>beaten by Nvidia in the high end
>suffer from poor driver optimisations and overhead in the lower end
Is any AMD product worth buying?
No.
AMD belongs in the trash.
>>55532774
Kek
>>55532774
Pretty much this
>>55532368
>comparing 4 year old cards
For what purpose?
>>55532774
>driver overhead
AMD really is trash
>>55532368
>driver overhead
lel like someone is giving a shit for dx11 nowdays
>>55534182
>>55532368
>AMD
>Drivers
Quit making new shitposting threads Nvidiots.
>>55532368
>Is any AMD product worth buying?
No.
>>55534262
>bringing light to real issues a company is trying to hide
>hurr durr stop shitposting
>>55534262
Fuck off AMDrone.
>>55534262
they are getting paid by posts
>>55534393
Yeah it's all just some big conspiracy...
Or maybe people genuinely hate AMD for making shitty products.
>>55534428
if you dont know that both companies use inactive accounts on forums to spread shill and shitposting then you dont really have place to anywhere...
go to ocn on the review thread and see how many of those shitposted while was inactive for 3-4 years and literally having 3-4 posts anandtech banned almost 40 of those accounts ocn i dont know but more than 10 i guess
>>55534474
Or maybe a lot of people hate AMD for scamming and making low quality shit that kills your motherboard.
>>55534428
Bought my HD6870 in late 2013 on sale.
Meanwhile, my old GTX260 still gets driver updates...
>>55532368
Kill yourself
>>55532368
It's sad that this has been true since fermi and AMD just doesn't give a shit about it's customers.
In 2014 they only had one driver update...
Really a shame since they release some genuinely good hardware.
>>55534515
>amd scams
thats nice
what you say about rotr and total warhammer dropping support for maxwell 1.5 of their release?
or the fact that nvidia still cant do async on neither vulkan or dx12? despite their constant lying http://www.legitreviews.com/nvidia-highlights-directx-12-strengths-amd_138178
or the 3.5gb
of the literall housefires gtx 480/590
or that they put a 2 years old card into legacy mode
or or or or or or or or or or or or
literally even 3dfx would be ashamed of what nvidia has been which is ironic since its nvidia
>>55534560
its sad that people thinks its a driver problem
literally being this dumb on a technical board..
>>55534579
>async
Nvidia supports software async compute, that's not a lie or scam.
>3.5gb
Gtx970 has no problems accessing and using all 4GB of vram and performs exactly as it should at that price point between the 980 amd 960.
>2 year old cards into legacy mode
Except they didn't, ffs even cards older than AMDs legacy shit like the 2012 HD6000 series are still getting updates.
>>55534615
ofc it is and vulkan is the proof of it since it supports too async on software level BUT the problem is vulkan use the cpu for it not the OS and that became laughably clear with the latest doom update lol
actually it has thats its have been proven a lot of times on beyond3d..
they didnt?
hmmm having support and having an active development is two different things you know
amd has support for 9700 pro that doesnt mean they are actively giving it perf boosts lol
i wonder how a gtx titan/black user feels to have his card no longer on active support after paying irresponsible amounts of money
>>55534697
Funny because fermi is still getting drivers for new games while AMD simply gave up on anything not GCN, which includes a chunk of HD7000 series cards which got rebranded, and correct me if I'm wrong, but the r5 230 too.
>>55534579
>literall housefires gtx 480/590
someone say house fire 590s?
Bitch dont know about my space heaters
>>55534830
moving goalposts again
no card below maxwell 2.0 is getting any perf boost anymore having support and being developed to it its different things
>>55534887
My gtx680 says otherwise.
>>55534887
Not the guy you're replying to, but getting "support" as in the card still working is way more important than some minor/non-existent performance boosts.
>>55534903
oh you are the shill from yesterday that said it was playing everything on ultra?
lel
please give me an improvements chart on how your card gained perf over time
>>55534917
What?
Nigga the fuck you on about?
No fucking way my 680 can handle everything on ultra, not anymore.
Still runs everything great on high though and gets monthly driver updates.
Ultra is for plebs who don't do anything better with their lives, I'll upgrade when my 680 can't handle medium.
