuff on the screen. It's why Fury scaled so bad, it had a fuckload of cores but they were sitting idle most of the time because it did not have enough rasterizing power.

Polaris can service 20% lower core count with HALF the ROPs, which means that it has much more advanced rasterizers.

>>59922209
You retard 1000 series also have a lot lower amount of rops as well like 1070 have 50% lower rops than 980ti, while having higher overall performance what now you fucking twat?
Fucking amd drone piece of shit

>>59919699

Nvidia marketing.

>>59922281
Im not even talking about 1060 vs 960 which have almost same amount of rops, while having 2.5-3x performance what now you piece of dog excrement?

>>59922281
Stop foaming you retard, the 1000 series run at significantly higher fucking clocks and ROP size is a thing.

>>59919613

WARNING

REVERSE SHILL THREAD

THEY HIGHEN YOUR EXPECTATIONS SO THAT YOU WILL "FEEL DISAPPOINTED" WITH AN OTHERWISE FINE PRODUCT.

>>59920729
>looks like a penis

>>59922327
yes just like rx 400 is running at significally higher clocks with lower amount of rops, 1000 series is running at higher clocks with lower rops, what now you fucking amd bitch?
1060 have 3x performance of 960 with almost same amount of rops and cores, what now? Where is your argument??? Fucking joker piece of shit

>>59922360
>higher clocks
>biggers ROPs
>better geometry engine
>higher IPC

Done, can you go cure your aneurism now?

>>59919722

>Nvidia still butt hurt

http://www.redgamingtech.com/the-vulkan-interview-exclusive-interview-with-the-khronos-group/

>Neil Trevett (Vice President of Mobile Ecosystem at Nvidia and elected President of the Khronos Group).

Nvidia is all out of excuses.

>>59922372
So pascal is maxwell but polaris isnt hawai, ye sure fucking thing retard

Fucking delusional indian curry shit

Wait, the 1080 only gets 60fps in doom vulkkkan 4k? LMAOOOOOOOO

>>59922360
Your autism is boiling mate

>>59922389

But Polaris isn't Hawaii, it's closer to Fiji looking at the frontend.

File: images?q=tbn:ANd9GcQqM_-EeP7wAGw4caxsuwd594TjA0P7jdpQzEpFjIgp5svvElYA.png (9KB, 223x226px) Image search: [Google]

9KB, 223x226px

>>59919722
>Goys! Look! The 970 beats the Fury X on Witcher 3 with hairworks and x64 tessellation! and look at Fallout 4 with ultra godrays!
>W-what? Vu-vu-vulkan? NO! NOT FAIR! YOU AREN'T ALLOWED TO USE AMD TECH!

>>59922406
atleast im autist and not some smelly paki mongrel piece of trash

>>59922350

It's the butt plug Nvidia has rammed up the industries ass to hold us back.

>>59922414
Fiji is same fucking thing as hawai
>it has 30% more performance with 30% more cores
WOW its literally not rebrand just like Polaris 20 is literally not polaris 10. Another curry eating and streets shitting indian talking

>>59922468
No it's not, for one thing Fiji uses the same power as Hawaii but is significantly faster.
Secondly their memory controllers are light years apart

>>59922498

Don't even bother with that retard. He's less informed then a YouTube box reading personality.

>>59922498
>>59922517
and drones think they know anything, piece of fucking curry shits

They had better fucking test it on Ryzen.

>>59922655
Ha ha no.

>>59922655
Can you imagine how Nvidiafags will respond when Vega tests are done with Ryzen and their 1080ti performance fucking tanks?

It'll be a bloodbath.

File: Download.jpg (10KB, 302x167px) Image search: [Google]

10KB, 302x167px

>Vulkan
what a fitting name for an AMD GPU

>>59920875
kek

>>59922655

I want to see the difference between the X and non-X variants as well.

>>59922693
That's an API you doofus.

>>59922693
vulkan is a graphics api

File: 02-star-trek-porn-sasha-grey-vulcan.jpg (93KB, 600x900px) Image search: [Google]

93KB, 600x900px

>>59922693

Illogical.

