>r9 280x faster than 380x
I think AMD gimped the wrong card, but don't worry, they'll gimp the 280x soon enough to give the 380x and 480 an even bigger edge :')
>>55580494
>rebranded 7850 faster than 780 Ti
Nvidiots will defend this.
>>55580494
>HD7950 just as fast as a 960
My OCed 7950 is equal to my friend's 960 in DX11. How much faster will it be in Vulkan?
The 280x is based off of the GOAT 7970, the 380x is based on the 285 with a lower memory bus.
>>55581330
the 285 also has significantly better tessellation performance.
weird shit mang
>>55581330
Lower memory bus, higher frequency and delta compression.
Shouldn't make a difference at 1080p.
>>55580494
>GTX TITAN
Nvidiot will defend this jewfuckery.
>>55581527
>an old, outdated, and no longer sold product struggles to run the latest graphics API and video games
WOW LOOK AT ALL TEH GIMPS YO
>>55581622
the titan X is, was, and will forever be a joke.
>>55581645
There is no good reason to buy AMD.
>>55581622
How come the HD79xx series is on the chart, even higher up then titan?
>>55580660
>Will defend this
Can we end this expected troll meme shit yet?
![red-herring[1].jpg](https://i.imgur.com/7pQq6Eh.jpg "red-herring[1].jpg")
>>55581660
m8
>>55581681
>7850 faster than 780 Ti
No seriously...why is the 7850 faster than 780 Ti
>>55581669
Because AMD built mangle around the GCN architecture that they're still rebranding after 4 years.
Vulkan is just a glorified mantle/opengl.
>>55580494
>TFW 780ti
YOU HAVE TRICKED ME FOR THE LAST TIME NVIDIA
>>55581669
because Nvidia cant into low level APIs
>>55581645
>>55581645
I remember when people would come to /g/ and brag about their Titan X...was not long ago....I think it was just 2 years ago. I bet they going to buy the 1080 next only to use it for another 2 years.
>>55581792
>tfw 770
>tfw getting outperformed by GPU that have 128bit bus memory
the gimpworks is real I'm getting a PS4 Neo this fall, fuck PC gaming the Steam store is full of indies and autism simulators like garry's mod amyway
>>55581823
What's wrong with that? They've got the money to spend on the best.
>>55581885
>>55581792
I don't get it, how can anyone be stupid enough to have such high expectations of an outdated product not even being sold anymore?
Plus funnily enough, my gtx 680 always manages to outperform the benchmarks available online even though my cpu is only an i5 3470. You need to wait for drivers before getting so butthurt. Unless you're really an AMD shill, of course.
>>55581779
>this level of damage control
Why are you shilling so hard? Buyer remorse?
>>55582022
Yes. I regret my HD6870 so much.
>>55581982
No OP but.
>770.
>AMD Shill.
Are you serious or you don't know who to blame for what pajeet?
who is pajeeting who ?
>tfw you bought a new 290X for less than the cheapest 970
It just keeps getting better.
>>55580494
if you sent this benchmark to nvidia and asked for an explanation , what would they say ?
i'm very curious
>>55582246
My GTX680 still plays most Modern games at high or ultra settings.
Glad I got the 4GB model 2bh.
Trick is to not play the game at launch, but to wait for prices to drop and drivers to mature. The game itself will usually get some performance and bug fix updates by then too.
>>55581982
>>55582398
Funny that the older nVidia owners have more sense than the AMD owners.
380X owner here. Even I know that it's a driver issue.
>>55582390
They wouldn't say anything, or they would say it's just a quirk of the architectures that is exaggerated in this particular version of Vulkan.
>>55582508
god damn semantics .
damn them to hell
>>55581982
>I don't get it, how can anyone be stupid enough to have such high expectations of an outdated product not even being sold anymore?
The 290 I bought was in direct competition with the 770/780. At the time of purchase I paid $275. When I see graphs like the one in this thread it tickles me pink. AMD only gets better with time. While Nvidia loses performance with time. So I would hazard a guess that many AMD users appreciate the extended lifespan of their purchases unlike the other side of the fence.
>>55582595
Behind close doors, they'd tell you that the GCN architecture is just better suited to Vulkan than nVidia's. Publically, they just can't say that because they'd get fired.
>>55582490
I'm glad I'm not the only one here with some common sense.
Also
>inb4 legacy
Keppler and fermi are still getting game ready drivers, which is more than can be said for HD6000 series cards. My friend went from a HD6870 to a gtx 970 after AMD dropped support for his card so quickly.