FYI, it's the 4gb model. And any improvements you think GCN is getting is from them testing on newer, stronger, CPUs like i7 6700k, which aren't affected nearly as much by driver overhead as older processors like 2500k or 3770k.
>>55534887
>no card below maxwell 2.0 is getting any perf boost anymore having support and being developed to it its different things
the 780ti and 970 actually gained quite a bit of perf in the last few months, in feburary they were both outperformed by the 390x in gta5 and now they're both considerably ahead of the rebranded 390x.
https://www.techpowerup.com/reviews/ASUS/GTX_980_Ti_Matrix/14.html
https://www.techpowerup.com/reviews/AMD/RX_480/15.html
>>55534949
http://www.babeltechreviews.com/hd-7970-vs-gtx-680-2013-revisited/3/
really hard to believe it
Open Source BSD Support.
>>55534972
cherry picking benches wont help really
http://www.babeltechreviews.com/nvidia-forgotten-kepler-gtx-780-ti-vs-290x-revisited/3/
>>55535039
>>55535072
Like I said earlier, driver overhead is less severe with newer stronger CPUs used in benchmarking newer games.
Are you retarded?
>>55535072
that's a no-name reviewer and they used setups completely different from the original results.
honestly wouldn't be surprised if they were being paid to peddle bullshit like this, amd is desperate to sell more of those 290x rebrands now that the 480 has flopped massively.
>>55535100
>driver overhead
literally being this stupid not realising that is a uarch native problem since gcn 1.0....
being this into damage control
>>55535072
>>55535039
>reposting the same shitty irrelevant link twice
>>55535125
used a better setup since 2013 to bench on 2016
>no name reviewer
its literally the only site that actually does this since 2007.....
but you probably didnt even knew that lol
>>55535142
>used a better setup since 2013 to bench on 2016
which pretty much invalidates the results. almost every game they listed was released in 2014 or 2015 and the CPU they're using was released in 2014. my guess is that the original benchmarks were done on a really old CPU like nehalem era or sandy bridge era.
>its literally the only site that actually does this since 2007.....
you mean it's the only reviewer being paid to do this. it's pretty much rule 1 of reviewing to do like for like comparisons (not changing the test system, comparing reference gpu vs reference gpu, etc).
>>55535142
>i7 6700k suffers less from driver overhead than an i5 2500k
No shit.
>>55535180
using a better system for both cards invalidates results
that has to be the most stupid comment i saw on 2016 so far
>>55535186
being this stupid that doesnt realise that a massive paraller gpu will always have a "driver" overhead simply because its not desing to run on a serial api..
6 weeks of lurking /g/ and i'm still no closer to deciding on what card i want to put in my new build that i've been running integrated graphics on for the last 3 months
i just want, for the first time in my life, to own a card that can play any game i throw at it nicely and won't set me back an extra $800
>>55535227
are you retarded? they're comparing old results to new results, using an identical test setup for stuff like this is basic and is the 'standard' for reviewing.
this reviewer is also the definition of shady. they have a very clear agenda, just taking a look at some of the newer reviews they've published shows a clear attempt at selling specific versions of nvidia and amd gpus:
http://www.babeltechreviews.com/devils-canyon-i7-479k-vs-skylake-i7-6770k-using-rx-480/3/
depicts the 970 exoc destroying amd across the board
yet one review below that one:
http://www.babeltechreviews.com/rx-480-reviewed-26-games/5/
completely different results, 480 and 390 actually beat the 970 exoc in some of the results instead of straight up losing across the board
>>55535227
The point of a gpu benchmark is to test gpu performance.
Swapping out other components makes the comparison invalid since of course a fucking i7 5960x will destroy an i7 920 in every single benchmark.
It's not a gpu benchmark if the gpu isn't the only thing that changes.
>>55535254
Yeah? How about pre GCN cards that suffer from the same issue
>>55535274
If you have an overclocked i5 6600k or better, then get AMD, otherwise always get Nvidia.
>>55535279
holy shit a better setup actually helps the cards potential?
who would have knew !
being this into damage control
>>55535312
>holy shit a better setup actually helps the cards potential?
explain how the 'better setup' helps one brand's performance in one review yet harms it in another one by this shady as fuck reviewer then
>>55535286
thanks m80
will a 390 handle something like GTAV? is a 390x a better option?
