Is AMD driver overhead just a meme?

File: hello.png (208KB, 640x409px) Image search: [Google]

208KB, 640x409px

>>55369744
What do you think would happen when you compare a fast processor to a slow one?
I dont know, maybe Nvidia has a driver overhead problem as well.

File: Screenshot_2016-05-12-18-03-32.png (2MB, 1920x1080px) Image search: [Google]

2MB, 1920x1080px

>>55369744
No. It's an actual issue.

>>55369848
>hairworks.
Useless shit that would bog a 5960x down
Thanks nvidia

>>55369744
>>55369848

It's not a meme.

Look at one of the Digital Foundry videos where they tested Just Cause 3 on like a 390.

It would keep up with a 970 until something exploded. The 970 would keep chugging, but the 390 dropped a lot of frames and the frame-time spiked.

Although, from the tests so far, the new RX 480 is improved.

>>55369744
Not just a meme, a Nvidia shilling meme

AMD's OpenCL implementation performs better than Nvidia's. And no one plays games or uses proprietary software.

>>55369744
>>55369848
>>55369888
>>55369894

Here it is, watch at 1:50

https://www.youtube.com/watch?v=4JHuyLoYU4c

it's also been documented in a lot of other games.

>>55369888
Are you legit stupid or trolling? It doesn't use hairworks. In actual fact it uses AMD PureHair.

yeah it's legit. most reviews don't show this though because they use shit like overclocked 5960x CPUs and similar setups that you won't ever see in real world usage unless you're rich and like to waste money

http://forums.guru3d.com/showthread.php?t=398858

File: 1445391083672s.jpg (4KB, 125x125px)

4KB, 125x125px

>>55369953
I don't know why this game is popular, but whatever.

Here it would be like on my system. Steady 60fps. Not exactly a rich setup.

https://www.youtube.com/watch?v=X_zASxtpRn4

Enjoy your 3.5.

GTA 5 on AMD seems stuttery at times.

>>55369744
yes

>>55369953
I never understood why they did that. I want realistic gpu performance, not ideal gpu performance

>>55370003
What card?

>>55369934
pure hair has long gone away from amd hands since the first rotr the devs said they are "created" their own version of it to ship it as a third party

>>55370008
It removes the cpu from the equation. Do you really expect all your favorite titles to be tested with 5 year old cpu's, or bargin bin AMD processors? You faggots hate upgrading, but hate your reduced performance, blaming it on others.

File: Hairworks.png (267KB, 798x460px) Image search: [Google]

267KB, 798x460px

Hairworks is shit, even on Nvidia products.

This is why AMD is pushing for DX12 so hard, they have give up on DX11, when you use a low level api you give up a lot of support and debugging tools that are inherited to a layered system like DX11, not to mention that going the low level api path you're risking breaking compatibility even within the same low level api once overhauls are done to the hardware, meaning that once new hardware is released the low level apis AND games will need to be updated or it will break compatibility, what are you doing to do once DX12/Vulkan becomes popular and hardware changes? games that are old will become unplayable unless the game developer stills supports it, Nvidia/AMD won't be able to change the source code for the game and the amount of work they can do on the driver side is very limited compared to DX11, there are huge implications and sacrifices on moving to low level apis specially with some many hardware revisions.

>>55370047
>It removes the cpu from the equation.

All it does is artifically boost AMD's performance in benchmarks. Nobody is pairing a 380 or 390 with a 5960x @ 4.5ghz with a $400 custom loop.

>>55369894
Can u explain frame-time to me, I don't get it

>>55370021
Hd 5450

>>55370095
I'm still just waiting on whatever point you are trying to make. Turn hairworks on, shit performance, on all cards, with all hardware.
Thanks Nvidia.

>>55370089
No shit retard, those are meant to be used by the %1 top of the line hardware, then once the technology matures combined with improved hardware it will be available for the mid range user, do you even know how technology advances you need to invest first to reap benefits years later.


Someone has to do it even if you don't like it and AMD keeps playing the emo card about how unfair it is, if AMD thinks it is unfair they should invest the money and developer the technology and push it to the public and try to push forward instead of complaining and whining.
Open source or closed source doesn't matter, what matter is that the investment is being done by someone, either Nvidia/AMD or game developers, in the end we all win.

