Support NeoGAF

[ethomaz]

About the ROPs. GPU-Z deosn't read that info from the card, but has a database that collects that information. And a lot of times when a card launches, it is incorrect.
See the 480: https://www.techpowerup.com/gpudb/2848/radeon-rx-480
And the 470: https://www.techpowerup.com/gpudb/2861/radeon-rx-470
And the 460: https://www.techpowerup.com/gpudb/2849/radeon-rx-460

First of all, they still have the old names, Ellesmere and Baffin. We can see the 480 is Ellesmere XT and the 470 is Ellesmere Pro, which means the 470 is most likely a cut-down 480.
But if we see the number of ROPs on the card, the 480 has 32 and the 470 has 48. Which considering the 470 is cut-down, is most likely wrong.
if we consider the ratio of Shading Units is the same as ROPs, 1536/2304 equals 2/3, which will give 72/48 or 32/21.3, and I hope it's the first one.

470 > 480 confirmed by GPU-Z lol

Anyway I do expect 480 to have at least 48 ROPs.

 

[DieH@rd]

Doom, MP, non-K i5, again custom preset :-/
https://www.youtube.com/watch?v=0u6t3l7Wts4

 

[JClarke]

For someone like me running a 560ti and strictly gaming at 1080p (mainly just dota)

Which card would you gents advise ?

 

[Locuza]

Anyway I do expect 480 to have at least 48 ROPs.

The ROP and Memory-Controller logic is tied together in even ratios, so 48 ROPs coupled with 256-Bit will not work.
It's either 32 or 64.

 

[Dezeer]

The ROP and Memory-Controller logic is tied together in even ratios, so 48 ROPs coupled with 256-Bit will not work.
It's either 32 or 64.

It hasn't been tied previously. Tahiti, 7970/7950, is 384-bit bus with 32 ROPs.

 

[Locuza]

You are right, AMDs GCN design is a more flexibel.
AMD says that they can build 1-4 Render-Backends per Shader-Engine.
Ranging from 16, 32, 48 to 64 ROPs if Polaris does not change anything and is using 4 Shader-Engines, what it probably will.

 

[SneakyStephan]

Doom, MP, non-K i5, again custom preset :-/
https://www.youtube.com/watch?v=0u6t3l7Wts4

Anyone with doom and a gtx 970 or r9 390 wana upload some footage of the same map at the same settings?

 

[Guess Who]

For someone like me running a 560ti and strictly gaming at 1080p (mainly just dota)

Which card would you gents advise ?

As someone with the same card, literally any of these cards will be a notable upgrade.

 

[cyen]

Anyone with doom and a gtx 970 or r9 390 wana upload some footage of the same map at the same settings?

Doom can be vou limited in certain cenários and he has a weak cpu.

 

[Tacklemacain]

And probably the wrong drivers.

 

[ethomaz]

The ROP and Memory-Controller logic is tied together in even ratios, so 48 ROPs coupled with 256-Bit will not work.
It's either 32 or 64.

That is not true because previous cards shows the opposite (not tied to memory controller).

 

[SneakyStephan]

Doom can be vou limited in certain cenários and he has a weak cpu.

He was at 100 percent gpu useage for the entire video

 

[Last_colossi]

Anyone with doom and a gtx 970 or r9 390 wana upload some footage of the same map at the same settings?

As I said earlier it's useless looking at these videos as benchmarks as the official drivers haven't been released yet and that can improve performance up to 50% like it did in the video I posted.

Right now all we can do is wait for official reviews (near the 29th) and until then everything needs to be taken with a grain of salt.

 

[Irobot82]

Doom, MP, non-K i5, again custom preset :-/
https://www.youtube.com/watch?v=0u6t3l7Wts4

Wow TSSAA at 8x. It's stuff like this that gives me hope because I don't need that much AA especially at 1440p.

 

[ZOONAMI]

Hmm so the dr got banned

Why is this guy who seems to be quite the pc enthusiast benching everything with a non k i5?

 

[icecold1983]

Wow TSSAA at 8x. It's stuff like this that gives me hope because I don't need that much AA especially at 1440p.

it has no performance hit

 

[tuxfool]

it has no performance hit

It does, but nothing significant. There is no reason to play doom without TSAA

 

[cyen]

He was at 100 percent gpu useage for the entire video

My bad. Didnt noticed, it happens to me on a 1080 and a 5820k.

 

[Coulomb_Barrier]

http://wccftech.com/amd-rx-480-gaming-benchmarks-leaked/

Performance looks fantastic around Fury to 980 levels. This is the performance I expected from the start. Imagine what the AIB overclocked cards will do.