>>55535336
390 is the better value option, but the rx480 overclocks better and beats it for the most part while usually costing less.
Also has half the power consumption and heat output, so aftermarket cards should run cool
>>55535361
thanks brah
>>55535361
>>55535336
Here's what you can accomplish if you use anything but the reference cooler, and cards from saphire or gigabyte will be even better because they'll have an 8 pin instead of a 6 pin.
http://www.gamersnexus.net/guides/2499-diy-amd-rx-480-hyrbid-results-with-bigger-overclocking
>>55535333
you moron this shows that nvidia DOESNT actively boost their older cards like amd does
this is called planned obsolescence
>>55535642
except AMD performance didn't improve, they're comparing AMD performance on newer systems to NVIDIA's performance from their 2013 system.
the only thing it shows is that the reviewer isn't impartial and can be bought, he has plenty of other obvious nvidia and amd shill articles that have strange performance discrepancies depending on what's being shilled.
>>55535642
HD6870 sure has gotten much faster in the past 2 years compared to gtx570
Fresh OC.
>>55532368
here's a (You).
>>55535397
Hold on.
That 1070 list is weird. I have a FE card and it clocks to 1898. What's the dealio?
>>55535766
The Founders Edition will clock up to nearly 1900mhz on it's own but only if the card is below 65c.
If you want to game on that, you need a custom fan curve.
>>55535694
you must be some dense mofo lol
are you that blind that you cant see they benched on 2013 drivers and 2016 drivers?
are you THAT DENSE?
this is some next level of shitposting lol
>>55535274
Get a 1070. AMD cards are all too weak to handle anything like that.
>>55535795
you are illiterate or shilling, maybe both.
they compared a stock 780ti review from 2013 to overclocked 290x (390x) benchmarks from this year.
>>55534903
780ti user here
The gimping meme is real.
>>55534615
>defending and denying 3.5gb
Kek.
My boyfriend's daughter enjoys the latest Nvidia(tm) card every year, it puts a smile on her face and she thanks me for being her most fun father yet.
>>55535875
Holy shit you're autistic
>>55535821
You don't update drivers, do you?
>>55535820
where the hell did you saw an overclocking 290x?
>>55532368
And yet somehow AMD is all the way at the very top in the very chart OP posted....
>480 now rivalling 1070 performance in doom
>>55535817
meanwhile 1070
https://www.youtube.com/watch?v=uxnFv8cnjZA
It's always the same. Benchmarks show 480RX being ahead of 970 gtx by 5-10 frames until fanboys realize those benches were performed on a $400 CPU and their 5 years old i3/i5/8350 suddenly are 15 fps behind 970 cuz of overhead. AMD drivers are so bad that this issue shows even in dx12 where GPU/CPU utilization is better.
>>55536020
>dx12/vulkan
>overhead
>amd
kek
maximum damage control
thats why nvidia cards even without async are behind
guess having a paraller gpu that works perfectly with a paraller api suddenly has overhead
>>55535989
>AMD shills actually believe this
>>55536044
Hey fellow AMD shill. I'm sorry to inform you that DX12 won't fix terrible overhead issues AMD still has. Here is a benchmark showing Hitman in DX12 in a GPU intense scene - As you can see 480RX shows its claws and fully supports async computing in Hitman blowing 970 gtx away.
>>55536097
AMD finally released something that can compete with the gtx 970?
It's about time
>ze germans showing an 8350 and nano going from 36fps to just over 50fps at 4k
While the gains from DX12 have been impressive in doom vulkan is absolutely unleashing GCN. On anything other than the mightiest i7 the performance increase seems to be at least 40% for a lot of systems (rough guess based on user results posted on the likes of OCN and anandtech).
AMD's own published results for the 480 (showing a 27% increase at best) were because the 480 was paired with a 5960x.
>>55536081
The screenshot shows the 480 getting 153fps correct? What is the 1070 getting in this chart?