>2017
>caring about muh gaming performance

ITT: Manchildren

>>59922730
There is nothing logical about wasting time watching pornography, especially "parodies" of shows like star trek that will interfere with previous memories, making you dwell on sex rather than star trek any time you think back. Really fucking logical.

>>59922759

Sex is bad?

NO SOURCE

>>59919613
for doom this is 100% possible, doom takes advantage of untapped amd potential, while this is very impressive if true, i'm more interested in how its dx11 performance is when it will get fucked over by shitty devs.

>>59920797
not to mention, when something new comes around that is good for computers, gamers, or just better then its competition, amd usually has their fingers in it some way or another.

>>59923668
Fury X's performance isn't all that amazing in Doom though. >>59921038, even taking advantage of AMD's hardware

>more fake slides

when will AMD faggots learn to not fall for fake benchmarks (like '480 beats 980ti for $100')? can't wait to see the hype train derail yet again just like it did for POOLARIS and zen.

>>59920178
doom is a game that is made by programmers that fully deserve respect, they threaded the game so it runs on both amd and nvidia as a net positive over ogl, where nvidia's ogl drivers are on par with their dx11 ones.

doom shows off both architectures in their best light possible.

>>59923717
got a 1080p for that? I don't know what doom wants at 4k or what settings that is at, but I dont think its a good comparison.

looking into it, doom wants 5+gb of vram at the highest settings, fury has 4gb

doom is also like old school games where the max settings had noticeable differences, so going 4k would reduce image quality more so than you gain from lowered settings.

not only that, you have the hardware bottlenecks that are inherent in the fury, and are un workaround able process ones. I would like to see 1080p on I think ultra, that should illustrate what doom does and can do for hardware without adding a bottleneck for one card.

for doom the better amd cards to look at would be the 480 that doesn't have an inherent hardware bottleneck compared to the 1060 6gb

File: Screenshot from 16-04-2017 21:21:46.png (95KB, 598x453px) Image search: [Google]

95KB, 598x453px

>>59923887
Only 1440p, but it does notably worse.

>>59923744
you site fake slides, but then you mention slides that were based off clock speeds that the china guy got, and lo and behold when it clocks that high it does.

doom is the most hardware agnostic game on the market, and takes advantage of everything you give it be it amd or be it nvidia. at the very worst, vega hits 1525mhz and at that speed matches a titan xp flop for flop, doom is the one/one of the few games that you can look at flops and gauge performance, at minimum vega without hardware bottlenecks would match a titan xp at just matching the mhz of the passively cooled pro line.

NOW, the overwhelming thought is that when actively cooled, amd can hit 1700mhz or there around, the slide very well could be fake but its based in reality enough to be believable.

but i'm also assuming vega is halfway between a 1080 and 1080ti till amd shows it running better.

>>59919613
Human eye can't see.

>>59922349
i think this is actually amd spreading bullshit so people will preorder vega
also that slide looks just too perfect in terms of quality, all the other slides look like ass and have watermarks all over them

>>59923924
yea, fury x hardware bottlenecks itself shortly after you take away the software one. bit sad to see.

>>59919613
fake, I believe

This could be a pretty well made fake, considering what's going in Vega and its size, this shouldn't be all that hard on something as optimized as Doom

>>59919613
honestly i bet that's the dual GPU version. would look about right for a single GPU version to beat normal 1080 by a bit.

>>59920729
intel amd back then was more to do with oems, amd made great sales in the builder markets, gpu wise... yea amd is effectively locked out of the consumer market mcm gpus will help by allowing them to iterate an entire line with no rebrands every 6-12 months

AMD supporter's average IQ must be surely below 90.

They are as dumb as one can get.

>>59920819
nvidia gets net positive over open gl in vulkan, vulkan isn't the same shit as dx12 is.

>>59920952
thats not implying a release date, that's implying a 'you can share this' date. if anything, its a release of all the info date.

