1060 vs 480 Benchmarks

>the 480 even loses in Ashes of the Singularity
Holy kek AMD is finished.

>>55684800
j..j..just wait f-for the d...d..driver update op.
> still believing catalyst has been or ever will be anything but the hallmark of shit since bitcoin mining with gpus.

>>55684800
oh cool another shitty review came out

I'll just stick with hardocp and guru3d instead of this shill crap I've never heard of

File: 4AcDL5D.png (108KB, 1596x635px) Image search: [Google]

108KB, 1596x635px

>RIP 1060

>AotS with 25 more FPS on the 1060
I sure believe your fake graphs, OP!

>>55684894
Seriously you know it's bad when AMD can't even win in their own sponsored games.

>>55684800
It's a shit card, shit at DX11 and even shit at DX12, without massive overclocking.
AMD dissapoints again and again, its pathetic.

>>55684970
It's not fake, source is here: http://www.hardwareunboxed.com/nvidia-geforce-gtx-1060-benchmark-review-25-games-tested/

>>55685001
To clarify, I didn't mean you faked the graph on photoshop. I mean the measurements themselves are fake.

>>55685027
Everything is fake when nvidia wins, right? The RX 480 is a ultra power card, it beats a 980ti and gtx 1070!

File: Screenshot_2016-07-21-10-05-45.png (685KB, 1920x1080px) Image search: [Google]

685KB, 1920x1080px

>>55685027

>>55684968
based sapphire. When are they going to release that shit

File: images (3).jpg (4KB, 203x248px) Image search: [Google]

4KB, 203x248px

>>55685063
>The RX 480 only 1-2% worse than the 1060 in Ashes
>While being 5-20% cheaper

Thats not to bad I think.

File: 1353170076811.png (151KB, 274x238px) Image search: [Google]

151KB, 274x238px

>>55684968
>18% clock increase increases performance by 21%
Its like you're actually retarded

File: Screenshot_2016-07-21-10-04-42.png (498KB, 1920x1080px) Image search: [Google]

498KB, 1920x1080px

Also, threadly reminder that the 480 only gets massive gains when using hyperthreaded i7 or $2000 cpu. It loses its lead when paired up with a older cpus which many people who are considering this card will have.

>>55685039
No, it's fake when it magically goes against every single other review where AotS runs way better on AMD.
The fact that you just posted more strawmen just cements my conviction that you're a shill paid to post slandering the 480.

>>55685101
>£10 difference between aib 1060 and reference 480

I know, another solid win for the 1060. You'd have to actually be retarded in the head to consider the 480 at this point.

File: value.png (43KB, 500x1210px) Image search: [Google]

43KB, 500x1210px

>>55685101
There's actually only a $10 difference between them now.

1060 is the #1 best value GPU you can buy.

>>55685103
>Reference card throttles hardcore in games, AIB card must do the same.

Undervolting ref 480's give a performance increase by reducing throttling in the same way. Retard.

>>55685105
i7: from 88 to 121= 38% improvement

7 year old i5 CPU: From 63 to 81=30% improvement.

You must have to many chromosome

Low quality reference 480 vs founder's 1060 with premium materials.

File: Ashes_01.png (73KB, 1304x1616px) Image search: [Google]

73KB, 1304x1616px

>>55685116
>i-it's all a lie!

>>55685173
Rekt

There's basically no reason to buy a 480 now that the 1060 is out

File: w8NAN3k.png (110KB, 1591x639px) Image search: [Google]

110KB, 1591x639px

>1060 clocked to the tits 1fps faster than an AIB 480
>clock the 480 a little more and the 1060 gets destroyed

Nvidia, the way gamer are meant to be played.

>>55685151
No one is claiming the 480 gets a boost. If you didn't lack basic reading comprehension and weren't blind, you'd easily see that the 1060 has a lot more to offer with these lower end cpu and that it takes a convincing lead.

File: Feels_good_man.jpg (20KB, 329x357px) Image search: [Google]

20KB, 329x357px

>post yfw you bought a Pascal GPU.

>>55685173
>Maybe if I pretend like he's stuttering he'll go away
Next time when you post lies try not to pick the one game that has been reviewed as running an order of magnitude better on the 480 to fake

File: maxresdefault.jpg (189KB, 1920x1080px) Image search: [Google]

189KB, 1920x1080px

>>55685209

>>55685211
>the one game that has been reviewed as running an order of magnitude better on the 480

Please, show me all these incriminating benchmarks from reputable sites which show it to be "magnitudes" better (eg 20+ fps). I'm expecting at least 5 sources as there have been about that many showing the 1060 is better in aots, in this thread.

>>55684800
1060 will be literally shit compared to 480 in 2 years, screencap this.

>>55685124
There is a 4GB version Rx 480 Retard. Thats 199. Now lets do the math. 199 to 250

File: ff1.gif (112KB, 255x231px) Image search: [Google]

112KB, 255x231px

>>55685137
>Not having the 4GB version because it would be off the charts

File: gtx-1060-bench-ashes-1080pfps.png (37KB, 712x482px) Image search: [Google]

37KB, 712x482px

>>55685211
>i-it's all a lie!

>>55685326
>4GB

What happened to all the amdrone shills who said 4gb is dead? There were literally threads on here for months back in the 970/390/980/390x time when they were all screaming from the rooftops about muh 8gb, yet now it seems they're contradicting themselves. I guess I'll take their own facts and say 4gb is dead too, I mean, they did say it after all. :^)

Tldr; rx480 4gb is worthless trash according to amd fans themselves.

>>55684800
Nice try nvidiot it doesn't matter because amd users only play ashes of singularity.

>>55685410
At least it's actually 4GB, instead of 3.5

>>55685116
What? Every reliable review has the 1060 either a couple of frames ahead of the RX 480 or a couple of frames behind. I haven't seen a test from a no-shilling website yet where the difference is huge.

File: fc,550x550,black.u1.jpg (31KB, 550x550px) Image search: [Google]

31KB, 550x550px

>>55685447

You mean
> At Least its 4GB instead of 8 Oh wait...

>>55685447
The exact response I was waiting for.

Can you please show me pictures of the 970 PCB with only 3.5gb vram modules? :^)

File: pepe1.gif (1MB, 640x360px) Image search: [Google]

1MB, 640x360px

T-THIS CAN'T BE HAPPENING! I'M IN CHARGE OF THIS BOARD! FUCKING NVIDIAFAGS CAN'T Y-YOU JUST LEAVE ME ALONE! YOU BULLIED ME IN SCHOOL FOR BEING A POORFAG, YOU BULLY ME ONLINE AND NOW YOU BULLY ME ON MY PRECIOUS IMAGEBOARD!!! REEEEEEEEEEEEEEEEEEEEEEEEEEEEEE FUCKING NVIDIAFAGS LEAVEEEEE
>>>/v/
>>>/v/
>>>/v/
>>>/v/
>>>/v/

File: 7e871585c3bfcffd78b50f81abb415c3.jpg (7KB, 236x236px) Image search: [Google]

7KB, 236x236px

>>55685484

>>55685484
ebin

>>55685500
kek amd poorfag

>>55685211
>>55685293

Where are my sources? I thought you said "it magically goes against every single other review where AotS runs way better on AMD."

...

>>55685435
kek

>>55685151
You're the retard here.
Going from a 6700k to a 750 with the GTX 1060, you lose 10% performance.
Doing the same with an RX480, you lose 29%.

>>55685700
Dat AMD driver overhead is a bitch

>>55685715
Yep.
Strangely, Vulkan doesn't appear to reduce the driver overhead penalty. It*s actually 33% in Doom according to that graph.

File: Twittercom+missangerist+with+replies+gt+had+enough+of+people+_7cd68fd68a320801a994e81f501a4eed.jpg (9KB, 200x200px) Image search: [Google]

9KB, 200x200px

>>55685700
>You actually thinking the left is is DX12 while the right is DX11
>Or you just that stupid with math.

Dark color=DX11
Light color=DX12

Why are you faggots comparing a $250 maximum (aftermarket cooler and branded) card to a $250 bare minimum card for a reference cooler and less RAM? Why are most of the people in these threads shit flinging retards?