>>55536097
However the overhead STILL shows up the moment you have a weaker i5 or an i3 CPU which suddenly makes 480RX barely better than 2 years old card in "Best Implementation Of DX12 Async Compute Yet" game.
Glad I could help you out, hopefully you feel bad for selling this garbage to poorfags with $500 budget rigs.
>http://wccftech.com/hitman-feature-implementation-dx12-async-compute-amd/
>>55536136
That bench btw was made on Sapienza map which is very CPU intense. So yes - Your cpu will bottleneck an AMD card way sooner than a Nvidia card unless you have a $200+ CPU - Somehow I don't think that's the case for most AMD owners.
>>55536136
wait wait this must be some kind of a joke right?
you are this depserate that you brought a dual core cpu into discussion for an api that benefits greatly from quad core and above?
is that some kind of next level nvidia stupidity?
whats next? pentium 3? amd k7?
>>55536179
I guess 4460 is dual core too right shill? 4460 is above 2500/3570 which still are most popular CPUs out there. Your GPU overheads mid range CPUs, deal with it shill.
>>55536172
>mfw r7 270x + i3 4160
dx12 will be better, right?
>>55536207
you cant really be this stupid now can you
>>55536315
kys. Stop shilling and face facts - Your AMD card is a meme card and every poorfag who buys it expecting excellent performance in dx12 will be doped unless he has a $250 cpu.
>>55536327
kek trying to justify nvidia poor perf
>>55536327
Wut? The joy of DX12 is that it reduces CPU overhead.
You're confusing your dumb memes
>>55536509
That's why AMD is forcing devs to rush dx12. They're going to refuse to sponsor any game that's still dx11
>>55536744
its not only amd thankfully..
its ms
its sony
its ibm too
its quallcomm
its nintendo
its samsung
and most of the devs already
literally nvidia will just stick with indie dx11 devs
>>55536744
There are plenty of DX12 games out and coming that aren't backed by AMD.
And how are they "forcing to rush"? Like they'd have even close to the leverage Nvidia has over devs.
>>55536803
i dont see any dx12 game that isnt backed by amd at this moment..
literally they all are
amd even snatched evolved from novidia
>>55536908
>tomb raider
>>55536916
well it took 7 patches to become a dx12 title pretty sure no one really counts it
unless 15 games are benchmarks and 1 is not
kek
>>55536935
Haven't you heard? If Nvidia doesn't win in a game its because either 1) its just a benchmark or 2) nobody plays the game.
As of right 10th july nobody plays tomb raider and as of the 11th july nobody polays doom.
>>55536974
this is getting so hilarious by the day...
>>55536984
Lets be fair here, Nvidia is winning in 50% of all vulkan games.
Battlefield hardline is dx11 or 12?
>>55536999
dx11/mantle
what is this new overhead meme
>>55537004
Fuck
I have an r9 390. It runs pretty much anything I want it to maxed at 1080p with close to 60fps. The exception is the new Hitman, which I suspect may just be poorly optimised. It also has started to crash like an absolute motherfucker after 3 mins of play, consistently.
That said, the software for installing drivers is a massive pain in the ass. I just want to download an .exe. I don't need some intermediary software to hold my hand and try to trick me into installing some bullshit amd-branded x-fire/social network/skype alongside the latest drivers.
As someone who used to own nvidia shit, that's the biggest advantage I've come across so far from a purely UX point of view
>>55536803
Heroes & Generals - The Way It's Meant To Be Played
Rust - No affiliation
Ashes of The Singularity - Gaming Evolved
Descent: Underground - No affiliation
Squad - No affiliation
Survarium - No affiliation
Caffeine - No affiliation
TESO - The Way It's Meant To Be Played
Rise of the Tomb Raider - The Way It's Meant To Be Played
GOW Ultimate Edition - Independent
Hitman - Gaming Evolved
Quantum Break - No affiliation
Forza Motorsport 6: Apex - No affiliation
Total War: Warhammer - Gaming Evolved
Let's see, AMD has 3 titles and Nvidia has 3 titles. Feel free to off yourself.