>>59919613
Clearly fake, but halfway plausible if D44M's postprocessing shaders were rewritten for 2x fp16.

>>59924058
Dual GPU versions are made like half a year to 10 months after the single version releases as they need to stock up on high quality bins

>>59921038
you know nvidia were retards right?

1080
titan x (what the press called it was titan xp)
1080ti
titan xp (the press now has no idea what to call it)

>>59924150
he probably means crossfire

>>59924170
XP & Xp

>>59921256
you are wrong.

amd needs api based threading, nvidia has a software scheduler so they can force games not threaded to thread to some extent, its the main reason why nvidia has as much a lead as they have with lesser raw hardware, they off load a fuckload of the process to the cpu while amd does not.

what amd needs is good software devs but in gaming id is really the only ones.

>>59924213
>id is really the only ones
Croteam i guess.

>>59919738
The future is now grandpa

>>59923998
the stars in the background could be different depending on outlet or place they sent it to so they could find out who is leaking shit and cut them off.

>>59924058
>honestly i bet that's the dual GPU version. would look about right for a single GPU version to beat normal 1080 by a bit.

nah, this is doom, a game that takes advantage of all hardware
the gpu is a gpu they developed focusing almost exclusively on dealing with bottlenecks
on a process that lets the pro version be passively cooled at 1525mhz

if amd dealt with bottlenecks, it would match the titan xp in performance, if they also got clock speed increases when actively cooling it, sky's the limit.

its an idealistic but within reason fake if it is one.

>>59924149
lol, if it was half precision then nvidia would fail to even launch the game considering that polaris has 50 times the half precision of the titan xp

>>59924129
Which always coincides with the release date.

>>59924292
no, amd is going to build a hype train for vega, especially if it is REALLY good.

part of building the hype train is leaking bits and pieces here and there over the coarse of months with a fairly big reveal but without a key question answered.

>>59924292
In recent cases NDA release and on shelves is different.

>>59919613
fake

>>59919613
Is TXP really 2x as fast as 1080?
Is any part of this pic realistic?

>>59921135
its not intel paying them, game developers are near universally dog shit programmers and should be fired.

when nvidia ditched their hardware scheduler they saw that when dx11 was properly threaded (civ 5) it crushes un threaded in performance.

because software is more flexible, nvidia can force threading while amd cant. though when something is properly threaded amd crushed nvidia due to nvidia having FAR more overhead than amd

>>59922759
I'd rather think about sex any time, than dwell on that amfattard aspie soap.

>>59923765
It is optimized quite well and sets a good example, but most games out there are optimized like shit.
I'd like to see how it will compare to Nvidia's GPUs at a wider range of titles.

Vega looks promising, but as with any unreleased product the hype tends to be overblown. Hopefully it releases next month and a few reviewers get their hands on some samples.

Been holding off getting a 1080Ti for a while now primarily because it doesn't quite max out 4K performance on most titles yet and costs and arm and leg here. I suspect vega's pricing will disappoint as well considering they are doing a low volume production run similar to the fury line while using new memory tech.

Similar deal happened back during the 300 series and fury debut.
I think next year might be a more suitable time for 4K price performance wise should HBM2 finally make its way to midrange.

>>59924432
No, should be more like 55% faster. But does seem like it might scale more than that at 4K.

>>59926441
Everything gets utilized more at 4k, the cores, the memory bandwidth, the memory controller, all of which the Titan XP has more than the 1080ti

>leak
Everytime

>>59919613
Will Vega offer competition to GTX 1070?

Fake

Aso doom/vulkan it like showing Nvidia performance on project car vs amd

I just buy the 1080ti, should I cancel it?

File: krypto-mainimage.jpg (49KB, 731x718px) Image search: [Google]

49KB, 731x718px

>>59919613
BUT CAN IT MINE KRYPTO?

>>59928523
Do nVidia cards perform better on Doom without Vulkan?

>>59921475
Look at the delusional fanboy.

Look at him and laugh.