GCN seems to be aging pretty well though and grabbing some customers back from nvidia.
>>55582508
Nvidia usually tries to stay quiet about this shit, they didn't say anything about the 970 until it was getting out of hand iirc.
>>55582645
The only reason that AMD dropped support for the 6000 series so quickly is because they switched to GCN, and didn't have the resources to keep the older architecture up to date.
The reason that nVidia can keep their legacy shit up to date is because their architecture hasn't fundamentally changed since Tesla (which is the reason that it's not really prepared for Vulkan).
AMD's GCN was in anticipation of architectures like Vulkan and DX12, while nVidia is trying to adapt CUDA to the model and hasn't got it right yet.
>>55582690
Oh I know why they dropped it, but I also understand why it cost them some customers.
One thing I don't understand though, why can't Nvidia use their Cuda cores for async?
A gpu core is so a gpu core, regardless who makes it. Or am I oversimplifying it?
>>55582726
>why can't Nvidia use their Cuda cores for async
Technically they can, but from what I understand, nVidia's scheduler is very limited in the number of requests it can handle, so is limited by the number of different tasks that it can perform at any one time, and as Async disabled or done in software with huge overheads.
AMD's scheduler can process any number of requests with ease.
I suggest you take the time to read through this.
https://www.reddit.com/r/nvidia/comments/3j5e9b/analysis_async_compute_is_it_true_nvidia_cant_do/
>>55582390
>>55582595
It's just an error on the game developer side god damn it you guys use you fucking brain for once, some Nvidia cars run worse on Vulkan than OpenGL, there is still some work that needs to be done by the game developer and not Nvidia.
>>55582850
Thanks (:
My R9 270 2GB at 1050 core 1500 mem never drops below 30 and usually keeps me around 40 but I've only got everything set to high, not ultra. Also this is on my old X4 955 at 3.7GHz since my 8320e is being RMA'd.
>>55582863
Tell that to a fundamentally 10 year old architecture.
>>55580949
Yea the 7950 OC's like a fucking beast. Card sometimes OC's better than a 7970 making it faster and cheaper. Insane.
>>55580494
All I see are GTX's at the bottom. Poor fuckers that bought that Nvidia shit.
>>55583038
How about using fucking logic you dumbass, if the high level api performance which is your baseline performance is higher than the low level api then something went incredibly wrong.
>>55583057
Well yeah, same reason the 980ti beats the Titan.
Less shaders, same ROP's == less heat, better OC.
For games that depend more on ROP's than shaders, it makes complete sense that a higher clock would benefit more than more shaders cores.
I would still expect the 7970
GHz Edition to outperform almost any 7950 though, as it was on a more mature manufacturing process, and most of those cards have better coolers.
>>55583071
Are your eyes okay anon?
Maybe you need glasses...
>>55583077
read this article: https://www.reddit.com/r/nvidia/comments/3j5e9b/analysis_async_compute_is_it_true_nvidia_cant_do/
nVidia's scheduler halves in performance when using Asynchronous compute with more than 31 queues operational. That's the reason that they're trying to offload the queue to the software driver, which is comping as a massive performance hit.
>>55583119
>reddit link
DROPPED
>>55583126
Ignorance is bliss for the idiot.
>>55583126
>i-i cannot comprehend to counter. so i toss meme response
>>55583126
Reddit is great as a reference archive. I don't like as a discussion medium because of all the politics (hence why I'm here), but there's some good info in there (amongst all the shit).
>>55583107
Maybe you do... The picture is all green at the bottom.
>>55580494
280X has more CU's and bigger memory bus than the 380X/285X, it has always been a bit faster though it has quite higher power consumption and lacks a few features
>>55583154
It would be better if reddit links were through archive instead so they get no traffic.
>>55583158
And the top.
What's your point?
God Dayum that 1080 score. The 1080Ti is going to be Titan priced at this rate.
>>55581389
The 380X/285X has more tessellation units than the 280X, Tonga was a rehearsal for Fiji and came with improvements to cope with The Witcher 3 retarded tessellation
>>55583166
I agree, that would be nice.
>>55583160
280X has exactly the same core config as the 380X (2048:128:32) and only marginally higher memory bandwidth (288 vs 256 GB/s)
the 380/285 is small Tonga, and 380X is big tonga.
>>55583172
Why does the top matter when the post you were replying to was talking about the bottom? You need reading comprehension bro.