>>55537736
Well fuck me, meant to quote>>55536908
and not myself
>>55537736
lets see
5 irrelevant games
2 already on dx11 since forever
no dx12 for nvidia
meanwhile
aots is amd
rotr is a hybrid really
hitman is amd
qb is amd
total warhammer is amd
doom is amd
deus ex is amd
star citizen is amd..
you keep forgetting one single thing..
whoever uses async/casting/shader intristics or porting in general is using amd tech by definition there isnt a single thing on dx12 that nvidia has added or can add nowdays..
>>55537968
>>55538187
AMD should just make better parts
>>55537968
You can blame only blame Nvidia for cutting out all things relating to compute so that they can increase their power efficiency.
>>55538233
They did - 4 years ago.
>>55537968
>doom is amd
You what now?
WHY ARE THERE NO BENCHMARKS FOR APUS IN DX12 TITLES FOR FUCK SAKE
I have the spare cash required to buy a GTX 1070 (MSI Gaming X, because of the quiet cooler).
Is there any reason I shouldn't buy one? The Price/Performance ratio seems pretty good.
>>55538267
sorry forgot to fill in about it
https://community.amd.com/community/gaming/blog/2016/07/11/radeon-graphics-takes-doom-to-the-next-level-with-vulkan-implementation
dont mind the shitty nvidia like graphs lol
>>55538313
Shiggy diggy doo. Nvidia also marketed with Vulkan and D44M a long time ago.
>>55538187
Microsoft is the Cancer of Gaymen
>>>/v/ you fucking child
>>55538384
no shit
problem is as tiago said
they CANT support it when async is on
nor do they have any flip queue capabilities lel
>>55538423
I'll point you to >>55538252
>>55538478
Let me write a summary.
>its AMD's fault
>its MS's fault
>its Khronos' fault
>its [developer]'s fault
>its not Nvidia's fault
>>55538546
nvidia never li.....oh wait no sorry
http://www.legitreviews.com/nvidia-highlights-directx-12-strengths-amd_138178
>>55535274
The amazing amount of nvidia asshurt and shilling, on top of the blatantly anti-consumer practices of the company, plus their scamming with 3.5gb was altogether more than enough to convince me not to touch their shit ever again
nvidia can go fuck themselves - their paid shills aren't convincing me to buy nvidia, they're convincing me to give the company the middle finger and go AMD or fucking intel integrated permanently
>>55535742
nvidia shits the bed on actually developing shit, instead opting to cripple their competitor with jewworks while running extensive shilling campaigns online
meanwhile AMD actually works on and develops an open source API that utilizes their architecture well, and then releases that for anyone to use, coincidentally ensuring that amd would lock in the console market as well
nvidia had their chance to get ahead and they squandered it on petty bullshit while amd pulled a 4d chess move on them
>>55538384
>gtx980 performance at 2x the power consumption
why would they brag about that???
>>55535274
Why the fuck would you come here for information?
>>55538659
This. AMD as an entire company may be tanking but it certainly isn't the fault of the graphics division.
>>55538719
>2x the Energy Efficiency
Did you read the entire line?
>>55538781
>This. AMD as an entire company may be tanking but it certainly isn't the fault of the graphics division.
It actually is, look at their balance sheet.
They lose money every quarter in desktop graphics (and desktop CPUs).
There is a rumor they are going to sell their desktop GPU business to Intel.
>>55538781
reading is for faggots
>>55538546
>>55538561
What's so ridiculous about the whole thing is that Nvidia had so much time to prepare for DX12 and async. The first AMD desktop GPUs to support it released how many years ago, 4 or 5? If that didn't put Nvidia on notice, the Xbox One and PS4 should have because both Microsoft and Sony adopted APIs that utilize asynchronous shading and compute.
>>55538851
they couldnt because they lost the consoles that they "didnt care" as they said
>>55538851
Nvidia thinks itself invulnerable.
Word is (from Ed from sapphire) is that the 480 has Nvidia scared as they expected it to be slower at that price point, so they are rushing the 1060 to market to have a competing product.
I also personally suspect Nvidia's shills (we know they emply some) are running damage control across the web to buy time for Nvidia to have the 1060 ready.