>>55370095
This is correct. My friend has a moderately cheaper build with a 390 and fx 8350 and when we play games like dayz or even some less cpu intensive games like dying light he always runs into issues due to that overhead. Like when he would come out of a cutscene on dying light his fps would tank like shit when it's re-rendering the world.

>>55370132
Are you homeless? Buy a modern card you poorfag NEET.

>>55369894
this has nothing to do with the drivers..................................
using gameworks on a game always results into lower framerates on amd simply because amd till 480 didnt had any way to cull sub pixel triangles
but now they have a primitive discard acc that is complete configurable

>>55370021
280x

>>55370147
Quit comparing benchmarks with useless visual fluff that do nothing to improve gameplay. >>55370148
You friend is a stupid fuck for buying a 8350, You have no reason to complain.

>>55370185
It's not about gameplay, is all about image quality ;)

The envelope must be pushed by someone.

>>55370169
swap to DX 10 or 10.1 will fix the issue. The API handles it differently in that game.

>>55370185
People with 970 and fx 8350 don't have that problem. I've checked many youtube videos.

Let's stop the denial already. Amd has higher driver overhead compared to nvidia and that can't be denied.

Guys, if I start a review site using weaker shit like i5 2400 or i5 750 would you guys take it seriously?

I have some old hardware laying around I'm tempted to benchmark out of boredom, and if it's successful I might add new shit like rx470 to it.

>>55370207
Haha holy fuck. Your image quality, (that you will only notice in static frame shots) is compromising your gameplay by crippling your frame-rate, and this is somehow progress?
Nvidiots, everyone

>>55370164
Every now and then these gpu threads actually have an interesting post where i learn something and this time it was yours, thanks

>>55370228
Use youtube, there are a lot of users that are always looking for actual gameplay videos on low end hardware.

>>55370232
Sacrifices are needed, if you aren't willing to make them then move aside and then the big guys work.

>>55370121
60 fps means a frame need to take at most 16.7 ms to render.

EACH FRAME.

But lets say you have an unusual load in your game. One frame renders in 6.7ms, but the next frame renders in 26.7ms.

Averaged out, that's still 60 fps - (6.7+26.7)/2 = 16.7ms aka "60 fps"

But since the last frame took more than 16.7ms to render, it would be presented late. The first frame would be presented twice while the longer frame is still rendering.

That's basically what frame-pacing/ frame time is.

So you want a nice and smooth frame-time graph of 16.7 ms or lower.

>>55370207
crysis did pushed the envelope because they knew it had to be done
gameworks just fucks with amd cards and nothing more

people already compared hbao+ with vxao and they literally find nothing better on vxao that is almost 8 years newer filter..........

>>55370254
Thanks, will give that a try

>>55370221
Oh yes, you are an expert.
They are equivalent cards, is this hard for you to accept?
https://www.youtube.com/watch?v=vSDQzlKDYq4

Your driver overhead is a meme.

>>55370274
It'll also be cheaper since I won't have to rent a site.

>>55370266
Thank u. For the viewer, is this what stuttering is?

>>55370232
Your entire argument betrays you, AMD expend 5 years downplaying tessellation and now that they released a architecture overhaul with improved tessellation performance they are even advertising it as a feature.

>>55369932
As much as I prefer AMD cards, this is why people buy NVIDIA.

They nail the little things even if their card is a bit slower.

File: 1447814157516.gif (1MB, 570x321px)

1MB, 570x321px

>>55370296
Oh my god you are a pajeet, your english betrays you.
How much do you get paid for this? Does is upset you knowing you are not influencing anyone?
Tessellation is a meme, especially at x64 levels. Totally unnecessary.

>>55370288

>i7 4790k

>he thinks goyim will be matching their budget GPU with a $400 i7

Does the rx480 suffer from it too?

>>55370321
Dude, what?
Tessellation is progress.

>>55370288
>I7 4790k

Welp there goes your argument out the window.

File: Cape-Benchmark-StandardTessellation.jpg (271KB, 1920x1080px) Image search: [Google]

271KB, 1920x1080px

>>55370296
yes what you would have done when the competitor just uses tessellation as a weapon
its like that they say "oh we are dx12.1 cards" like people are so stupid that they dont know that they barely can go dx12 let alone having a full dx12 card..

>>55370294
>is this what stuttering is?

Yes.

Average frame-rate, aka "fps," is just that, an average and will hide (not quantify) stuttering.

Think of fame-time as the instantaneous frame-rate.

ALL stuttering will show up on a frame-time graph.