>>59928967
No, the 10 series scales about as well as the Rx 480 on OpenGL vs Vulkan.

>>59920875
>didnt even see the last part
someone isnt getting sheckels tonight

>>59921678
>needing a dual gpu to beat amd
AHAHAHAHAHA

>>59919613
Nice.

Any DX12 benchmarks out as well? I hate it, but DX12 is probably going to be as, if not more, common than Vulkan.

>>59921704
And where do you work, anon?

>>59919613

This slide style is nothing like something AMD would use, especially not 20% transparency on the bars.

You need to learn how to duplicate visual style if you're going to make fake shit.

>>59920641
>Vega will beat even the 1080Ti at 4K so hard.

Into the collage you go. Can't wait for the tears on announce day.

>w-we only wanted 1070 performance!
>60 FPS is a meme! 30 is all you need!

>>59921724

But that's all AMD shills have EVER done. Cherry pick 1 game (Ashes in vogue now) and cling to it until it's 9 years old and look for a new one.

File: 1406108591955.gif (516KB, 500x240px) Image search: [Google]

516KB, 500x240px

>>59919722
> Vulkan
> made by amd

>>59920776

>b-but you gotta include the stutter during loading screens, it's not fair otherwise!

>>59919613
>doom
Call me when it breaks 60fps in any game that matters

>>59928523

Haven't you heard? It's absolutely okay when AMD uses partnered games for benchmarks, but don't you dare do it for nvidia.

>>59919613
Except the linx drivers for amd sucks, so no reason to not get nvidia for Linux gaming.

>>59919613
Estimated price per graphics card, power consumption per hour and Hashrate speed please? I'm thinking of using it to mine cryptos.

>>59921678
The 4850 WAS destroying nvidia. What you're missing is the prices. The 280 was like $400, while the 4850 was $150.

>>59919613
big if true. but there's no point in speculating over a leaked screenshot. Wait for some actual benchmarks.

>>59919613
NVIDIA IS FINISHED!!!! IT'S ALL OVER BOYOS!!!!

>>59930000

>>59924278

>the gpu is a gpu they developed focusing almost exclusively on dealing with bottlenecks

Then why the fuck does thier second fastest card has 4gb memory?

>>59919653
ryzen "leaked" too how did that turn out?

>>59931341
exactly how it leaked?
stop operating on day0 benchmarks

GPU is different story though, stop hyping this shit up, it's going to be 7-10% under 1080ti/titanxp first 6 months, then it will pull ahead but nvidia will release volta 2018 and Navi is 2019. Same cycle will continue.

>>59930088
>Haven't you heard? It's absolutely okay when AMD uses partnered games for benchmarks, but don't you dare do it for nvidia.

The difference is AMD uses all open technology that NVIDIA is free to take advantage of and write drivers for. The same came not be said about Gameworks.

>>59920775
>Once games begin to use more than one core, Ryzen will become more and more powerful.
bulldozer pls go

besides most of the bullshit core loading people have found is in nvidia's driver because of its meme software scheduler and other shit

>le insane high resolution benchmarks

Where's the reality?

>>59930681
good thing the 260 was faster than the 4850

>>59919613
>titan xp almost twice as fast as 1080
no

>>59920742
>but muh gayms though

>>59931251
because, they showed off the feature during the last conference they were at where they vram limited it to 2gb and played deus ex mankind divided at max settings and suffered almost no perceivable hit.

thats 2gb and on screen you couldn't tell

If this feature works as intended, it removes vram as a necessity, and more as a nice extra, and allows lets say an 8gb card to last FAR longer as the max everything card, where as the moment a game requires more then 8gb of ram currently, everyone not on a titan x, xp 1080ti would have to reduce settings.

This also allows for a 4gb full vega to be a viable option, I mean /g/ cant shut the fuck up about how expensive hbm is, this allows one die to be very useable apposed to 2+

>>59930030
Share it anon. Can't wait for another massive disappointment.