File: 249-1060.png (103KB, 978x324px) Image search: [Google]

103KB, 978x324px

>>55685769
Uhh the aftermarket 1060 is $249, pic related

Why do people keep posting these threads? Nvidia is better than AMD? Oh wow, no shit, Sherlock. Why not compare BMW and Lada while we're at it.

>>55685758
Nothing I said in that post had anything to do with DX11 vs DX12 (it's not even DX12 but Vulkan you twerp).
Learn to read and then learn to think logically because clearly you aren't capable of either.

>>55685782
What the fuck, where did I hear that 250 was reference only. Well fuck me huh.

>>55685758
he's right for the most part. going from the 480 + i7 at 121 fps to the 480 + i5 at 81 fps is a difference of 39%.

doing the same thing with the 1060, going from i7 + 1060 at 104 fps to i5 + 1060 at 103 fps is only a 0.9% difference.

basically the 480 loses nearly 40% of performance in vulkan when shifting between the newer and older cpu whereas the 1060 barely loses 1% of performance, also in vulkan.

>>55685829
Aftermarkets are €269 in Europe too.
Same as the shitty rx480 refs.

I actually ordered a 1060 today after I heard that there won't be a price drop and Sapphire Nitro will cost €299.

Pajeet fucked it all up.

NVIDIA ALWAYS WINS BABY

>>55685963
in my eastern euro cuntry, the 480 and 1060 cost the same, but is it worth it buying the cheapest gainward or zotac brand?

>>55686027
Zotacs have decent enough coolers for 15% OC.

Yes, they cost the same, but the point is that nvidias cost the same as the reference radeons with the shitty coolers.

Boy I'm glad I'm not poor. Its so good getting rich, you just get the best shit and don't have to fuck around and count pennies and see which GPU you have to get for your money and which is best etc etc.

I'll be getting 2 1080Tis when they come out. It'll be better than anything AMD offers.

>>55686027
Zotac is fine. I don't know much about Gainward.

File: Shekels_distribution_per_shill.png (153KB, 1298x604px) Image search: [Google]

153KB, 1298x604px

Shekel distribution in an arbitrary unit multiplied by 6 thousand shekels.

Here we can see TH's in the leading position for the maximum absolute amount of shekels collected, while, on the other hand and quite surprisingly, [H] losing as many shekels as the average profitability for the majority of reviewing sites.

Tom's Hardware coming out on top with a whopping total of 87,6k Shekels
Hardware Canucks with 54k shekels, which looks quite meager in comparison to the first place and followed very close by Hexus with 53,16K Shekels in the third place.

File: behindthispost.jpg (135KB, 576x1024px) Image search: [Google]

135KB, 576x1024px

>>55686163

>>55686163
If anything common logic says that if 8/9 reviews favour 1060GTX then HardOCP is the shill website.

> 480 recommended price by AMD: 199$
> 480 actual price by vendors: 249$
> 1060 recommended price by nVidia: 249$
> 1060 actual price by vendors: 249$

Nice bribing.

>>55686209
By that logic both low and high outliers should be ignored, which means the result stays pretty much the same.

>>55686163
Tom's Hardware have always been shills, these fuckers actually shilled for the GeForce FX back in the day and recommended it.
The one generation ATi completely and unquestionably dominated with better hardware and software, and TH shilled for Nvidia. All you need to know.

>>55685137
You missed the fact that the Rx 480x inputs were normalized with the current prices (average price gouging). While they decided it was a good idea to levae the MSRP for the 1060, the top score @ 120%.

It's the same kind of shillery than using Vulkan openGL in order to skew the total average performance charts - hint a 20% makes a huge dent on a mean hovering around -5% to +5% (on 15 or 20 sample points)

Well duh 480 is half

>>55685189
This is fake bullshit, you'd think AMDummies would at least learn their lesson from the last time they bought into this FUD

>>55686373
lol i still remember when those fake benchmarks came out from some chink site claiming a 480 was faster than a titan x at 1400 mhz in dx11.

>>55686209
Ignoring the voice of reason amongst the shillmongers.
Ignoring the only angel lost in the midst of devils.

yeah, I know, got lazy to quirky a quirky entry for them. And they baffled me with their review, they're probably shilling FOR AMD to get back in line to exchange for sweet review samples. Which is weird, because I doubt they don't have the funds to get the cards themselves. maybe there are other things going on.

>>55686373
Every review is fake.
Too much money involved to let things like neutrality get in the way.

You should get a sample of 5 AMD shills and 5 Nvidia shills and renormalize the scores to get a good grasp on things.

The only problem is that we don't have 5 AMD shills...

>>55686470
we have like half the board on /g/ who are AMD shills

>>55686491
Popular, well known ones.
On youtube the only one I can remember is Adored, as for tech sites I don't think I could name a single one.

>>55686396
Lol seriously. So much fake AMD hype. Then the real results come out and it's always disappointing. That's reason enough to buy Nvidia, at least they actually back up their results.

Im new to graphics cards and I want to get a 1060. There are different versions like MSI, Zotac, and EVGA. Whats the difference?

File: 56qckDx.png (35KB, 1053x278px) Image search: [Google]

35KB, 1053x278px

Now, this is some next level shilling.
Publish review where card A roflstomps card B.
Release an errata a day later with corrected results.

http://imgur.com/a/AtDcb

Conclusion: the Rx 480 looks a lot more competitive now. But it's a shame because no one will bother anymore.

>>55687086
You can check the reviews, they all have little differences.

EVGA is a very solid brand, can't go wrong with them.

>>55687086
Evga = best nvidia vendor. best looking cards, generally cheaper than the gigabyte, msi, asus alternatives but perform just as well or even better.

msi = expensive compared to evga and zotac but offers barely anything more other than some software. you're essentially paying more for the brand name.

zotac = the best budget nvidia cards you can buy. good cooling and overclocking. you're paying less than everyone else but you do lose features like dual bios or custom pcb which helps with overclocking. you don't really need that stuff anyway unless you're an enthusiast.

tldr; if you're willing to spend more, buy evga. if you're only willing to spend a little amount then the budget zotac cards are for you.

File: download (7).png (68KB, 642x463px) Image search: [Google]

68KB, 642x463px

Asus Rx 480 Strix review
https://www.computerbase.de/2016-07/asus-radeon-rx-480-strix-test/2/#diagramm-call-of-duty-black-ops-iii-1920-1080

>>55687203
>stock 1060 slower than the stock 480 in witcher 3

damn, the more computerbase benchmarks i look at the more i lose trust in them. every other review so far has shown the 1060 to have a nice lead over the 480 in this game.

>>55687267
computerbase is a known Mohammadian AMD shill site

>>55687294
the more i see of their benchmarks the more i believe this. this is like the 10th benchmark from them that's been posted in the last couple of days that have really weird results. i mean,. since when was the 390x faster than the 980 in gameworks witcher 3? even amd fanboys won't claim this. also gtx 970 at 30 fps? the only way this could be happening is if 8xmsaa hairworks was on but this would in turn cripple amd performance too, instead of giving them a lead.

1. computerbase are either lying about their results
2. they're running with 8xmsaa on hairworks but they reduce tessellation in amd catalyst so the amd cards don't take a hit, but the nvidia cards do.

very fishy. i think i am not going to believe any computerbase benchmarks from now on. there's definitely something weird going on with them.

File: 13-1080.3289219759.jpg (389KB, 1620x1080px) Image search: [Google]

389KB, 1620x1080px

Battle of heatsinks / PCB's.
This is something we all can play together.
Except Sappire, XFX and EVGA.

Which one is better?

>>55687267
>>55687294
>>55687398
see
>>55686470

>>55687398
They aren't lying.
They're simply cherry picking best case scenarios (e.g., that one scene/level/whatever where the card they want to promote either performs the best or the one they want to discourage, the worst).

And don't cry foul, everyone does this. You're finding it strange because in this case they're doing it against your prefered brand.

>>55685347
+1 fps AMD masterrace, nvidia got rekt!