>>55370321
>Tessellation is a meme, especially at x64 levels. Totally unnecessary.

'64x' tessellation is just 16x with 4x MSAA. Not unnecessary at all if you care about your game not being a jaggy mess, though perhaps AMD cucks perfer a degraded experience.

>>55370296
>>55370328
>>55370330
>>55370336
N/v/idiot shills belong back on /v/
>>>/v/

>>55370321
x64 on nvidia cards doesnt mean true x64...remember nvidia cards till today they could cull those pixels in a way it was like a true x32 for them but on amd a x64 was an x64 because they didnt had any tech (a stupid decision from the former ceo)to cull the triangles

>>55370341
Thank you

>>55370328
970 and 390 were midrange card that anyone that bought could afford a 4790k.
But here is an i5 for you poorfags, still keeping up perfectly:
https://www.youtube.com/watch?v=04fRG5UixFM
>>55370330
It is literally useless at the x64 levels Nvidia is pushing

>>55370352
>>55370330
>>55370321
>>55370296
>>55370232
AMD Firepro Q9100 32GB is best graphics card.

No doubt.

>>55370352
see >>55370358

>>55370369
fuck, typo.
W9100

>>55370339
I can't tell if you're actually thick or just pretending. Nvidia cards support dx_1 and if you were educated enough you'd realise that dx12 is build on different tiers. Dx11_3 and dx12_0 require async support which is why amd 300 series is classed as a dx12_0 support. The maxwell cards which support dx12_1 support async (to a shitter extent but support is still there somewhat) and also conservative rasterization (which amd 300 series doesn't support). So maxwell does in actual fact fully support up to dx12_1.

>>55370368
Lots of people are still using i5 3470 and earlier.

You don't upgrade to the latest cpu every year...

>>55370368
>thinks everyone is using i5

Check the steam hardware survey. Everyone uses 970 with shitty cpus most of the time.

File: 11.png (72KB, 1022x584px) Image search: [Google]

72KB, 1022x584px

>>55370390
Quit bitching about driver performance when your cpu is 20% behind anything on the market.

>>55370336
Here is with an i5, it still stands

File: 1446932679415.jpg (85KB, 400x535px) Image search: [Google]

85KB, 400x535px

>>55370397
So people that buy shitty cpu's buy shitty 970's? Got it.

>>55370386
Writing a driver to accept Async commands but ignore them isn't quite "support."

>>55370386
At the hardware level, or just a software implementation, or hybrid?

>>55370403
Why? I won't have this problem with nvidia, why should AMD be allowed to do this?

I'm actually on an i7 870 which is plenty for me, as long as I don't use AMD.

>>55370440
You would have more of a problem with the horrible driver overhead Hairworks takes on your shitty cpu. Complete denial over the trure reasons of your poor performance: old ass cpu.

>>55370386
lets be clear here
dx11.3 was the last dx version of dx11
dx12 is dx11.3 with literally the mantle
dx12.1 NEEDS to have a full compliance over dx12 because it doesnt ADD anything new it just expands the dx12 envelope to full heaps

the maxwell cards never supported async the cuda cores do but if nvidia activated them they would have a hard time explaining as to why their power usage would skyrocket by using the hardware sc they have for the cuda core..
yes infact but dx12 isnt only rasterizers is it now(which they do have on tier 2 only) and since they dont have any pixel shading even on pascal nor do they have any dynamic core to use resource heaps id say they are far from being called even a full dx12 card.. let alone a dx12.1....
not to mention that the whole dx12.1 started right after august 2015 when people shittied on them because of the async fiasco

>>55370468
Then disable hairworks?
It's so simple.

>>55370403
https://youtu.be/XoKu0_2ozAc?t=18

File: slav.png (184KB, 1035x686px) Image search: [Google]

184KB, 1035x686px

>>55370505
Why don't you fuck off with your old as 770 slav videos. I thought is was a 390 970 showcase here?

>>55370486
bu but hairworks is da future
it gives outstanind hair!
ITS THE FUTURE
literally 3dfx had something called boobjobs back then and they didnt even cared to have it on their boxes as a feature..
and now we get pubic hairs

>>55370360
Why does it kill performance on Nvidia cards, also? Driver overhead?

>>55370544
Are you fucking retarded or what.
Even if you ignore the 770 you still have Mantle vs DX11 on the AMD side.

You can see the 390 DX11 getting an average of %60 GPU usage even a low settings.