>>55583172
I think the point was that all of nVidia's mid-range cards perform like shit compared to AMD's mid-range (in Vulkan).
That's a big deal to people that don't spend all of their money on GPU's
the bottom are all kepler. except 750 ti.
>>55583261
Yeah, poor 760. That is disgusting.
>>55583261
>>55583246
>>55583071
>>55583158
See
>>55582398
The game will receive performance updates for kepper, and nvidia will release driver updates.
No need to get your panties in a twist.
>>55583261
Not relevant considering the HD7k series is the same age. 4D chess pajeets.
>>55580494
They should've tested the R7 360 too. For Science.
>>55583287
Like Project Cars right?
>>55583366
SCIENCE
>>55583384
Yep.
https://youtu.be/n3zUMhxz8V8
>>55583409
>gtx 670
Because it performs really well on AMD gpu's!
>>55583415
That's AMDs problem.
Nvidia is still optimising for Kepler, and even Fermi.
>>55583261
>muh gimping
Wait for drivers you idiot.
>>55583291
The AMD HD 7XXX series is OLDER than the Nvidia 7XX series and was built to go up against the Nvidia 6XX series.
>>55583445
That's a good point.
>>55583432
>that's AMD's problem
because Vulkan performance is nVidia's problem right?
>>55583462
Technically is up to the developers to optimise their low level API for all architectures, but nvidia is guaranteed to release drivers to help improve performance too.
Unlike AMD, they actually care about their customers.
>mfw dx12
>>55583485
>nVidia
>caring
whatever lets you sleep at night.
>>55583604
I wish fanboyism would end, AMD has fixed most of that with drivers already.
>>55583636
Thanks (:
>>55580494
>>55583670
Got a link senpai?
Were they using the same cpu to compare older benches to newer ones?
i love my rx480
>>55583697
>>55583773
What? I'm confused.
>>55583721
Neat
>>55583721
Good for you mate
>>55581779
I dont even like both sides, but anon you make nvidia fanboys look dumb and ugly.
>>55581885
>fuck PC gaming the Steam store is full of indies and autism simulators like garry's mod amyway
I cant agree more, the only reason i still game on pc is because of the mods and editors. All the indie shit pulled the whole pc gaming down into cuck zone which made gaben their bitch and sex slave.
God damn i can feel gaben's breath from my screen.
Talking about editors, which is ironic. My favourite UE3 turned UDK attracted shit tons of indie fucks, mark rein was right.
It all started with "now everyone can make games!" Idea.
>>55583445
Im still using a 7790 2gb and it performed well with current games with some settings low, some mid and some high. It was a small jump from my old gtx460 but this is good.
My friend's gtx650, 750 barely holds the cadle for some reason.
Might get amd 480 since i frequently game on my pc now.
>>55582645
>so quickly
>5 years
>completely different architecture
Nvidia hasnt changed their tech in any fundamental way since the 8000 series. AMD went through two different types of VLIW before GCN.
He should he happy a shrinking company with little money and manpower kept support for five years.
>>55580494
>960 beating a titan :^)
>>55584015
4 years actually.
>>55584026
Wait for drivers (:
When Vega Comes It Will Beat The Fuck On Off All The Cards Too Holy Shit
>>55584015
>>55584031
http://support.amd.com/en-us/download/desktop/legacy?product=legacy3&os=Windows+7+-+64
5000 - 7600 series just got an update to Radeon Crimson..
>gamegpu
this is how you know the results are not credible. either that or op is the same retard with broken english who has been posting photoshopped benchmarks trying to shill amd rebrands for the last few months
>>55583208
The 280x/7970 might have negligible bandwidth increase but the difference is in the higher mem controller amount, a 280x can handle more simultaneous data accesses than a smaller 256 bit bus, depending on frequency. For instance, GDDR5 at 2000 on a 256 bit bus might be able to match accesses while providing less bandwidth. Tonga by the way doesnt have 256GB/s bandwidth. You're quite lucky to break 1600mhz/205GB
>>55584015
>AMD went through two different types of VLIW before GCN.
AMD's VLIW architectures are much better than their SIMT-knockoff architecture. AMD is struggling so much with GCN's terrible perf/watt and utilization that it might actually be better for them to make an entirely new architecture at this point.
>>55584049
>1st March
>just got an update
Okay.
>>55584161
Hey man, on the timescale of 5 years, a few months is a small length of time.
wonder whether AMD still loss this quarter.