>not waiting until 2018 and buying a volta

You idiots are just throwing your money away.

File: techreport.png (325KB, 570x1352px) Image search: [Google]

325KB, 570x1352px

techreport bench

>>55686163
>not six million shekels

File: 9a8245fcaa.png (9KB, 659x115px) Image search: [Google]

9KB, 659x115px

>>55687693
Come on, even thought nvidia is richer than top 20 countries by GDP combined that's a lot of dosh to spend shilling on a mid range product.

File: amd dx12.png (2MB, 1450x3400px) Image search: [Google]

2MB, 1450x3400px

>mfw dx12

File: fuefueprices.png (299KB, 1171x535px) Image search: [Google]

299KB, 1171x535px

Hues gonna pay as much for the 1060 as they did for 970's special editions at launch or 980's at a discount...

>>55687586
what? i'm simply stating that they get skewed results which are massively different to all other reviews out there. how do they get very different results to everyone else which is suspicious. take their claim of the 480 stock being faster than a 1060 stock for example;

http://www.eurogamer.net/articles/digitalfoundry-2016-nvidia-geforce-gtx-1060-review
>480
61.2
>1060
68.4

https://www.techpowerup.com/reviews/NVIDIA/GeForce_GTX_1060/22.html
>480
51.4
>1060
57.8

http://www.guru3d.com/articles_pages/geforce_gtx_1060_review,23.html
>480
53
>1060
59

http://www.hardocp.com/article/2016/07/19/nvidia_geforce_gtx_1060_founders_edition_review/8#.V5ErJvmANBc
>480
44
>1060 (+ hairworks on)
47.6

http://www.techspot.com/review/1209-nvidia-geforce-gtx-1060/page2.html
>480
68
>1060
73

https://www.pcper.com/reviews/Graphics-Cards/GeForce-GTX-1060-6GB-Review-GP106-Starting-249/Witcher-3-Wild-Hunt
>1060
11% faster than the 480 on average (don't know how to get exact frame metrics from these shitty graphs)

http://www.hardwareunboxed.com/nvidia-geforce-gtx-1060-benchmark-review-25-games-tested/
>480
55
>1060
61

here are just 7 benchmarks from the first few pages on google which all unanimously show that the 1060 is faster than the 480. computerbase are getting such different results it's actually suspicious. stop being a fanboy here.

File: wtf.png (235KB, 1007x822px) Image search: [Google]

235KB, 1007x822px

>>55687831
scrap that, that's well within 390x and 980tis range, at current prices.

File: 15 - 1.jpg (108KB, 480x605px) Image search: [Google]

108KB, 480x605px

>>55687203
>They did not even try overclocking it to maximum
wtf is this shit?

>>55687917
>They did not even bother overclocking both cards to their maximum.
fixed

File: Amaama to Inazuma - 03 [1080p].mkv_snapshot_09.52_[2016.07.19_00.19.14].jpg (205KB, 1920x1080px) Image search: [Google]

205KB, 1920x1080px

>>55687863
it's no fair when they do it to meeeee~

>>55687966
who? i own an amd 390. i'm calling for fair benchmarks here. everyone keeps accusing benchmarkers of favoring nvidia yet all results i get in my own testing are generally on par with what benchmarkers say. fanboys like you are just deluded and are the cause of hate for everyone who dislikes amd.

File: 1442008494063.gif (46KB, 300x356px) Image search: [Google]

46KB, 300x356px

>>55687863
you're bossu must be really proud of you.

>>55688022
nvidiots belong on /v/ this is an amd board

>>55688068
i literally just said i own a 390. are you this dense? amdrone fanboys just get worse and worse.

>>55688022
oooh. The "I'm a proud AMD card owner" card mixed with an ad hominem and bandwagon.

Next one: burden of proof, for sure.
>>55688046

File: spec.png (103KB, 723x601px) Image search: [Google]

103KB, 723x601px

>>55688109
you mad?

>>55688068
>this is an amd board

Time for you go to go reddit

>>55688068
It's not about being a nvidiot.
I just don't like being lied to my face.

Pajeet said the card will cost $199 and perform at gtx980 level. Now performance issues aside - there are no 4gb versions at all and 8gb versions start at 269 euro, which makes $297.

That's outright false advertising.

>>55687966
Nice taste in /a/ there.

Help a guy out here, evga 1070 imported at £360 with an American friend discount, evga 1060, or Zotac rx 480 ?

I'm paranoid of warranties and want 5 years as I run stuff down to the bone, Ifigure once in my life I want to go balls to the wall on graphics card budget.

>>55688147
>those temps
AMD housefires

>>55688193
>zotac 480

are you high?

>>55688147
Damn, I forgot the obligatory timestamped pic/speccy.

You're just too good at it anon.

>>55688190
>´there are no 4gb versions at all


that's not true. in germany, mindfactory has them up in stock from time to time. last time i saw was yesterday at 12 oclock but they are sold out so fast that it dissappears again

it's still the 8gb bios modified version as a reddit user confirmed

File: 1444402921467.png (266KB, 680x583px) Image search: [Google]

266KB, 680x583px

>>55684800
S U M M E R

>>55688227
why so salty? because i'm an amd owner who doesn't shill for amd and that i wan't fair benchmarks? stay mad son.

File: w1q2w6bdbp9x.jpg (235KB, 2048x1367px) Image search: [Google]

235KB, 2048x1367px

>>55684800
With the GTX 1060 get me 60 fps in Dark Souls 3 with everything turned to maximum at 1080p?

>>55688259
want*

>>55684968
Anything over 60 fps but under 120 fps is the dead zone of meaningless performance differences.

>>55688193
you hue too? or yuropoor?
tfw balls to the wall is a burger mid range at best. At SLI titan x total cost in burgerslandistan.

I'd say RX 480 because I'm AMDrone.

>>55686211
>Being this stupid

File: 1374185783396.gif (719KB, 500x281px) Image search: [Google]

719KB, 500x281px

>>55688259
i'm not salty at all.
i'm just shitposting while you're on the clock.

>>55688255
As I said, that's half the issue. The second half the is the 269 euro is the reference price - no backplate, shit cooler, 6 pin power.

And AIB cards have been announced ranging from 289 to 319 fuckin euros. I can add a 20 and get a Sapphire R9 Fury. I can buy an EVGA OC 1060 for 10 euro less.

I'm a Radeon user and I'm mad enough to buy an nvidia.

>>55688276

Hm... The bossu is a grammar nazi.
>>55688046

Nice, Q&A is very important after all.
Shilling is serious bussiness.

>>55688225
Ignore me, I had a dull moment and mixed up things in my head, Zotac gtx 1060. Anything with 5 years is good but I don't know of any AMD card makers that do such a thing.

>>55688295
Neither as of June 23rd, I listed pounds and am a cheapskate.

It's like I want a good warranty because I don't believe in changing graphics card every 3 years or so. I don't give a shit if by year 3 everything is nearing 30fps.

>>55688326
i saw a msi r9 fury x at mindfactory yesterday for 350€

i was about to pull the trigger but i so desperately want 2 nitros for over 1070 performance

i really hope they won't cost more than 249€ or i will probably get a good deal on kleinanzeigen/craigslist with a small footprint fury card

>>55688313
>>55688338
>i-it's a shill pretending to be an amd owner

neck yourselves. amd is objectively the worse company out of the two and if i wasn't a poorfag at the time of purchase i would have never bought this shit tier housefire hardware. amd even jewed me out of my free game.

>>55686247
And by that logic we should ignore any standard deviation from the mean, which means the result also remains the same.

>>55688364
>for over 1070 performance
enjoy your handful of non-stuttering games

>>55685105
x4 955 still working good. nice

>>55688364
The fury nitros were 337 euros at mindfactory yesterday.
The only thing stopping me is that my PSU won't be able to carry it without reducing my CPU overclock.

Pajeet fucked up big with the 480. I'm just gonna wait a bit to see if he drops the AIB price to compete, if not nvidia it is.

>>55688373
The memes now.
You forgot the shit driver support, shillanon.