>>55370568
I fail how to see how this is a problem when the 390 clearly is better even at it's 60% vs all the other competition.

>>55370587
Holy shit what?

File: driver overhead.png (186KB, 1045x694px) Image search: [Google]

186KB, 1045x694px

>>55370568
muh driver overhead

>>55370587
It's gone from

>driver overhead doesn't exist
To
>driver overhead doesn't matter

Lmao these fucking retards

>>55370606
https://www.youtube.com/watch?v=S546TL2LWNY&feature=youtu.be

>>55370606
>I7 4790k

Another opinion thrown out the window.

>>55370553
Idgi

>>55370615
failure to detect sarcasm is a common characterization of autism.

>>55370625
I'm unsure of the point you are trying to make. Is it:

Dont buy a slow cpu

or

Dont use an old cpu

and expect to have optimal performance?

>>55370606
>gtx 970 average is below its minimum
wewlad

>>55370660
or don't buy AMD and you won't have to upgrade your CPU.

>>55370561
because gameworks affects nvidia less than affects amd
look at project cars the dev knew this is going to happen and yet he used a version of physx that
1) you couldnt disable it
2) was an older one that was running on only the first thread of the cpu
3) he offloaded everything conserning the filters of gameworks into the cpu while you had a amd card
he literally made it unplayable a 950 was passing a 390 and he thought it was fine and when we called him out he started to ban our accounts and gives us refunds
i dont want a future that we will choose a game depending on our gpu...this would be VERY BAD for everyone..

>>55370671
No one that has a fairly recent cpu will get horrible performance out of any gpu they buy today. This is really a non-issue, unless you want to run at ultra with all the eye candy. Quit the dramatics.

>>55370689
Too bad, this is the path AMD is going with DX12 they are trying to push their hardware on everyone and from now own you're going to get games that run fine on AMD but worse on Nvidia, fine on Nvidia but worse on AMD.

Welcome to AMD's vision of PC gaming.

>>55370660
Don't buy amd at all is the solution. I can run an nvidia card on a 5 year old cpu just fine without getting bottlenecked in cpu intensive scenes like >>55369848 because of amds driver overhead. If I had a 2500k right now and a 390 I would be getting fluctuating frame times compared to if I had a 970 in cpu intensive games. Simple as that.

>>55370698
>i7 4770k
>Not recent


Simply impossible to have a discussion with you, you lack basic logic.

>>55370689
You don't blame skeevy game developers, you blame the hardware manufacturer that had nothing to do with it?

I don't care because I have a i7 3770k. AMD could be better but they're not as terrible as people make them out to be. Nvidia is jewing everyone hard and they need to be stopped.

>>55370296
a feature that preemptively discards useless tesselation isn't exactly flattering the awesomess of tesselation.

>>55370730
Impossible to have a discussion with you when you only:

>i7 4770k

and expect to make any scene at all. Do you want to show me something to make your point?

>>55370729
It would be perfectly playable for having a 6 year old cpu in it. More dramatics.

>>55370296
Didn't AMD introduce dedicated tesselators in the first place?

>>55370756
*muffled linking park plays on the background*

>>55370767
>i-it would b-be perfectly playable
>literally ignoring the solid evidence showing it's not prefectly playable >>55369848

Neck yourself

>>55370774
I don't listen to Linkin Park. Or any kind of edgy music for that matter. Nvidia is charging too much money and is involved in too many scandalous business practices, such as forcing tessellation in heavy games to lower the performance of their previous gen cards, or straight up LYING about the specs on their GPUs. I'm not saying AMD is some benevolent corporation, they just got caught jewing out their reference cards with 6-pins instead of 8-pins, either to save money or to try to wave that "muh power consumption" ePeen, but they are nowhere near as bad as Nvidia, and their cards are fine. There needs to be a more healthy balance of market and mind share between the 2.

>>55370841
AMD isn't an option until driver overhead is resolved.

>>55370468
I get okay performance with hairworks fully enabled in tw3 on i5 750 and Gtx 980 and all settings maxed. Over 30 Fps!

>>55370783
I am glad you are poor and have to worry about one frame of one game with one card with one cpu to have this much anxiety over other peoples graphic card choice. The 390 is really the better all around performer to the 970. I'm sorry you can't enjoy things the way they are meant to be played.

>>55369888
hair works generally looks better then normal hair, the question is if the performance hit is worth it.

personally i don't think so.