>>55538837
I should have elaborated. They've gotten murdered in discrete GPUs, yes, but what has kept their entire company afloat is the APU driven IP that they inherited and finished developing after the ATi purchase. This has given them custom design wins and stuff like the past two console generations as well as future consoles. Not high margin stuff but it's at least brought in some cash while the CPU division has just floundered for a decade.
Also, if their dGPU business is so bad then nobody, not even intel who just laid off something like 12,000 employees worldwide, is going to buy them out before they go bankrupt on their own and are forced to liquidate assets and IPs.
>>55538883
>they couldnt because they lost the consoles that they "didnt care" as they said
It was obvious that they were sore as fuck, being completely out of the current-gen consoles was a huge loss for them, especially as the hardware features on AMD cards were rolled into DirectX 12 as a result. Still though, they should have sucked it up in 2013 and adopted the new hardware features the new DirectX implemented.
>>55538909
>I also personally suspect Nvidia's shills are running damage control across the web
Probably true, as soon as the RX 480 released and benchmarks came out "leaks" that the 1060 was coming soon started to release claiming it would be out soon and faster than the RX 480. These "leaks" seemed coordinated to harm RX 480 sales.
>>55538909
I don't understand why they're fucking about and ignoring dx12/vulkan still
At this point, even when they DO put together a card that's designed for the API, it's going to be full of bugs/issues that AMD's managed to work out and fix over the past 5 years.
>>55539106
because they wanted to milk the dx11 cow
if it was on nvidia hands they would have 300000mm2 gpus accommodated for gameworks
>>55539020
>>55538909
>>55538851
I also should have mentioned that one of the reasons they haven't done well in dGPUs is due to GCN's inherently shitty performance in DX11 games. GCN is very much an architecture that was introduced ahead of its time and is only now going to start to reap the benefits of its design. Sound familiar? AMD has a history of doing this to themselves but if they can hang in their for a little bit longer then they'll be fine.
Nvidia is starting to look like the company that is behind the curve at this point now that the DX11 honeymoon is overwith.
>>55539106
Historically AMD's engineering talent has been superior to Nvidia's - making gpus that compete despite having a muc, much smaller die. Its only been the fury series where AMD has tried ot match Nvidia with a monster die and it really didn't work that well.
Even going back a generation - hawaii is considerably smaller than full fat kepler and provided competitive performance. This is why there where rumours of some vega cards being dual die - if you can make it work like the theory goes its an enormous advantage for AMD as they typically can't meet Nvidia head on at huge die sizes.
>>55539106
Nvidia is way behind AMD in research related to scalable, multi-GPU architectures that are eventually going to dominate the future so they have naturally been opposed to anything that will move graphics technology away from discrete single GPU solutions. Read: DX12, Vulkan.
The writing is on the wall. Silicon shrinks can only go on for so much longer before they aren't economically feasible and Nvidia has nothing but dGPU tech at this point. They will fight anything multiGPU as much as possible. Note how they are discouraging SLI at this point for proof of why they're getting more defensive about supporting multiGPU solutions.
AMD is trash, we need a different company to compete with Nvidia
>>55538909
>Nvidia
>Scared of the 480
One day before the 480 was released.
Do I have to remind you that Shappire depends almost entirely on AMD success? if the 1060 is better than the 480 their profits will be marginal.
You're a fool if you think Nvidia couldn't predict AMD's 480 performance, pro tip; companies have entire teams to analyze their competence products and expected release products, not to mention that Nvidia hired lots of ATI engineers AMD fired when they bought ATI.
>>55539477
Sapphire actually has a sister company, Zotac.
>>55539106
Because low level API are still immature, not to mention developers are seriously lacking.
The performance you get in DX11/OpenGL should be yo baseline performance, if you get lower performance with a lower api even without async you can immediately tell someone fucked up big time that the Nvidia driver can do a better job than the human who wrote the code for the DX12 path.
>>55539477
Seriously Nvidia is just laughing at AMD at this point
>>55539494
And?
>Do I have to remind you that Shappire depends ALMOST entirely on AMD success?
>>55539494
Aren't both owned by PC partner?
>>55539543
Yeah.
>>55539536
That specific branch does, but the main company does not.
Sapphire is basically THE AMD brand, though, that's true.