Bossu gonna be mad at you.
>>55688046

>>55688398
but muh multi gpu. also i gotta freesync monitor

i'm pretty much gaming only casual and vr stuff so i won't have a problem with stuttering as i either can run just a single card for rocket league or use both for some downscaled virtual reality with liquidvr

any other than that, i'm stoked about battlefield 1 and it's announced "premium multigpu support"

>>55685068
august 5th

File: 1342194175493.jpg (159KB, 520x601px) Image search: [Google]

159KB, 520x601px

AMD IS DEAD:

https://docs.google.com/spreadsheets/d/1Q4VT3AzIBXSfKZdsJF94qvlJ7Mb1VvJvLowX6dmHWVo/edit#gid=0

HOLY SHEET

>>55688418
yeah i also already thought about a pricedrop on pajeets side.

since the release of the 1070/80 i thought the rx480 seems a bit overpriced for what it is.

>>55688373
kys nvidiot

>>55687446
Zotac

>>55688453
I ain't clicking that shit nigga

File: click that shit.jpg (25KB, 540x960px) Image search: [Google]

25KB, 540x960px

>>55688475

>>55688475
>I ain't clicking that sheet nigga
ftfy

>>55686209
HardOCP's results stem from their selection of games. When you only benchmark 4 games, one of which is Doom under Vulkan and one is Hitman, it's pretty natural that RX 480 would wound up ahead. That said, their Tomb Raider results seem a bit weird, as they place 1060 only slightly ahead of 480, regardless of whether DX12 or DX11 is used, while all the other sites I've checked (guru3d, hardwarecanucks, pcgh, for example) place 1060 about 20% ahead with DX12.

wow. another video card thread. cool now you guys can game ever harder bro

>>55688227
>pretending to be nonchalant about getting rekt

I like your style.

>>55687893

Those MSI prices are fucking stupid, you can get a 1070 with that money

>>55688423
>owns amd but not bumlicking them
>shill but not a shill

amdrones everyone. and people wonder why i hate amd.

>>55688453
amdfags so far:

Denial ()
Anger ( )
Bargaining ( )
Depression ( )
Acceptance ( )

File: nvidia user.jpg (276KB, 1325x743px) Image search: [Google]

276KB, 1325x743px

>the zotac shill
never ever heard of this fourth rate cosmetic brand

File: nvidiadefenceforce.jpg (63KB, 574x328px) Image search: [Google]

63KB, 574x328px

>>55688285

>>55688537
fuck, shit

>Denial (x)

>>55688540
But the 980ti supports dx12, it lacks AC, you moron.

>>55688540
me on the left

>>55688540

>good warranty
>good overclocker
>great value

nope, better pay more for fuckall
fuck off EVGA/MSI shill

>>55688553
>But the 980ti supports dx12
how good of microsoft to invent a separate direct3d path and numeration just so nvidia can be "dx12 capable"

>>55688453
This is pretty nice.

I've been claiming that 480's supposed DX12 dominance has more to do with cherrypicking and posting the same two AMD sponsored DX12 games. 1060 and 480 being basically even in DX12 is as I expected.

Also illustrates the point I just made pretty well too >>55688509

>>55688461
It's kinda like seeing a vampire disguised as a jehova witness politely asking permission to enter your house. When you deny him, he loses his shit and bares his fangs.
>>55688373

>>55688540
Owned a Zotac 760 and now 980Ti. I never heard anyone shilling them before but the deals were so good I got them. Never had an issue with either.

>>55688453
>https://docs.google.com/spreadsheets/d/1Q4VT3AzIBXSfKZdsJF94qvlJ7Mb1VvJvLowX6dmHWVo/edit#gid=0

what am i supposed to be looking at here?

>>55688627
>Average % Increase from 480 to 1060 13.92%

So which is the better card to get the most bank for the buck? Which one is the most VR ready?

>>55688634
The 1060 in both cases as of now.
We have yet to see the 480 AIBs and if AMD drops the price.

>>55688627
1060 is about 14% faster than 480 overall.
1060 is about 15% faster than 480 in DX11 games overall.
1060 is about 1% faster than 480 in DX12 games overall (basically an insignificant difference).

>>55688654
So if I were to upgrade my graphics card now I should be fine with the 1060? I have a 550ti if that helps

>>55685105
>1060 wins even in vulkan for most users.

Do we even NEED more nails for AMD's coffin?

>>55688668
Yes, you would be fine. But I suggest waiting for monday, Pajeet will be having a speech and the AIB radeons will be rolling out.
If he shits himself again get a 1060.

>>55685105
What's the technical reason why AMD is so CPU dependent?

>>55688669
found the idiot that cant tell which bar is longer

>>55684800

>12% on average faster

We know this since the fucking beggining. What are you trying to achieve?

>>55684968

One game.

>>55688720
those two cpus on the bottom are more than 7 years old

>>55688723
Yeah, the idiot, its you.

>Average % Increase from 480 to 1060 in Vulkan 27.13%

>>55688747
why are you allowed to use a computer

>>55688285
Finally someone with a brain on /g/
But should probably be altered to ~70fps and ~150fps.

>>55688720
The 480 is so slow, the CPU has to do all the GPU work with its onboard.

>>55688762
Because Im smarter than you, you failed, you shill somewhere else amdfag

>>55688764
That's not how it works, sadly. There is never enough FPS.
Unless you own a 144hz monitor with freesync/gsync 60 fps is pretty unplayable.

>>55688681
Actually you should wait at least a month or so for the prices to stabilize, if you can.

In doing that you'll allow AMDrones time to scavenge some FUD to shit on the 1060's. If that's not possible, the reviewers will readjust their scores later on, because they'll already have spent their hard earned shekels and naturally fleece nvidia for some more.
>>55686163

File: Command-Buffer-DirectX-121.png (167KB, 959x541px) Image search: [Google]

167KB, 959x541px

>>55688720

Their driver isn't multi-threaded and the gpu doesn't (basically) cache many instructions so it has to keep requesting work from the cpu and only on thread 0. The new command processor in polaris is aimed at reducing this overhead simply by being able to cache mroe instructions as it were so it won't keep slamming the cpu for work. Being single threaded is still a major issue under DX11 and older which is (partly) why AMD is all aboard the vulkan/DX12 hype train.

Nvidia's driver is actually multi-threaded in a way (despite in theory DX11 not really being suitable to it). In fact really optimising the threaded nature of their driver is how Nvidia gained so much perforemance in their infamous wonder driver a few years back (I forget which specific driver it was though).

>>55688785
nothing is unplayable realistically, people adapt and live with subpar performance
jarring at first would be a better phrase

>>55688787
Honestly I don't care about the benchmarks, it's clear that 1060 is a tad faster.
I just want 480 to sell at the MSRP price. If nvidia can sell at MSRP, even for the AIB cards, why doesn't AMD?

File: 249-1060.png (106KB, 1018x485px) Image search: [Google]

106KB, 1018x485px

>>55685782
can you actually buy it though?

>>55688793
Thanks m8, this is a very easy to understand a well written response

The 480 doesn't seem like a very good fit in a budget machine if you need to spend upwards of 300 dollars on a cpu, you'd save a lot more money just getting a 1060 and not having to upgrade your cpu for more performance

That's how it looks anyway

>>55688793
>Nvidia's driver is actually multi-threaded in a way (despite in theory DX11 not really being suitable to it).

You shit eating retards need to stop talking out of your asses about things you're woefully ignorant on. How D3D handles draw calls between the CPU and GPU has literally nothing to do with the interaction of the GPU and its drivers between the system.

DX11 uses 1 CPU thread to communicate with the GPU, and uses others to hold queues. DX 12 improves upon this by letting more threads pass draw calls.
This is API level. It has literally nothing to do with Nvidia's drivers.

Jesus Christ, end yourself.

looks like AIB won't fix the 480, it still overclocks just as poorly with a better cooler

https://www.techpowerup.com/reviews/ASUS/RX_480_STRIX_OC/25.html

this will be remembered as the first generation where amd failed to even put out a passably competitive product, the 480 is completely obsolete now that you can get a card that's 10% faster than it out of the box for only $10 more. OC it and the 1060 is close to 25% faster.