I have the suspicion that this card might force Nvidia to get their retail partners to start selling the 1070 closer to the official RRP. Seems to me there will be a lot of 480 stock at launch.

 

[Beerman462]

Can we get the RX480 added to the thread title since this is this is where we discuss it now.

That or

http://wccftech.com/amd-rx-480-gaming-benchmarks-leaked/

Performance looks fantastic around Fury to 980 levels. This is the performance I expected from the start. Imagine what the AIB overclocked cards will do.

I have the suspicion that this card might force Nvidia to get their retail partners to start selling the 1070 closer to the official RRP. Seems to me there will be a lot of 480 stock at launch.

New news new thread?.....nope

 

[ethomaz]

http://wccftech.com/amd-rx-480-gaming-benchmarks-leaked/

Performance looks fantastic around Fury to 980 levels. This is the performance I expected from the start. Imagine what the AIB overclocked cards will do.

I have the suspicion that this card might force Nvidia to get their retail partners to start selling the 1070 closer to the official RRP. Seems to me there will be a lot of 480 stock at launch.

What article... lol

He says the test avg. ~100fps then you watch the video and it goes down to 50-60, have some picks of 100-120 and stay most of time in ~80 (to be fair only at beginning and near the elevator of the video it reach 100+ fps).

 

[MagicWithEarvin]

Has anyone experienced both FreeSync and G-Sync? I have a FreeSync 4K IPS panel and was planning on getting a 480 8GB so I can actually use the FreeSync.

 

[Marmelade]

I have no idea how they can say it's around 980/Fury performance from these vids...
Doesn't look like it at all

 

[ethomaz]

I have no idea how they can say it's around 980/Fury performance from these vids...
Doesn't look like it at all

Because the dude that wrote the article (and made the claim) used max fps like it was avg. fps... Doom for exemple in the video runs at avg. 80fps while the article says it is running at avg. 100fps.

There is close to 20% difference between what he wrote and what the videos shows.

 

[Caayn]

Has anyone experienced both FreeSync and G-Sync? I have a FreeSync 4K IPS panel and was planning on getting a 480 8GB so I can actually use the FreeSync.

Here's a comparison between Freesync and G-sync with a high speed camera and some other equipment.

 

[JRW]

I'm curious how the RX 480 would compare to my R9 290X performance wise.

 

[TheFatOne]

http://wccftech.com/amd-rx-480-gaming-benchmarks-leaked/

Performance looks fantastic around Fury to 980 levels. This is the performance I expected from the start. Imagine what the AIB overclocked cards will do.

I have the suspicion that this card might force Nvidia to get their retail partners to start selling the 1070 closer to the official RRP. Seems to me there will be a lot of 480 stock at launch.

Need a better source. WCC is pure garbage.

 

[roadrunn3r]

http://wccftech.com/amd-rx-480-gaming-benchmarks-leaked/

Performance looks fantastic around Fury to 980 levels. This is the performance I expected from the start. Imagine what the AIB overclocked cards will do.

I have the suspicion that this card might force Nvidia to get their retail partners to start selling the 1070 closer to the official RRP. Seems to me there will be a lot of 480 stock at launch.

Pure wccftech quality article right there. They say "which is significantly faster than what the R9 390X delivers at even lower settings" and then link to a benchmark done with different settings, different sections of the game and most importantly with the 390X showing a minimum of 65 frames when in this video the 480 goes as low as 57.

As much as I'd love the 480 to be good, these articles are pure bullshit and any random forum user could do the same really or probably better. How that clickbait site isn't banned around here (it is on a lot of other places) is beyond me.

To add to that, drivers aren't final also and even well done benchmarks right now wouldn't tell much let alone a guy with a 2.7 Ghz i5 running games with custom settings.

 

[opticalmace]

Has anyone experienced both FreeSync and G-Sync? I have a FreeSync 4K IPS panel and was planning on getting a 480 8GB so I can actually use the FreeSync.

The 480 is going to struggle at 4k, depending on which games you play and what settings you use/are okay with.

 

[roadrunn3r]

Unfortunately nothing official. Just rumors. But Primitive Discard Accelerator is AMDs solution of getting around the exorbitant amount of unnecessary tesselation GameWorks introduces to cripple performance, at least that's my take on it.

I think the more telling story is the additional headroom left untapped that Raja Koduri pointed out when he demoed AoS benchmarks at Computex. He said additional driver improvements and developer fine tuning will increase performance. So the 10% performance improvement from newer drivers isn't entirely out of the question.

Yup the tessellation improvements are very interesting to me, can't wait for a proper architecture review from the usual sites.

 

[Irobot82]

it has no performance hit

since when does AA not have a performance hit? Is this something similiar to SMAA or something?