>>55539571
If the Shappire branch doesn't delivers I don't have to tell you what will happen to it next do I?
>>55539313
I agree. The irony of AMD struggling over the past 5 or so years is that they've actually had a more forward thinking philosophy than Nvidia but GCN was too far ahead of its time. If RTG survives for another 5 years or so it'll be interesting to see how GCN adapts to the DX12 landscape.
>>55539417
Get out of the thread Jensen.
>>55539477
You act like it doesn't go both ways. Both companies share and have shared tons of employees.
>>55539510
This. DX12 has the potential to expose a lot of "developers" as shitty coders.
>>55539587
What the fuck are you on about? Who cares about Sapphire? They've had a special relationship with ATi and AMD as an AIB but literally who the fuck cares about that right now?
>>55539604
>What the fuck are you on about? Who cares about Sapphire? They've had a special relationship with ATi and AMD as an AIB but literally who the fuck cares about that right now?
Do you have autism and can't follow a conversation or what.
Shappire says Nvidia was scared of the 480
Twitter from high rank Nvidia's employer making fun of the 480 the day before it went on sale or any review was out.
It's pointed out how Shappire depends entirely on AMD and they are know as "THE AMD" Branch.
What is this driver overhead meme? I just tested in Firestrike (4690k 4,5ghz vs stock). And got roughly the same score.
>>55539726
Try it again at 2.2GHz (effectively about as fast as an FX chip), then you'll see what the "driver overhead" meme is all about.
>>55539762
But FX chips were always horrible for gaming, the last decent one was the phenom II. Ever since the Q6600/E8400 came out AMD processors were shit for gaming.
>>55539783
NVidia cards still perform fine on an FX chip though.
>>55539604
>This. DX12 has the potential to expose a lot of "developers" as shitty coders.
Even though I dislike a lot of stuff about Windows 10, DX12 is a step in the right direction and its adoption should be fast given most games are developed for Xbox One, PS4, and Windows.
>>55539726
tl;dr
AMD drivers use more CPU compared to Nvidia's so since the CPU is used by the driver the games have less CPU to use since it is being used by the driver performance hit is more noticeably specially in games that are CPU heavy.
Pairing a AMD GPU with a weak CPU will give you lower framerates than pairing a weaker Nvidia GPU with the same weak CPU.
You need a strong CPU to get the most of your AMD GPU.
http://forums.guru3d.com/showthread.php?t=398858
Weaker Nvidia card pushing more FPS than the stronger GPU on the same CPU.
https://www.youtube.com/watch?v=S546TL2LWNY&feature=youtu.be
https://www.youtube.com/watch?v=XoKu0_2ozAc&feature=youtu.be
>>55539853
This is the reason why the GTX960 sold more than the 380 too, not everyone who is buying a 960 or 380 are running last gen CPUs. Unlike reviewers that benchmark the 960/380 with 4790K OC(where the 380 wins no CPU bottlenecks) people who buy budget GPUs usually run slow CPUs worse than a 2500k so they instead use Youtube to search for someone with their same build so they compare gameplay.
>>55539897
You're giving customers too much credit, most people are buying based on price and brand affinity.
>>55539762
Depending on the clockspeed of the fx chip it will be considerably faster than a 4690k at 2.2ghz.
>>55539853
fuck i wish i knew this before going for my poorfag build a few years back with fx 6350 and 7850, the 7850 died pretty fast for some reason and now i am with 290 so i hope it kinda makes up for the cpu overhead
>>55540016
If anything that makes it worse. You need to either overclock the shit out of that CPU, switch to Intel or switch to NVidia for the best performance.
>>55540034
meh, i can play pretty much anything on 60 fps and high-ish settings so i am okay for now, i slapped an evo 212 on the cpu because the stock one was crazy loud so i doubt i can get anything with OC that's not gonna be like 2-3 fps in exchange for loads of heat
>>55539945
You're underestimating people, people who buy by individual parts actually put some research on it, it has been like this since the 8800GTX days and it hasn't changed to this day and the only people who just buy for brand are usually people who buy prebuilts.
Not to mention than in plenty of 3rd world countries Nvidia is cheaper than AMD.