>>55688844
yeah, you can even get a $10 mib if you buy directly through the likes of evga

So can we all agree that 4xx line products are the worst products in consumer history?

>>55688881
>how do GPUs work

There is nothing to "fix." How well a die clocks depends entirely on the architecture itself, and binning. Certain dies before better than others with less voltage, some can function at higher clocks but with increased leakage. There is a reason why ASIC quality rating exists.

GCN has a short pipeline, it'll never be a high clocking arch. The RX 480 cards OC to about the same maximum that first gen Southern Islands cards could reach. A cherry die will hit about 1400mhz, anything higher than that is basically a legendary rare.

>>55688793
>Posting a picture and not even having a look at what it does
If you didn't look at it, it's literally spreading the runtime and driver (which is what AMD themselves write) load evenly across cores.

DX11 limits communication between CPU and GPU to a single thread and partly because of this both AMD's and Nvidia's drivers relied very heavily on a single thread. DX12 doesn't have this and as a result drivers are now not only leaner, but also much more multithreaded (like the picture you posted).

Seriously, please stop acting like you know shit because you've skim read something on hardocp...

>>55688881
HAHAHAHAHAHAHAHA

what happened to "all aib cards with 8pin will reach 1500mhz and btfo the 1060"?

File: 31a.jpg (204KB, 1024x696px) Image search: [Google]

204KB, 1024x696px

>>55687893
This is what a 838,10975 usd PCB in Huelandia looks like.

*exaggerations may apply - I took the highest price listed there divided by the current exchange rate - 2.479/3,28. The pcb is obviously for the reference card you silly. Couldn't find Ref cards yet, nvidia site doesn't sell FE down here and I don't think we'll get rebranded refs.

>>55688928

Do you not see it clearly states that is under DX12?

>>55688720

Not exactly CPU dependant, more like DX11 API calls are handled by a single thread, which is ironic considering their cpu's rely on more cores.
It's not that they don't have the horsepower, it's that since they developed GCN they're simply focusing on the newer API's that inherently offload work onto multiple cores

>>55688793
preemption =/= Async

When 40 percent looks like 80 percent

>>55688881
We already know Asus did fuck all to improve the 480. It's cooler doesn't even sit on the processor correctly. Wait for Nitro+ OC and Powercolor Devil benchmarks.

>>55689002

None of that post is related to either preemption or async compute.

>>55689038
>Nvidia's driver is actually multi-threaded in a way

>>55689018
>Nitro+ OC

OCuk announced it to be 260 british pounds - 343 US fucking dollars.
Just get a R9 fury or a 1070.

>>55688900
I think Nvidias Fermi, 4XX and 5XX line was worse, but AMDs 3XX, Fury and 4XX is pretty shit tho.

>>55688999
So, it still benefits from the faster core speed of the CPU, so in a way still CPU dependent? I'm just seeing that their GPUs suffer greater penalties with slower CPUs compared to Nvidia

Looks like RX 480 vs. GTX 1060 is decided in the Nvidia´s favour. Now I´m just hoping the Vega cards will be competitive, or I´ll probably go with the 1060.

gg no re AMDelusionals

It's another Nvidia NEET thread. Best part is figuring out who has the most college debt and thinks they're a system builder.

File: XFX-Radeon-RX-480-Black-Edition-10.jpg (278KB, 790x642px) Image search: [Google]

278KB, 790x642px

>>55689018
>>55689086
I didn't see Nitro's PCB pics, I think.
Kinda bummed that XFX just slapped a (good) heatsink on the reference board.

Asus at least redesigned the PCBs. And they're even marketing on the pci-gate bandwagon.

File: ASUS-RX-480-Alloy-Power-2-1-900x506.jpg (103KB, 900x506px) Image search: [Google]

103KB, 900x506px

>>55689161

>>55688881
'Overclocking results listed in this section are achieved with the default fan and voltage settings as defined in the VGA BIOS. We choose this approach as it is the most realistic scenario for most users.'

No! Fuck off! Shitty site can't do a proper OC. Doesn't adjust power limit. Doesn't up the mV. What a pile of shit review. If we cannot see what it's potential REAL overclock is how are we to know? Jesus H fucking Christ I hate that site.

AMDpoors literally on suicide watch

>>55689117

Nothing wrong with that, but personally I'd wait for actual results from aib 480's to come out first. If the rumors hold true, the gains on polaris from increasing frequencies should be a lot closer to linear than pascal.

File: 193223229083.jpg (21KB, 203x249px) Image search: [Google]

21KB, 203x249px

>>55689154
>AMDPOOR DETECTED

>>55689114

In a way, yes. Slower core speeds should hold the gpu back more.

>>55689184
Haha, wow. It's like they can't be seen putting nvidia's competing card in bad light.

File: RX-480_RED-DEVIL-F.jpg (569KB, 1613x2288px) Image search: [Google]

569KB, 1613x2288px

>>55689161
The devil cards also look like they're going to be reference pcb as well.

>>55689154
me on the left, senpai

>>55689184
they do the same with all cards. they did it with the msi 1060 too. i'm sure the 1060 has much more to offer if they increase the voltage and power targets.

https://www.techpowerup.com/reviews/MSI/GTX_1060_Gaming_X/27.html

>>55689086
Cheapest 1070 I can find is £431.99. I could get a Fury for £300 but it will be running only on 4GB and power hungry and huge! I want something going forward with more memory for larger textures etc.

File: smug3.jpg (93KB, 567x567px) Image search: [Google]

93KB, 567x567px

>>55689254
>litarally targeted 12yo with xmas money

>>55689018
>wait

Ah yes AMD's favorite mantra.

>>55689272
Then get a EVGA 1060 and buy a new HDD for the change.

>>55689272
>lying on the internet

https://www.overclockers.co.uk/kfa2-geforce-gtx-1070-ex-oc-8192mb-gddr5-pci-express-graphics-card-gx-091-kf.html

>>55689207

Yeah, if the AIB cards level the playing field with the 1060, I´ll probably with one of those.

It´s quite dissapointing that the AIB cards weren´t released on launch day though.

>>55689295
>edgy 12yo's.

When I was 12 I'd probably be scared senseless of that card.

>>55689254
>8 pin power connector
>Clear changes to power delivery
Yeah... It's clearly not a reference board. They may not have completely changed the layout of the board, but they have fixed both of the main issues with the reference card. i.e power and cooling.

>>55689267
Which is what I want to see. A GTX 1060 AIB pitted against a RX 480 AIB at *ACTUAL* overclocks. Not this power limited boost clock bullcrap.

>>55689314
If I am gonna spend over £350 I want something that's well received by most people. The MSI Gaming X would also go with my MSI gaymen mobo but that's not the most important aspect. I can cheap out on other brands I know.

>>55689308
The only EVGA 1060 I can see on OCUK is the pre order single fan. Fuck that. If I get a 1060 I will push it to the maximum clock speed so it will need a better cooler than a single fan.

>>55687636
>techreport bench

Pretty sure that pic is just the one in the Nvidia marketing brochure they sent with the card.

>>55689254
>midrange card, 10.5 inch long

for fucks sake.

>>55689354
yeah, i'm not sure on that one.

The PCI connector shouldn't prove to be much difficult to change, especially considering that the reference PCBs already came with 3 12v circuit/power traces. The voltage regs seem awfully close to the REfs, but the 6+1 got me wondering (the refs are 4+2, iirc). The lenght also appears to be quite similar, sans heatsink.

No substancial proofs, I know. Just going by the general looks on a semi hidden crappy pic.

>>55689391
>I can cheap out on other brands I know.

just because the brand is less well known it doesn't mean you're buying a cheap brand. kfa2 are the same company as galax, which make some of the best nvidia cards with their HoF series. there's a palit gtx 1070 on ocuk for the same price as the one linked previously and palit are the same company as pny who make reliable quadro cards for nvidia.

>>55689437
techreport is very neutral site. even scott wasson founder of the site currently work with amd.

>>55688900
>So can we all agree that 4xx line products are the worst products in consumer history?