 

[icecold1983]

since when does AA not have a performance hit? Is this something similiar to SMAA or something?

its a post process temporal aa.

 

[Sir Abacus]

since when does AA not have a performance hit? Is this something similiar to SMAA or something?

The samples are generated by subpixel jitter not by multisampling a pixel.

 

[Rushersauce]

http://wccftech.com/amd-rx-480-gaming-benchmarks-leaked/

Performance looks fantastic around Fury to 980 levels. This is the performance I expected from the start. Imagine what the AIB overclocked cards will do.

I have the suspicion that this card might force Nvidia to get their retail partners to start selling the 1070 closer to the official RRP. Seems to me there will be a lot of 480 stock at launch.

Jesus, the comments are cancer.... Nvidia fanboys are cancer (so is any fanboy, really)

 

[Renekton]

Kyle from HardOCP is going apeshit again on AMD Polaris, dropping bombs.

"Cannot reach 850Mhz"
"Delayed to October"

http://i.imgur.com/hxJP39k.jpg
http://i.imgur.com/2Iyr9tM.jpg
http://archive.is/a9wqA
http://archive.is/wjtIO

Edit: the 2 quotes were copypasta from reddit, I cannot find them after some searching so cross them off.

 

[Last_colossi]

Kyle from HardOCP is going apeshit again on AMD Polaris, dropping bombs.

"Cannot reach 850Mhz"
"Delayed to October"

What is he talking about?, It's already been pretty much confirmed the 480 runs at 1266Mhz and stores like newegg have already started listing the card for launch?

I'm honestly confused as to what your talking about.

 

[opticalmace]

Kyle from HardOCP is going apeshit again on AMD Polaris, dropping bombs.

"Cannot reach 850Mhz"
"Delayed to October"

Link to him saying those 2 quotes?

 

[ethomaz]

Kyle from HardOCP is going apeshit again on AMD Polaris, dropping bombs.

"Cannot reach 850Mhz"
"Delayed to October"

I can't read anything in this image


Edit - I did some search and this 850Mhz comment was made before May that could be rumor for even before that... well AMD had worked in the clock until mid June I guess.

Edit 2 - I could read the image and there is nothing related with your quotes.

Edit 3 - Now I understood the images... thanks... he is saying AMD GPU will collapse because Raja near the Polaris launch was talking about Vega in a event yesterday... too much assumption to my head lol

 

[Cow Mengde]

AMD confirms VEGA 10 GPU reaching development milestone

http://videocardz.com/61361/amd-confirms-vega-10-gpu-reaching-development-milestone

 

[ethomaz]

AMD confirms VEGA 10 GPU reaching development milestone

http://videocardz.com/61361/amd-confirms-vega-10-gpu-reaching-development-milestone

That is what Kyle was talking... Raja is teasing Vega before launch Polaris... he thinks he needed to show Polaris "strength" in this event.

While so agree with him about that his assumptions (collapse, etc) are really crazy talk.

Edit - Raja said that too "long way to go before you see it’".

 

[FLAguy954]

Newegg listings are up (towards the bottom of the page).

 

[ethomaz]

Newegg listings are up (towards the bottom of the page).

I did filter the 7 to make it easy.

 

[kami_sama]

I did filter the 7 to make it easy.

For now it seems only the XFX one has a backplate.

 

[WiiU0706]

So GAF if NX went with the 460 , would it still be better than current gen???

 

[DieH@rd]

So GAF if NX went with the 460 , would it still be better than current gen???

If clocks are high, it could match and beat PS4. But I dont expect the rest of the HW to match it (cpu & memory bandwith).

And eho knows what NX really is...

 

[roadrunn3r]

So GAF if NX went with the 460 , would it still be better than current gen???

I expect the 460 to at least match GTX 950 performance so it would be roughly 40% better than current gen.

 

[Irobot82]

Fresh look a the new Overclocking tool in the big Crimson update that will come with the new cards.

LOVE the design!

 

[ethomaz]

Some crossfire with (I guess) AIB high clocked.
Looks a bit lower than GTX 1080 at Firestrike Ultra (I didn't compare the others).

BTW I read (I won't take any word from this) that the card struggles to go anything over 100Mhz OC with default voltage.

 

[FLAguy954]

Some crossfire with (I guess) AIB high clocked.
Looks a bit lower than GTX 1080 at Firestrike Ultra (I didn't compare the others).



BTW I read (I won't take any word from this) that the card struggles to go anything over 100Mhz OC with default voltage.

If those temps are true, I'm waiting for 3rd party coolers for sure.

 

[JohnnyFootball]

The 480 looks like the card I have been waiting for.

Although, I am skeptical at some of the performance claims thus far and fear we may be setting ourselves up for disappointment on the 29th.