RX 480 post mortem - the good, the bad, and the ugly.

In all reality, 8gb on this card means jack shit. It does not have the power to be effective at larger resolutions so memory is a mute point. It performs near a gimped memory GTX970 so that right there tells you the 8gb model is a pointless card. About the only big draw to this card would be DX12 support but still, price is too high IMO. $199 for 8gb model seems very reasonable and would be an excellent seller.
Memory is not consumed as much by resolution, as it is by textures. Rise of the Tomb Raider with highest textures setting takes some 6GB VRAM even on 1080p.

Latest games are taxing enough, that maxing all settings will not guarantee 60fps smooth gameplay even on 480/390/970 class of cards.
 
I still wonder how they show it beat 1080 in CF mode, but couldn't touch the regular GTX980, when 980SLI is just little below the 1080, Or maybe I just don't understand the CF or SLI.

Because the scaling with 2 RX 480 is fucking out of this world in Ashes of the Singularity.

Robert is talking about performance in API and mentions scaling of RX 480 :)

 
I think Ashes uses DX12's multi-adapter mode which is not "CrossFire" (or "SLi" for that matter).
Making that comparison has always been a false equivalency.
 
Every review I've seen, Linus Tech Tips, TOT ect. the 970(Stock) beats the 480. Can you provide me a review that shows me otherwise?

Most reviews I saw ([H], PCPer, TechReport, Anand, Ars, PCGamer average summary, Digital Foundry) show the following thread of 480 vs. 970:
- Pretty much even at DX11, though there are a few outliers both ways
- 480 is much faster in DX12 (Hitman, Total War: Warhammer, Gears of War)

Also, as you linked that video, since when is EVGA 970 SSC stock 970?!
 
Last edited:
I think Ashes uses DX12's multi-adapter mode which is not "CrossFire" (or "SLi" for that matter).
Making that comparison has always been a false equivalency.

That is so weird ? because the person asking the question was mentioning how they beat 1080 ? And the only benchmark available that did that was AotS.
He used the wrong terminology.
Or are you going to show me more benchmarks where 2 RX 480 beat a GTX 1080 ?

Clearly posted about scaling in Ashes not anywhere else .......
 
Hardware canucks did Quantum Break if future DX12 games are similar 970 is a horrible Idea. C'mon Deus Ex release already :mad:
 
Wow the 390 is actually 5-10% faster than the 480, according to PCGamer's avg.

1HYZKaB.png

IDK most reviews its up there neck and nect give a frame or two. We call write our own story from the review we pick lol.
 
  • Like
Reactions: Remon
like this
IDK most reviews its up there neck and nect give a frame or two. We call write our own story from the review we pick lol.

my RX-480 is slower than my R9-290. My XFX BE OC @ 1400mhz/2250mhz is 5-10% slower than my R9-290 @ 1250mhz/1625
 
Wow the 390 is actually 5-10% faster than the 480, according to PCGamer's avg.

1HYZKaB.png

I would freaking love to see a 290 on that chart :eek:

my RX-480 is slower than my R9-290. My XFX BE OC @ 1400mhz/2250mhz is 5-10% slower than my R9-290 @ 1250mhz/1625

Well played AMD, now all those people who bought a card three years ago have a reason to upgrade! Oh wait...
 
Wow the 390 is actually 5-10% faster than the 480, according to PCGamer's avg.

I was curious about those results, so I read the review. They used a stock 480. They didn't use a stock 390 or 970. They used an AIB model with out of box OC.

TPU's numbers, the ones that I cite, are reference only unless otherwise indicated. This has the benefit of showing how reference cards stack up. But, this also runs into the opposite problem. If you ONLY read TPU's performance summary and nothing else, you'd think that the 380 is absolutely faster than the 960, and the same for the 390 over the 970. The 390x even tops the 980 in their charts at specific resolutions. But AMD's cards from the R9 300 series didn't have high factory OCs from AIB cards, whereas NV cards did. Comparing like for like in those situations would even out or even flip the chart.

Bottom line is that between the results from TPU and elsewhere, the sotck 480 seems to slot above the stock 970, below the stock 980, but closer to the 970. It's on par with average factory OCs, but can't keep up with your faster (Zotac AMP, EVGA SSC, etc.) 970 models. I'm curious to see what kind of OC's we'll get with AIB cards.
 
IDK most reviews its up there neck and nect give a frame or two. We call write our own story from the review we pick lol.

Depends on what games they are running. Something like Farcry Primal that requires a wider memory width will run worse on the 480 than the 390.
 
Well played AMD, now all those people who bought a card three years ago have a reason to upgrade! Oh wait...

Well they can always buy a 1080 or a 1070. The Fury X is on sale for a like 430$ on newegg right now as well.

In all honestly I will likely swap my 290, in favor of the RX-480. It consumes half the power. Will have to wait for water blocks though, and lets someone else do the leg work on increasing the tdp limits of the card in bios.@1500mhz it will beat my 290 pretty easy.
 
You are right. Why compare similar performing cards ? Regardless of generation.