Nah, for all its faults it actually performs okay.

Worst blunder is the Intel Pentium 4. Took 5ghz overclock on the $1000 extreme edition, to be able to match and beat the 2.6ghz Athlon64 ($200) in every test.

>>55688927
>RX 480 cards OC to about the same maximum that first gen Southern Islands cards could reach. A cherry die will hit about 1400mhz, anything higher than that is basically a legendary rare.

Have we gotten cards with proper power delivery to test that? With a 6-pin connector the card is power constrained, the chip just cannot be few with enough watts for it to hit its OC limit.

>>55689254
>Card comes with its own hexagram
should make installing amd drivers easier

AMD quarterly financial results are in.

http://ir.amd.com/mobile.view?c=74093&v=203&d=1&id=2187336

>>55685063
>>55685105
im running an old A10 5800k, so would I bottleneck the RX480, or could i make the APU and GPU work together for an added benefit?
doe the GPU need to be from AMD to crossfire ti with APU?

>>55689494
I became an AMDrone after a printed paper, well known, magazine convinced me to buy a P IV instead of Athlon.

My cousin had an AMD and was butthurt jelly of my pentium. He only discovered that he had the way better processor long after both machines became scraps.

File: rx480-13b.jpg (157KB, 1920x1080px) Image search: [Google]

157KB, 1920x1080px

>>55689507
>the chip just cannot be few with enough watts for it to hit its OC limit.

You don't magically cheat an 8 stage pipeline.
AMD designed the die to be as simple as possible, tried to keep transistor count as low as possible, and even sacrificed density to ensure lower defect rate, balancing it vs total die area. Part of the simplistic layout of P10 is its 32 ROPs, and to compensate for that they pushed clocks as high as they could without limiting the number of dies that would hit those clocks. Thats also why the ASIC quality of cards on the market is so high, and why the voltages for the pstates are so high by default.

They did everything humanly possible to ensure that Polaris 10 had stupid high profit margins.

Its a factory overclocked card. There isn't much OC headroom beyond the stock 1266mhz boost state. Power delivery has literally nothing to do with it.

>>55689618
>Thats also why the ASIC quality *variability* of cards on the market is so high

fixed
50% difference between high and low according to the CPU-Z database

>>55689583
Your setup sounds so old it would probably just commit suicide with either GPU.

>>55689592
P iv + tnt riva ultra, iirc.
384mb ram

>>55689643
So the only way to ensure a good ASIC is to buy the expensive AIBs like Nitro?

>>55689646
just over 3 years old.
It handles most games on medium-high, some are playable on low (but not enjoyable) the real problem are those that require a GPU for physics simulations

>>55689757
AIB's probably have higher standards of compliance for the chips. Reference are usually bottom of the barrel binning.

>>55689757
That remains to be seen.
AMD didn't do any clock binning with the dies used in reference cards. They set margins so wide that most anything would pass validation, and set their default voltages high enough to account for the massive variation in all these dies. Just to ensure they'd have tons of cheap dies with healthy profit margins.

Manufacturing partners like ASUS and Sapphire could do their own clock binning, and hand pick some of the better dies. That would let them reach higher clocks at equal or lesser power than the reference RX 480s, or they might just say "fuck it."

>>55689804
Except at periods close to their launch, first to market ref cards.
I think they put out higher binned chips to sway the first wave of user benches and general discussions.

No proof whatsoever. Just a tinfoil suspicion.

>boot up tomb raider for the first time in months
>pc instantly crashes and resets

and this is why i fucking hate windows 10. dx12 can go fuck itself till it works properly. even hitman dx12 had boot problems which took like 30 mins to boot up.

dx12 = dead. the only people who will support this garbage are massive amd shills. vulkan is no better.

>>55689954
"werks on mah peesee" ;)

>>55689954
I love seeing nvidiots turning into paradoxical ludites when they're faced with Dx12/Vulkan.

>Always buying the best and latest gpus.
>Always inhibiting software development.

>>55690015
i own a 290 you fucking retard. why would i be playing in dx12 mode if i had nvidia you utter spastic?

>>55685410
Actually RX 480 4GB can be unlocked to 8GB.

>>55685124
I'm still going for the 480 or 490, I'm not paying 60€ more for for aib 1060 when I can get the sapphire turbo+ for almost reference prices.

-Yurofag

File: 1250060657412.jpg (58KB, 298x443px) Image search: [Google]

58KB, 298x443px

>>55690041
>AMD just puts 4GB of extra RAM on the card

>>55685410
4GB RX480? Never heard of that before, I only know that there is a 8GB version and a 8GB version that costs slightly more

>>55690070
It was easier and cheaper to manufacture only the 8GB version and flash a 4GB BIOS on some cards than have 2 production lines, one for each model. So yeah, they put 4GB of extra VRAM on the card.

>>55690041
No.
The first lot of reference cards were artificially limited to 4GB in VBIOS just to simplify production, and make it easy for reviewers to test both configurations.

Not all 4GB cards are like this.

>buying amd
>ever
https://www.youtube.com/watch?v=8lTGk0R12gs

>>55684800
>>>/v/abies
remember to install gentoo

>supposedly the benefit of the rx 480 is that it's "future proof" because of vulkan/dx 12 support
>people suggesting 4gb version

doesn't make any sense senpai

>>55690211
funny amd has atrocious linux drivers

>>55690265
Because you can unlock the 4GB version to become 8GB

>>55690301
That's not guaranteed.

File: 757322d247dee07bcc8b755e197129e4.png (40KB, 671x527px) Image search: [Google]

40KB, 671x527px

so which one should i buy? pic related is my current setup, will be buying an i5 6400 + graphics

>>55690358
>1920x1080

Whichever one is cheaper for you.

>>55690358
which ever one is available and better priced for your needs while also providing performance you want in games you play because they're so close to similar that only fanboys are splitting hairs at this point so just fucking pick one and roll with it

>>55690368
im planning to buy a 1440p dell in the future

I usually stick with Asus for graphics cards, if I go Gainward or Phoenix on the GTX 1060, am I buying a ticking time bomb or are they manageable as long as I don't overclock.

>>55684894
>>55684984
its the dx11 version, not the dx12 version

the 1060 loses in dx12 mode.

amd doesn't give a shit about the dx11 version since the dx12 version exists.

>>55690426
>the 1060 loses in dx12 mode.
Seems that they're tied if the 1060 is not a little faster.
http://www.guru3d.com/articles_pages/msi_geforce_gtx_1060_gaming_x_review,17.html

>>55689774
its the cpu. the cpu has similar single threaded performance of processors from 2009.

example is a single 4ghz amd piledriver core has the same single threaded performance of a single 2.6ghz nehalem i7.

your 5800k is based off of amd's piledriver core.

>>55685883
So, either AMD's driver is presenting a lot of overhead (In Vulkan, it really shouldn't), or there's potentially a lot of untapped power in the 1060. As I own neither card I wouldn't be able to verify this, but knowing the CPU usage during the test would be good.

File: guts980ti.jpg (4MB, 2745x1893px) Image search: [Google]

4MB, 2745x1893px

>>55690515
very interesting.

though its actually sad. with the 1060 having more brute force than the 480 it can't beat a 480 in dx12 while it does beat it silly in dx11 mode.

just more proof nvidia is garbage in dx12 / vulkan.

>pick related
i'm a nvidia shill and its sad to see nvidia do so awful in dx12.

>>55690593
>with the 1060 having more brute force than the 480

no. the 480 has more physical cores than the 1060.

>>55690592
the way gcn works, the cpu's cannot feed the 480 fast enough to keep up with gcn's aces and cu's.

its why the 6700k gets a huge boost, it can.

this has been known for awhile.

regardless all 480's still received a boost, so dx12 low level is working as intended, but you won't see the bigger boosts until you get a faster processor.

all things considered, that 6700k turned that 480 into a $400 card from a $200 card.

the 1060 does fine because its the way pascal handles compute + graphics. software scheduling and preemption. the slower processors received a bigger boost because the 1060 kept up with them while received a lower boost because the 1060 cannot keep up with the faster processor.