>>55370841
t. Adoredtv bumlick

Reddit is that way --->

>>55370841
Amd problem is worse.

>>55370850
3.5 is never an option

https://www.youtube.com/watch?v=ZQE6p5r1tYE

>>55370857
>can't prove me wrong
>i-I'm sorry...
>waaaa stop being right
>"Nobody has hardware older than 2 years and I know this because amd told me"
>waaaa

Neck yourself.

>>55370851
I have hairworks(tm) maxed on tw3 and get 60 fps with the 1070

it's a stupid gimmick though

>>55370885
Prove you wrong about what? Hairworks gimping older hardware? You are proud of this?

>>55370850
Or you have a good CPU.

>>55370862
I don't watch AdoredTV or browse Reddit. I'm not saying driver overhead isn't an issue, it's just not an issue for me. I use my wallet to vote for what I believe is better for myself and everyone else in the end, but if you disagree, that is fine, it is your choice to do whatever you want with your money. I am explaining my perspective on the matter.

>>55370896
What the fuck are you talking about? Who the fuck is talking about hairworks? You said >>55370857 and I replied.

Neck yourself.

>>55370889
Looks pretty good to me.

>>55370266
async displays also make this seem fucking smooth, this is the thing i wasn't able to wrap me head around before i got to test an async display as VERY low fps. the display i used was a free sync one, cant remember the number for you to look it up, but it was advertised 40-144, however a fuck ton of people noticed it still works at around 20fps, some put it at 18fps some at 23 at the lowest it can handle.

point it, downclocked the card to test a worst case scenario load in battlefield 4, got 23-28fps, the game still was smooth enough to play, something that without freesync just would not be the case.

not everyone has 500$ to toss at a monitor, but freesync is worth it if you can get it and the range is low.

>>55370341
micro stutter is different, you see this in multigpu setups where one gpu is going to show you a bit that came before the current frame was rendered,

>>55370914
You trying to showcase one specific frame of one game with one cpu to show this fabled "driver overhead" meme. Its funny how shallow your argument it.

>>55370169
Bruv i have 270x and best driver for gtav for me was 15.7something though fo4 sucked a bit with that one

>>55370862
>>55370885
Haven't you been told to go back to /v/ yet nvidia shill?
>>>/v/

>>55370903
Or you just buy nvidia for progression within the industry instead of replying on a company who rebranded a card for like 4 years in a row and can't even develop an architecture that's better than their competitors 2 year old 28nm based architecture. If amd were in the lead right now we'd all be fucked.

>>55370164
amd gpus on dx11 have something like 1.1 draw calls per second, while nvidia is able to handle 2.3~ i don't know if this was a specific game or what the scale was benchmarked or if its across them all, but this is what we mean when we say driver overhead.

with dx12 and vulcan... here
https://www.youtube.com/watch?v=Fs9fzxeNkm8

don't know how good this is, but it should show off what i'm talking about. draw calls are a bottleneck for pcs, close to metal apis can pull more draw calls, its why in prior console gens hardware that was similar to a pc at the time would be overshadowed, or even overtaken and what use to give consoles long legs.

currently this can be mitigated with a fast cpu on amd side, but being able to get that driver overhead down would speed the cards up and improve the playability on them by a significant amount.

>>55371005
Progression? Nvidia literally just shrank Maxwell and added some new features like software asynch compute, which is trash.

>>55370983
>1 specific frame

Are you to retarded to even read the fucking graph properly?

>one specific game

You think there is only one cpu intensive game? Even the people who made the video said it was because of amd driver overhead. You think they're lying? Of course you think they're lying because you're braindead in the head.

>>55371039
so dx12?

so AMD hardware is actually going to be better with future standards?

>>55371005
AMD innovates as well. I'm not saying AMD should be in the lead or is better, nor am I saying Nvidia should not exist or offers nothing of value,
but 85/15 is very bad for everyone. Nvidia is getting away with murder. It can only get worse with time. Refreshing an entire series is bad but the cards themselves are still fine, I don't understand these exaggerations. AMD and Nvidia both make good cards who's blunders frequently get blown out of proportion by fanboys and trolls. My only problem with Nvidia is LYING about specs and paying devs to make their card perform worse on previous gen cards. I personally cannot support these business strategies.

>>55371055
>I know nothing about pascal or its multiple hardware advantages which make it drastically different to maxwell

>trash

You mean like polaris which is fundamentally flawed as an architecture? Enjoy frying your motherboard.