Just saying, if you're going to do it, don't complain when I call you out by doing the EXACT SAME THING to point out the flaw in your logic. You pointed out the flaw in my logic. What you failed, and still fail to realize is that I wasn't using my logic. I was using yours.

I have a different version of this that I use on people in person. It goes like this. "I'm going to repeat to you what you just said. Now, YOU tell ME how stupid I sound."
 
Just saying, if you're going to do it, don't complain when I call you out by doing the EXACT SAME THING to point out the flaw in your logic. You pointed out the flaw in my logic. What you failed, and still fail to realize is that I wasn't using my logic. I was using yours.

I have a different version of this that I use on people in person. It goes like this. "I'm going to repeat to you what you just said. Now, YOU tell ME how stupid I sound."
There is no logic on the internet. And no one wins. But cool story bro.
 
It might in power consumption too, :ROFLMAO:

LOL it wont be that bad lol. Some guy has it tested at 1425 all with a better cooler without even touch the voltage. It was doing 63c with aftermarket cooler. Ran firestrike in all modes and it was above the fury nano. Around 180 watts.
 
my RX-480 is slower than my R9-290. My XFX BE OC @ 1400mhz/2250mhz is 5-10% slower than my R9-290 @ 1250mhz/1625

that r290 must be like at 330+ watt at those speeds no? Damn that is actually a good overclock on that.
 
I want more cores/threads but I'm not sure I'm willing to give up IPC for it.
It looks like my next CPU will be one of the premium Kaby Lake chips ($300+). 6c/6t minimum.


Kaby Lake definitely will not be 6 cores unless you are referring to KabyLake-E (if that makes it to production), which is theoretically years off.

This years kabylake (14nm) and next year's cannonlake are both just evolutions of skylake on the 1151 socket.

The earliest 6 core mainstream cpus could come would be 2018. I really doubt they will.
 
Is this comparison done at the same clock speeds for each CPU? Seems a bit fishy but they (is that GTA?) could be using some instructions, I don't know.
Can't watch the video, only judging by the leading image where it doesn't say.

Anandtech and other reviewers found there to be a 25% IPC difference. In that video the 2500k and 6600k are at their stock (and very similar) clock speeds. The 6600k has a 6% base clock advantage and a 5% turbo advantage, and is showing a 20+% gain over the 2500k. That also doesn't take into account PCI 3 vs 2 for modern video cards.
 
I expected more from this card. I wasn't expecting anything crazy, but thought it would match 980 performance.

Also, doesn't it seem like Nvidia could embarrass them big time with the 1060? If it's priced right...

There was a ~40% perf difference between 970 and 960

The same difference from 1060 and 1070 would make it like +10% over the RX 480

Not to mention, Nvidia is probably making some healthy profit margins on the 1070/1080 pricing, so they can afford to get really aggressive with the 1060. We'll see what happens I guess.
 
I expected more from this card. I wasn't expecting anything crazy, but thought it would match 980 performance.

With all due respect, your expectations were a bit too optimistic. Generally, the next upgrade takes the existing price (in this case, $199/$229 for the R9 380) and bumps the performance to the next tier (R9 390). So, 970/390 was the most realistic expectation and anything else a bonus. I expected 970 but hoped for 980.

Also, doesn't it seem like Nvidia could embarrass them big time with the 1060? If it's priced right...

They could, or they could embarrass themselves. Too soon to tell. What I find hilarious (and you're not doing this), is that some users guarantee that the 1060 will be faster based on nothing more than the brand name. At least those speculating on 480 performance had leaks to go off of.

There was a ~40% perf difference between 970 and 960

With the GTX 960 as the base, the GTX 970 was 61% faster at 1080p. (SOURCE) That's quite a gap.

The same difference from 1060 and 1070 would make it like +10% over the RX 480

Using the same source, the 1070 is is 50% faster than the RX 480 at 1080p. So, if MY numbers (TPU's numbers, really) are correct, then if the GTX 1070 has a similar 61% gap over the 1060, then the 1060 will be slower than the RX 480. For what it's worth, the gap between the 960/970 was astronomical for a $200 and $300 card. So I don't expect that to occur again.

Not to mention, Nvidia is probably making some healthy profit margins on the 1070/1080 pricing, so they can afford to get really aggressive with the 1060. We'll see what happens I guess.

Depends on sales and volume. But yea, I think that NV can afford a price war more than AMD can. AMD had better hope that Vega is a success. If AMD ends up fighting for the low end again, they'll lose the price war and ultimately get pushed out, just like they did on the CPU side.
 
I expected more from this card. I wasn't expecting anything crazy, but thought it would match 980 performance.

Also, doesn't it seem like Nvidia could embarrass them big time with the 1060? If it's priced right...

There was a ~40% perf difference between 970 and 960

The same difference from 1060 and 1070 would make it like +10% over the RX 480

Not to mention, Nvidia is probably making some healthy profit margins on the 1070/1080 pricing, so they can afford to get really aggressive with the 1060. We'll see what happens I guess.

The 480 I think is headed for a flop. It simply is drawing too much power compared to Pascal.
 