>tldr
take a look at this thread to learn more how nvidia and amd architectures work with dx12.
http://www.overclock.net/t/1606224/various-futuremarks-time-spy-directx-12-benchmark-compromised-less-compute-parallelism-than-doom-aots-also

>>55690593
>awful in dx12
Competes evenly while using less power is your definition of "awful?"

>i'm a nvidia shill
Just like Falcon is an "nvidia shill" because he has a 970? I wonder if you pulled this image off the web to try and bolster a bullshit claim. From say a post on Overclocker's forum from March. . .

>>55690663
amd's shaders are not the same as nvidia cudas.

thats like saying amd's 8350 is better than a 6700k because it has 8 cores instead of 4 or a 2x 4 core 2 quads are better than a 6700k because they offer 8 cores instead of 4.

it doesn't work that way

>>55690677
even in time spy, where the 1070 and 1080 topped the charts, still did awful compared to amd.

all amd cards received a higher boost with async on than off than nvidia did. let alone maxwell received no benefit what-so-ever.

every single dx12 & vulkan benchmark shows how awful nvidia is. even tomb raider, where nvidia still won, still received either no boost or a fraction of a boost while amd had massive gains.

http://www.extremetech.com/gaming/231481-rise-of-the-tomb-raider-async-compute-update-improves-performance-on-amd-hardware-flat-on-maxwell

nvidia cannot get into dx12 & vulkan.

>>55690696
Shut the fuck up nvidiot.

>>55690668
Alright, so the RX480 is actually a very good card held back by hardware. As more games and faster hardware is released, the RX480 will scale to take advantage.

The 1060, OTOH, already operates at full capacity on even very modest hardware, and can't be expected to scale any going into the future.

This sounds to me like while the 1060 is the faster purchase *now*, but the RX480 will be more "future-proof".

I'm still in a wait-and-see approach, but if the situation hasn't changed very much before September I'll probably pick up an RX480 and OC my CPU as best I can.

>>55690696
Hence why the claim of the 1060 being a brute force card is wrong. The 480 is the brute force card just like all current amd cards. The fury x is the best example of this. It had over 4000 cores to compete with the 980 ti.

Basically it requires beefier hardware from amd to compete with nvidia ie brute forcing to get the same results.

>>55690731
architectures are what matters. shader count does not when comparing two different architectures to one another. they're only useful to compare in that cards generation.

File: 1468280350117.jpg (73KB, 960x960px) Image search: [Google]

73KB, 960x960px

>>55690070

>>55690728
>even in time spy, where the 1070 and 1080 topped the charts, still did awful compared to amd.
Where are you getting your information from?
https://www.youtube.com/watch?v=dTJPp7WdEBs
The 1060 beats the 480 in TS though only by about 6.5%.

>all amd cards received a higher boost with async on than off than nvidia did.
I wonder if that might be because Nvidia did a better job of efficiently using its cores than AMD. Concurrency only brings much if there are "empty bubbles." Anandtech has a pretty good write up on that.
http://www.anandtech.com/show/10325/the-nvidia-geforce-gtx-1080-and-1070-founders-edition-review/9

>>55690776
Seriously, just shut the fuck up, you have no fucking idea what you're talking about. You literally just pulled that shit out of your ass.

>>55690765
> the RX480 will scale to take advantage.
as long as the game is coded for async compute and to the specs of dx12 and vulkan.

see this for more info:
http://www.overclock.net/t/1606224/various-futuremarks-time-spy-directx-12-benchmark-compromised-less-compute-parallelism-than-doom-aots-also

on what i mean by:
>to the specs of dx12 and vulkan.
even if its done nvidia's way to help make up for pascal's short comings, amd will still receive a boost. time spy proves this since amd still received higher boosts than nvidia did. but amd can easily receive more when fed more compute + graphics.

>>55690785
>GEFO**RCE
Shit shoop is shit. Good idea, poor execution.

>POWERCOLOR RED DEVIL RX 480 8GB GDDR5: 269 USD, available July 29th

+ import tax +19% VAT in Europe.

Just great, I'm looking at €309 here.

Well done AMD, Pajeet just gave the mid range market to nvidia for free. Who thought it's a bright idea to hire a fucking indian?

>>55690800
He's pretty much right. It really is down to architectural choices like raster units, texture units or bus width etc which make one gpu faster than another. Nvidia does better with less because of their architectural design. Amds arch is designed to work better with more cores hence why they included ace's which enable full utilisation of those cores under these new api.

File: LL.png (109KB, 623x900px) Image search: [Google]

109KB, 623x900px

>>55690794
look at the percentage of gains with async on & off.

amd across the board received higher boosts with async on than nvidia did.

>>55690800
have your balls dropped yet or did daddy touch you when you were a kid?

>>55690895
>amd across the board received higher boosts with async on than nvidia did.
Where did I say that they didn't? Are you responding to my post without reading it?

File: gg-small-discordcert.gif (52KB, 500x362px) Image search: [Google]

52KB, 500x362px

>>55690677
Everybody is a nvidia shill by default.
You only get to call yourself an AMDrone after five thousand confirmed posts.

There's a huge party and a certificate delivery for it. Then, a nice hotdog buns feast.

>>55690895
Nobody is saying amd don't receive a bigger boost. I think the point being made here is that even with the boost on amd cards, the nvidia competitor is still faster. Kinda similar to
>>55685105
>>55685063

>>55690949
>Everybody is a nvidia shill by default.
Stopped reading here. Only a foolish AMDrone would believe anyone who doesn't agree with him is an Nvidia shill, ie. projecting his fanboy failings onto others.

File: 0b807cb9fb4f0b500dff33cdf6f6cc18ebc8a08d_hq.jpg (44KB, 600x450px) Image search: [Google]

44KB, 600x450px

>>55690830
...you simply can't be real.

>>55690949
Lol even called this amd owner an nvidia shill >>55688147

It's fucking crazy. You either shill for amd or you're an nvidiot...even if you own amd..

File: 1335329902641.jpg (60KB, 530x715px) Image search: [Google]

60KB, 530x715px

>>55690976
Nvidia shills are known to have poor reading and comprehension skills plus a short attention span.
indeed.

>>55690953
that doom benchmark shows the 480 shitting all over the 1060

ashes is the only odd ball out.

>>55690915
i wasn't talking about the 480. i was talking about amd cards as a whole. amd's architecture does a far better job at async than nvidia does. even that video you showed mentioned the 480 receiving a higher boost than the 1060.

the 480 overall will last longer as a dx12 / vulkan card than the 1060 will. the 1060 requires nvidia to heavily invest in their driver optimizations while amd doesn't. the tables have turned.

>hurrr its a throw away card
for a lot of people, they keep even their $150 cards for years to come.

>>55691081
>i wasn't talking about the 480.
Let me see if I understand this correctly. You responded to a post without regards to the topic of the post? Because the discussion before you started talking about something else was comparing the 1060 to the 480.

>>55691081
>that doom benchmark shows the 480 shitting all over the 1060
>loses 40% of performance gains when dropping from a top of the range i7 to an older i5
>1060 loses less than 1%

>shitting all over 1060

Top kek. You aren't stupid enough to assume that people who are buying budget midrange cards like these all have i7 6700k with hyper threading (which is what gives the 480 the lead), right?

>>55691081
>muh async
http://www.anandtech.com/show/10325/the-nvidia-geforce-gtx-1080-and-1070-founders-edition-review/9
>all amd cards received a higher boost with async on than off than nvidia did.
I wonder if that might be because Nvidia did a better job of efficiently using its cores than AMD. Concurrency only brings much if there are "empty bubbles." Anandtech has a pretty good write up on that.

>>55691081
>that doom benchmark shows the 480 shitting all over the 1060
you know, the 1060 would actually be doing worse than what was shown in there. async is disabled on nvidia cards. for nvidia, the boost is coming from less driver overhead. if they had async performance would of probably gone down. this is actually why ID disabled async support until nvidia releases their drivers for it because nvidia cards where facing regressions.

btw nvidia hasn't released their async drivers for doom vulkan yet and nvidia hasn't said a word.

amd on the other hand has async enabled.