>>59919613
Now, for once, lets just assume these leaks were true. Don't think about the specs and whatnot, let's just assume. How glorious this would be...
...but sadly this is never gonna happen.

File: 1487720349817.png (492KB, 900x600px) Image search: [Google]

492KB, 900x600px

>>59919613
DELET

File: 1491930562423.jpg (395KB, 2400x2000px) Image search: [Google]

395KB, 2400x2000px

>>59920742
>day 1
meanwhile in reality

>>59919613

price?
red staple cherry picked bench
HBM2 vs GDDR5
power consumption and temps, oc headroom

>>59930058
They used Mantle as a starting point when drafting up Vulkan.

File: C5QPfezUoAE4r-1.png (690KB, 914x898px) Image search: [Google]

690KB, 914x898px

>>59920688
>Intel being BTFO
>Nvidia being BTFO
>2017 still isn't the year of the Linux desktop though

>>59932360
>Intelaviv shills about to swallow the AyyyyMD redpill

they must be suffering so badly right now

>>59932476
becuase amd is more viable then autism: the os

let that sink in for a second

>>59932490
OY VEY ANUDDA SHOUA

>>59930088
AyyMD makes their tech open, nYidia makes gimpworks.

>>59932360
Oh hey please use less demanding area for your GTA V

>>59932546
what like starring at a wall until some sheenie magic makes intel better?

Why do people get so riled up over AMD vs Nvidia? How do those corporations affect you?

>>59932735
nvidia is jewish and i hate jews

>>59932784
That's weird. Seems like Nvidia is an american company and is not in fact based in israel

>>59932784
how Nvidia jewish, i thought it just intel

>>59932735
they don't. amd have this corporate funded shitflinging scam going on where they pay shitty youtubers like adoredtv to spread hate towards anything that isn't amd. the only reason people get riled up is over all the misinformation that these people like adoredtv spout. he literally just made a video about how bad nvidia are when it comes to rebrands back in 2002 yet he failed to mention the most recent amd rebrands from literally 2 years ago being the 2xx and 3xx series. if you're going to talk about something at least give the complete and unbiased picture. he gives amd a free run but hates on everyone else for doing the same thing.

this is what amd are paying people to do, and it's truly sad.

>>59932899
>adoredtv
As amdfag I also hate this pajeet

>>59919613
>In one game
>On an API optimized for AMD
>At a resolution where no monitor can even display this amount of frames anyway

Irrelevant/10

File: image.jpg (247KB, 796x805px) Image search: [Google]

247KB, 796x805px

>>59919613
DELET DIS

>>59932899
>"Why do people get so riled up over AMD vs Nvidia?"
>Proceeds to shill for nvidia

>>59931686
Ryzen reduced Intel to two skus, shekelchaser. Please feel free to kill yourself.

File: kizuna_ai___face_expression_test__sfm_by_lemon100-db0frmf.png (1MB, 1080x1920px) Image search: [Google]

1MB, 1080x1920px

>tfw Vega is only tested on a 7700K on 90% of tech sites.

>>59932899
AMD rebrands maybe once a year but usually two years. nYidia rebrands nearly every fucking quarter.

>>59919613
So do you mean AMD finally fixed stuff and DOOM actually runs with the latest drivers?
Really makes you think

>>59919613
>leaked
>AMD slides
You retards wouldn't trust promo slides from nvidia or Intel, but you'll cream your pants at ones from AMD. You morons are such bad consumers.

>>59932784
>Intel AMD and NVIDIA are all American companies with manufacturing in China or Taiwan
>somehow they're jewish
/pol/ everyone

File: 1420865723416.jpg (331KB, 1279x1020px) Image search: [Google]

331KB, 1279x1020px

>>59935081
>t. Goldstein

>>59919613
if Vega does this I will actually buy it day one for $700

>>59933193
>API optimized for AMD

File: 1491487687825.jpg (71KB, 552x661px) Image search: [Google]

71KB, 552x661px

GOOD JOB AMD

>>59932360
All that still dipping below 60fps at 2k. I thought 2017 was supposed to be the year of 4k and vr.