>>55581982
It might have something to do with the fact that the 780ti was a top of the line card just 2 years ago. Anyone who bought it has to be pissed as fuck about how nvidia has completely left them out to dry.
>>55582364
Hawaii is very, very good. Tahiti is GOAT however.
>>55584182
>old hardware is never superceded nor does it depreciate
the 780ti is still between 970 and 980 performance even at the shitty reference clocks, they're still very capable even if they aren't the best value.
>>55584173
A driver every few months wouldn't be bad for legacy cards, but this is unfortunately their final driver update.
>>55584182
Wait for drivers, this has happened before and nvidia has fixed with drivers then, as it will now.
Or did they never provide an update for the witcher 3 and project cars?
>>55580494
>gtx 1070 doesn't even break the top 28 cards
Nvidiots on suicide watch
>>55584182
they still doing great in opengl and dx11.
>>55584207
>Hawaii is very, very good.
decent performance but terrible if you don't want your gaymen room to turn into a sauna.
>>55584209
Stop sucking on that chink dick for a second and look at the chart. The 780ti is down near the bottom.
>>55584242
>he thinks gamegpu is a credible source
they are russian (rogue, nuclear armed terrorist state) and their numbers don't agree with any other reviewers. it's pretty much entirely spammed here to shill AMD products.
>>55584213
Expecting your highest paying customers to wait for drivers when you've clearly already spent time optimizing for everyone else is a really shitty thing to do.
>>55584273
Why do you hate the Russians, Chang?
>>55580494
>using the smiley with a apostrophe nose
>>55584026
>using the smiley with a carat nose
>>55584279
Highest paying customers wouldn't be using a 4 year old gpu.
That being said, are you disappointed that they're providing driver updates and fixes for legacy cards when most gamers just upgrade every 2-3 years?
>>55584341
>saying using the smiley face with the carat nose :^)
>>55584355
Planned obsolescence is such a great thing.
>>55584383
Yep. That's why the performance gap between the 560 ti and HD6870 has increased
>>55583697
>>55584436
That's not a link...
>>55584086
kek terrible utilization...a card from 2011 is still relevant at 2016
and a card from 2015 cant even beat a card from 2013...
epic
>>55584404
You are comparing between 2 diff screen resolutions though
>>55580494
because you little shit, the 280x or the 7970 was a powerful top en card at its time.
the 380x is based on 285 tonga, its not supposed to be faster but its extremely more power efficiant.
>>55584445
No but its a different source.
>>55584503
I'm not taking a series of cherry picked images put together in paint or gimp seriously. And anyone who does is an idiot.
>>55584404
>That's why the performance gap between the 560 ti and HD6870 has increased
AMD crippled the 6000 series pretty hard. They even tried to force upgrades in mid-2012 and 2013 with gimped driver updates, like making CCC instantly crash/close when ran (you had to reinstall OS completely to fix this 'bug'), same with the random BSODs and fan stoppages.
>>55584517
Don't forget the rebranded 390 and 390x getting exclusive driver improvements over the 290 and 290x :^)
>>55584554
Jesus Christ...
No wonder AMD fans are always so angry, kek
http://www.pcworld.com/article/2947294/components-graphics/asus-strix-radeon-r9-390x-review-keeping-up-with-the-high-performing-joneses.html
>>55584554
They sure did. Oh wait, no they didn't.
It really is the exact same silicon at higher clocks.
>>55584868
>y AMD no give us a nano
>>55584897
>>55584868
Holy shit AMD, get your shit together
>>55585875
Fun fact: during the 300 series AMD gained marketshare against Nvidia. Not much admittedly, but some.
Rebranding really did work.
I really want DX12 to become the industry standard by the end of 2017 and for Vulkan to get some traction.
And before nVidiots yell at me "SHILL" no, I am not. But if AMD becomes competitive again the industry will advance faster and we will get better prices. Look at the CPU market and see where AMD incompetence has brought us. Stagnation and gouged prices.
>>55584372
>>55584554
>using the smiley with a carat nose
>>55586266
There is literally nothing wrong with Intel. They've improved their CPUs year after year while keeping prices low. Meanwhile AMD has done nothing.
>>55587809
While a direct comparison isn't possible due to lack of L3 cache, roughly speaking the cpus in AMD's latest apus are considerably more powerful clock for clock than their FX lineup.
>>55587773
It's spelled "carrot", dumbfuck.
>>55587845
I guess the spelling depends on how jewish you are.