>>55690668
is right, amd shines when it has heavier async load, its why the 480 skyrockets with the 6700k weaker processors, it still shines, but not as much, because they couldn't keep up. gcn does better when under a heavier async load, which is actually cool. fill her up more and she'll kick more ass.

with nvidia it was eliminating some of the driver overhead since they are heavily dependent on using a software scheduler for everything since they lack hardware. the 6700k showed less gains because the 6700k was more than capable of handling the driver overhead.

ironically, what amd faced with their dx11 drivers is now what nvidia is facing with their dx12 drivers. the more you know.

>>55691081
>that doom benchmark shows the 480 shitting all over the 1060
I wonder if that might be because Doom is using specific AMD shader optimizations that do not benefit Nvidia cards at all or that Doom doesn't support Nvidia's async at all?

>>55691174
http://www.overclock.net/t/1606224/various-futuremarks-time-spy-directx-12-benchmark-compromised-less-compute-parallelism-than-doom-aots-also

&

http://www.overclock.net/t/1605899/various-futuremark-releases-3dmark-time-spy-directx-12-benchmark/650#post_25363376

&

http://www.overclock.net/t/1605899/various-futuremark-releases-3dmark-time-spy-directx-12-benchmark/120#post_25348152

&
so much more

btw this guy was the one who helped bring the light about maxwell's lack of async.

File: eris.jpg (108KB, 1024x746px) Image search: [Google]

108KB, 1024x746px

>>55690949
Eris' only ever please when you're a Nvidiot in AMDrone threads and vice-versa.

>>55691224
see this guy:
>>55691192

then see all of these:
>>55691242

>>55691242
I'd believe a reputable reviewer who has given wins to both amd and nvidia in the past rather than a guy who many people have admitted is a salty fanboy who is obsessed with bringing down nvidia on some random forum

https://youtu.be/dTJPp7WdEBs

Good day.

>>55691242
see this guy:
>>55691192
Okay. So what? Nvidia hasn't said anything because ID admitted that there is a lack of async support for Nvidia's hardware.
>then see all of these:
>>55691242
What does TS's lack of vendor specific optimizations, or should I say vendor neutrality, have to do with the shader optimizations or lack of async support for Nvidia hardware in Doom?

>>55691174
>Now Nvidia has been claiming that they support Async Compute since Maxwell. Several developers have spoken out and told everyone countless times that Nvidia GPUs cannot support Async Compute, but that isn’t stopping Nvidia from having presentations on their “Enhanced Async Compute” with Pascal. So basically Nvidia created and actually showed “Enhanced Async Compute” while using DX11. As hilarious as that may sound, Nvidia is serious. They still have GPUs to sell and they must stay competitive, even if they have a large share of the market already.
>Nvidia “Enhanced Async Compute” is basically a newer and better form of preemption. The issue with preemption is that it is not asynchronous so therefore it cannot be async compute. Preemption is not parallel at all and can cause overhead issues. Preemption basically causes data to stop for more important or higher priority data that must make it to the GPU immediately.

http://www.overclock-and-game.com/news/pc-gaming/50-analyzing-futuremark-time-spy-fiasco

>>55691258
Oh, and I should have pointed out in regards to >>55691242 that you are a retard if you are trying to sincerely make a point about Nvidia's performance in Doom, which is Vulkan, by raising a comparison with TS which is not Vulkan but DX12.

Given how it seems you seem to be trying to defend AMD like a good AMDrone I guess that wouldn't be surprising.

I hate this fuckinh shill threads.

>>55691286
that video actually proved all of this:
>>55691242

that same video even says the 480 received a HIGHER boost than the 1060 did.

that same link you just dismissed in here:
>>55691242

states exactly why. time spy doesn't use much graphics async, mostly compute and even with the amount of graphics they used wasn't much. so it didn't overload pascal's handling of dx12.

but even with the less async usage amd STILL came out on top in the % of increase with async on than off. >>55690895 pic shows this and does that SAME EXACT VIDEO you just linked to did.

this fool >>55691192 specs truth. gcn shines with async. it shines even more with increased async load. maxwell can't into both graphic & compute at the same time. it has to stop one, then load the other, then stop again. pascal helps lessen this with preemption, but it still cannot fully handle a high load of graphic + compute. more than maxwell, shit anything is better than maxwell, but nothing compared to gcn.

File: trashman6.png (591KB, 778x1018px) Image search: [Google]

591KB, 778x1018px

>>55691292
>random blog

File: download (2).jpg (4KB, 200x144px) Image search: [Google]

4KB, 200x144px

Leave Nvidia to me.

>>55691368
Yeah i doubt you'd be able to read it anyway.
It has more than 2 paragraphs and not that many pictures.

>>55691306
vulkan and dx12 share virtually everything together and operate nearly exactly the same.

the perform the same identical function. so close you can actually share doc's between the two.
even id came out to state why they went with vulkan over dx12:
http://www.overclock3d.net/news/gpu_displays/id_software_explains_why_they_chose_opengl_vulkan_over_directx11_12/1

> DirectX 12 and Vulkan are conceptually very similar and both clearly inherited a lot from AMD’s Mantle API efforts. The low-level nature of those APIs moves a lot of the optimization responsibility from the driver to the application developer, so we don’t expect big differences in speed between the two APIs in the future.

>On the tools side there is very good Vulkan support in RenderDoc now, which covers most of our debugging needs. We choose Vulkan, because it allows us to support Windows 7 and 8, which still have significant market share and would be excluded with DirectX 12.

>On top of that Vulkan has an extension mechanism that allows us to work very closely with AMD, NVIDIA and Intel to do very specific optimisations for each hardware.

>>55691251
>Not going one step further and arguing with yourself in non-biased threads over stupid shit until other anons join in to fanboy and continue the argument without you

File: facepalm8bu0.jpg (183KB, 716x494px) Image search: [Google]

183KB, 716x494px

>>55691401
>vulkan and dx12 share virtually everything together and operate nearly exactly the same.
You just won the award for the stupidest post on /g/ today. Yes, both APIs provide "low level natures but how they go about it are not functionally vastly different at all . . .

File: tfw.png (91KB, 566x585px) Image search: [Google]

91KB, 566x585px

>>55691407
but I was talking to myself.

>>55691407
And that's anarchists cookbook stuff isn't it?
principia discordia is before the internet.

>>55691353
>shit anything is better than maxwell

What? Maxwell was amazing for what it offered in dx11. Retards seem to forget than when it released dx12 didn't even exist in 2014. Everything was dx11 (or ogl) till mantle came out and slowly became vulkan/dx12. Maxwell was brilliant for the api it was primarily designed for. And to be honest it still is brilliant today. Only spastics claim dx11 is dead when literally all the games coming out this year and some which have been announced for next year, support dx11. Maxwell owners will be just fine for a few years yet and pascal is a good enough stop gap till Volta comes out with full dx12_(3?) support. Feel free to hate them but nvidia have got their marketing and product lineup spot on in these last few years.

>>55691452
This

File: 1461932114132.png (108KB, 1415x954px) Image search: [Google]

108KB, 1415x954px

>>55691407
>>55691453
>implying /g/ isn't one person

>>55691452
It's pretty much like saying that OGL and Direct3D work the same..

File: 1452136383601.jpg (125KB, 800x363px) Image search: [Google]

125KB, 800x363px

>>55691546

>>55691407
>>55691495
TAZ
temporary autonomous zones.
Subversive poetic terrorism and shiz.

good times.

>He actually wants AMD to go out of business

Enjoy paying $500 for the GTX 1160 or whatever Nvidia shits out next generation.

Cue the poorfag name calling.

>>55692147

if amd goes out of business it's their own fault for not outputting a proper competing product.

>>55692147
Not paying 250 USD for 480 either. It is already overpriced comparing to 1060 as it is. If AMD doesn't cut it down to 200 then fuck them them it would not be my fault for not giving them my money in overpriced shit, it will be their god damn fault of not being able to be competitive.

>>55690830
I never noticed that, that is a pretty shit shop. The spacing between the O and the RCE is all fucked up.