>>59920742
Intel Shilling

The 4c8t beats the 8 core which beats the 10 core which beats the 6 core which beats the 4c4t.

> All of them using the same dye
> makes sense.

>>59920928
>i5

For only 999

>>59919613
>leaked
Just wait for real benchmarks.

This shit is fucking stupid.
If it's that good, we'll know soon enough.
If it's not, overhyping it now bullshit leaks will only make it look bad.

OP you're an moran

>>59936105
>all of them using the same die
No.
115x i3/i5/i7 share the same die.
for 2011-v3, 4/6 core share the same die, and 8/10 core share the same die.

>>59929839
I'm in college

>>59919613
Although I want Vega to succeed, this chart is faker than the Holocaust.

>>59936425
Former Holocaust supporter here, I'm really glad to see that people are starting to believe it wasn't real. But seriously we can't let people forget the 6 million.

>>59919613
> Another AMD-boi's wet dream.

Jesus, just wait for actual benchmarks. How many times are you going to post false shit and then be disappointed by the reality of AMD.

>>59936425
>Vega supporters are Holocaust denialists
wtf I love vega now

How much Vega will cost?

can someone explain fanboying? shouldn't anyone who likes amd or intel products be happy that the competition is strong? we as consumers reap all the benefits of the back and forth battle. Better products, lower prices etc. Some of the shit I read on here is funny as fuck. Why would anyone want AMD or Intel to go under? That is retarded.

Benchmark is probably sitting on the menu screen or staring at a wall or something.

>>59937431
>need a new GPU
>GTX 970 advertisement looks good, gayming yeah
>everyone tells me to get the 290X
>fuck AYYMD housefires
>2 years later
>GPU is shit
>everyone insulting me for getting the 970
>can't hold the pressure
>buy new NVIDIA top gear GPU
>NVIDIA still NO. 1
>AMD fags BTFO

>>59919613
Tottaly unbiased coming from AMD themself.
Also Vega is going to show up when?
What was the original planed release end 2015 to mid 2016?

>>59934825
>tfw Vega is only tested on a 7700K on 90% of tech sites.
:(
It'll happen. Despite Ryzen+AMD GPU working so well together.
In many cases the 7700k will still probably be better, though. But I'd like to see both.

>>59937670
AMD hasn't put out a benchmark of their own products that wasn't real, at least since I've been paying attention for the last 5 years

>>59937687
I hope Ryzen doe snot hold any performance advantage when combined with Vega. Not ready to ditch my x99 yet

>/g/ being consistently wrong in predicting what AMD will do

Increasing their shares by nearly 10 times in a year is a sign of bankruptcy is /g/ is to be believed.

>>59938156

Infinity fabric.

In every Zen CPU and Vega GPU. Read up on it.

>>59937687
>In many cases the 7700k will still probably be better, though.
in what metric? Minimums hell no, Ryzen does better nearly across the board. Average? Maybe, but they're neck and neck there. Max? Well yeah, but who the fuck cares if your 1% and .1% lows are atrocious.
Smooth gameplay > very high FPS with bad stutters

>>59938652

Can we officially add minimums to the list of things that no longer count for cpu performance?

>>59938749
>minimums don't count anymore because they aren't in intels favor.
10c has been deposited into your account.

>>59938977

Don't forget to pay me for the following-

>cinebench multi doesn't count
>cinebench single doesn't count
>blender doesn't count
>AES encryption doesn't count
>power draw doesn't count
>IPC doesn't count
>high resolution doesn't count
>only clockspeed counts

>>59938156
>I hope Ryzen doe snot hold any performance advantage when combined with Vega. Not ready to ditch my x99 yet
It's minor. <5% advantage generally. And x99 still wins sometimes.
Might be noticeably smoother and lower latency on Ryzen too but eh. >>59921203

>>59934975
So I'm not the only one who can't start it
How do we fix it

>>59935762
Ooooh 'ZYKLON Ben' oh you...