At least we concluded you're a shill and not someone whose buying because of muh industry and mind share. Everyone shows their true colors with a small push.

I hope Nvidia releases the 1060 at a price of $250 to finally kill off AMD and then someone else can buy their assets, image if it was Samsung vs Nvidia.

AMD as a company is okay but the fanboys really ruin it.

>>55370296
>>55370773
amd did tessellation first before it was a native api thing, and when it was going to be in dx10 they had a card ready to go, then nvidia threw a fucking fit to microsoft till they took it out of 10.

when tessellation was re introduced into dx11, nvidia had what, 4-5 years to get their shit together and make something, and amd still beat them, then with the 500 they beat them a bit, then with the 600 by a bit more, the 700 was able to beat it by a fair amount, then with the 900 nvidia was a dx11 asic, and could tessalate like a motherfucker to boot so they used tessellation as a weapon, crippling prior gens in their cards to hurt amd a bit more.

>>55370321
its not a meme, however its used in meme ways, look at a barrel, that could use tessellation, model it with 8 sides and have tessellation expand that to effectively perfectly round with tessellation

but no, its ALWAYS used in the most retarded ways possible, nvidia could literally fucking shit on amd in a way that no one could deny, nvidia looks better, but they never fucking do, they ALWAYS use it in a way that makes everyone ask why bother.

it hurts everyone when they don't force each other to greater heights, amd should not be able to get away with a tessellation limiter, but because 64x literally does nothing even at 4k for image quality, you can bump it down to 16x and see no difference, and 8 to see a slight difference but even better performance.

>>55371067
Well we should all be glad when this dx11 trash is gone. It was really the cause of all your driver overhead you are so obsessed over. I'm sorry the developers of that game are trash and dont care about your specific experience.

>>55371086

Fundamentally flawed? That reference cards are shit? Or that it is the only card fully dx12 capable? That's something nvidia dosent have. The future.

>>55371073
yes, AMD GCN actually doesn't make sense for DX11, it's impossible to utilize it fully in high level APIs. But we won't see any great result until devs kills OpenGL and DX11 support for their games. Right now AMD performance boost in DX12 is more due to better driver and less about API itself

>>55371098
I've been saying this for ages. Amd need to die or sell of ati to samsung already. I'd actually buy a samsung gpu since I've found them reliable like I have a samsung phone, laptop and tv and it's not because I love samsung its because these were the highest recommended products at the time I purchased them.

>>55371086
N/v/idiot shills belong back on /v/
>>>/v/

>>55370266
frame syncing makes this not matter as much, screen gets a frame when the frame is ready without possibility to miss a screen scan. I think freesync/gsync is the most important GPU tech to come out in the last 10 years.

>>55371104
>with the 500 they beat them a bit
>tessellation was used in crysis 2 to totally gimp the performance on AMD cards
You can only pick one.

>>55371120
>being this much in denial

From your posts everyone can tell you're blindly shilling for things you don't even understand.

File: 1462660091644.png (91KB, 1716x441px) Image search: [Google]

91KB, 1716x441px

threadly reminder even amd owners hate amd

>>55371187
It hurt AMD cards more than it hurt Nvidia.

>>55369744
What do you expect, AMD is practically broke at this point, they have no money to optimize this shit anymore

>>55371005
DX12\Vulkan
HBM

vs?

inb4 CUDA, OpenCL is a thing, arguably works better.

>>55371208
All these
>(you)'s
tho

is it me or 90% of all these bencmarks are useless for plebs? 3d mark my ass
most users don't have 6core intels/i7, any oc or even 4k monitor
and sometimes there are not that big difference in graphics at high vs maximum but there is huge difference at fps
implying you would notice how good looks grass driving over 200kmh
i really like videos like from digitalfoundry with 2500k

>>55371221
I don't understand, what's the problem?
I went from an r7 260x to a gtx960 and I enjoyed both products...

>>55371280
>dx12
Enjoy upgrading to a botnet

>vulkan
Current being contributed to by nvidia engineers the most. Hell, the ceo of the khrobos group is also a head at nvidia.

>hbm
Saying amd invested this is like saying sony invented blu ray. Multiple companies were involved with hynix doing 90% if the work.

There isn't even any point in me mentioning things nvidia contributed to such as in the enterprise space because I know you'll just full on deny them without even refuting them.

>>55371388
Khronos *

Invented *

>>55371281
>he still can't argue against my solid evidence

i guess i've won here chaps.

>>55371334
This.

And the i3 vs i7 videos

>>55371388
>Enjoy upgrading to a botnet

No, it's not you fucking faggot.

>>55371388
>botnet
Without Mantle there would be no Vulkan. nVidia develops nVidia extension for Vulkan, just like they did with OpenGL.
>hbm
it's a joint venture.

You're basically unsuccessfully trying to nullify AMD accomplishments without providing nVidia ones. Also, 16nm nVidia uses is developed by Samsung(TSMC got sued successfully last time I checked), so no "innovation" there as well.

i don't understand why hairworks and purehair exist when you can just use verlet integration on hair mesh. wouldn't this be waaaay more efficient on the gpu?

>>55369932
This video has literally nothing to with driver overhead. Literally nothing whatsoever. The whole driver overhead "issue" is only remotely relevant when you're using a budget CPU. That video is using a 6700K, which isn't even remotely close to being maxed out during Just Cause 3 under any circumstances.

Whether driver overhead is a meme or not, posts like this prove that most people here are too tech-illiterate to understand it, or the circumstances where it could be a problem. Here's a fucking hint: If you have a 6700K, you don't need to worry about it.

>>55371481
>it's a joint venture.
So you admit it's a moot point because they did fuck all during r&d. Ok.

>Also, 16nm nVidia uses is developed by Samsung(TSMC got sued successfully last time I checked), so no "innovation" there as well.

The fuck does this have to do with either amd or nvidia? They hire out these foundries to produce their chips. Nobody has even claimed nvidia invented the 16nm tsmc process. Are you really trying to say amd invented samsungs 14nm process? You are one deluded little child.

>>55371416
Solid evidence of hairworks gimping your performance.

>>55371559
>5000 posts later and he's now trying to change the subject after I BTFO'd him

good to know you're running scared :^)

>>55371532
>So you admit it's a moot point because they did fuck all during r&d. Ok.
and now you're just spouting shit,

>that tone
wow, superiority complex much.

nVidia has more money to throw around, yet they iterate. If you're trying to say nVidia spent more money on optimizing Maxwell, sorry Pascal, then you're right. If you want to talk about progress and moving forward even nVidia uses technologies co-developed
developed by AMD, while making their own little extensions almost no one will use, maxing out revenue in the process. Which is great, but has nothing to do with progress.

>>55369826
> physics simulation
> driver overhead
That's two different things.
But OP is stupid. Driver overhead is a thing, and it exists on both cards. Under DX12, it won't be an issue. That's why DX12 was invented, and that's why consoles had some advantage - because they had no driver overhead.

File: 7b.jpg (58KB, 1198x654px)

58KB, 1198x654px

Confirmed $250

Amd are finished.

>>55369744
amd itself is a meme

>>55371777
Just like rx480 was confirmed $199 and gtx1070 was confirmed $379?
Yea...

>>55372114
If you couldn't get an RX 480 for $200 at release and actually wanted one, you were just slow. You only have yourself to blame.

1070, on the other hand, hasn't appeared for less than $400.

>>55371280
>DX12\Vulkan

glnext was planned before AMD had even started planning mantle

>HBM

nvidia was working on their own stacked, wide-bus memory pretty much identical to HBM. hynix beat them to market:

http://www.theregister.co.uk/2013/03/19/nvidia_gpu_roadmap_computing_update/

I fucking love my 480.
Stay mad guys

>>55372414
K

>>55372114
the rx 480 is 199

>>55372966
Lol

File: 2016-07-02_07h53_48.png (110KB, 699x607px) Image search: [Google]

110KB, 699x607px

>>55373746
fuck off shill

>>55374027
>out of stock
Lol

>>55374027
Don't bother, nvidiots are too retarded

File: 1381843555073.jpg (41KB, 251x242px) Image search: [Google]

41KB, 251x242px

>>55374042
And you think the price will go up when they get more stock?

Is that how business works? More stock = higher prices?

>>55370138
The numbers would be more useful if the GPU was paired with a CPU it's actually likely to be paired with.

>>55374080
>sells out
>drops price to msrp while sold out to piss off other sellers
Lol

>>55374144
>this much denial

Lol

>>55372131
>$399.99 on Newegg right now
CHECKMATE. Nvidia wins again. Please ignore that $3.99 shipping and handling. Please.