Using the same source, the 1070 is is 50% faster than the RX 480 at 1080p. So, if MY numbers (TPU's numbers, really) are correct, then if the GTX 1070 has a similar 61% gap over the 1060, then the 1060 will be slower than the RX 480. For what it's worth, the gap between the 960/970 was astronomical for a $200 and $300 card. So I don't expect that to occur again.

I think writing is literally on the wall of how this will play out. The 1060 will be a little faster, a little more expensive and TON more power efficient than the 480.
 
The 480 I think is headed for a flop. It simply is drawing too much power compared to Pascal.

Historically, power draw doesn't matter. Power draw causing a flop is just what fans of the brand with the better efficiency always spout. Whenever AMD had the power draw advantage, they'd spout this same line, and yet NV still maintained >50% market share. The majority of users are simply going to look at their budget and get the fastest card within that budget. Well, actually, no, that would be the majority of enlightened users. The vast majority of users will succumb to marketing :)
 
I think writing is literally on the wall of how this will play out. The 1060 will be a little faster, a little more expensive and TON more power efficient than the 480.

Based on...? (I'm not saying that you're wrong, but it would be nice to base that on something of substance other than the brand name behind the product).
 
Historically, power draw doesn't matter. Power draw causing a flop is just what fans of the brand with the better efficiency always spout. Whenever AMD had the power draw advantage, they'd spout this same line, and yet NV still maintained >50% market share. The majority of users are simply going to look at their budget and get the fastest card within that budget. Well, actually, no, that would be the majority of enlightened users. The vast majority of users will succumb to marketing :)


The only time this is happened is with Fermi, and that is because it still was the fastest card by a large margin. Actually Fermi 480 launch nV lost marketshare for one quarter..... So yeah power matters.
 
Historically, power draw doesn't matter. Power draw causing a flop is just what fans of the brand with the better efficiency always spout. Whenever AMD had the power draw advantage, they'd spout this same line, and yet NV still maintained >50% market share. The majority of users are simply going to look at their budget and get the fastest card within that budget. Well, actually, no, that would be the majority of enlightened users. The vast majority of users will succumb to marketing :)

I think it's one thing to draw a lot of power at the high end versus the low end. A cheap card that draws a lot of power that's significantly slower than top end models just looks bad and hurts the value proposition of a low end card.
 
I think it's one thing to draw a lot of power at the high end versus the low end. A cheap card that draws a lot of power that's significantly slower than top end models just looks bad and hurts the value proposition of a low end card.

And yet, it hasn't stopped the faster card (or even the slower card sometimes) from outselling the more efficient competition. See - Fermi.

Power only becomes relevant when your favored brand is suddenly the most efficient. Because I don't have a preferred brand, I don't care about power draw. I don't care if the RX 480 draws 50W, 100W, 200W, or even 250W. My PSU can handle it. I care about the performance and the price.
 
Conclusion - This is AMD Maxwell. Take a GTX 970, drop the price $130, give it a separate 8GB option, and paint it red. Once we have AIB models with custom cooling and power delivery to fix AMD's engineering faux pas, this card will be highly recommended. Until then, wait patiently.

Yeah, pretty much.

but I think a lot of people would've ponied up $330 for a GTX 970 in september 2014 instead of $230 for a RX 480 in 2016, two years later if they knew that the medium end wouldnt move as much.
 
Based on...? (I'm not saying that you're wrong, but it would be nice to base that on something of substance other than the brand name behind the product).

There a are number of power benchmarks out there showing the 480 drawing more power than the 1070 and not a lot less than the 1080. That's a huge warning light going off.
 
And yet, it hasn't stopped the faster card (or even the slower card sometimes) from outselling the more efficient competition. See - Fermi.


Fermi lost marketshare for 1 quarter after the gtx 480 was launched lol, it wasn't till nV introduced the mid range and low end models did Fermi do better.
 
There a are number of power benchmarks out there showing the 480 drawing more power than the 1070 and not a lot less than the 1080. That's a huge warning light going off.

Again, you said that the GTX 1060 would be FASTER than the RX 480. Power draw on the GTX 1070 is not relevant.

That's like gauging how fast a Corvette will run a 1/4 mile based on the size of the gas tank. But anyway, this is pointless. Thank you for confirming that you have no evidence that the 1060, at this time, will be faster than the RX 480.
 
I don't care if the RX 480 draws 50W, 100W, 200W, or even 250W. My PSU can handle it. I care about the performance and the price.

And this totally kills the point of a performant budget card. That kind of product simply shouldn't be in the same power envelope as the fastest cards out right now.
 
Yeah, pretty much.

but I think a lot of people would've ponied up $330 for a GTX 970 in september 2014 instead of $230 for a RX 480 in 2016, two years later if they knew that the medium end wouldnt move as much.

And if they didn't know about dx12, sure. But it's not 2014.
 
The 480 I think is headed for a flop. It simply is drawing too much power compared to Pascal.

Here is how how it goes. People see $200 and $229 and they see anothe card at 299. Even if the review says oh its good performance for the money but it uses alot of power. At that point people forget about power trust me. Money is always first!
 
Again, you said that the GTX 1060 would be FASTER than the RX 480. Power draw on the GTX 1070 is not relevant.

That's like gauging how fast a Corvette will run a 1/4 mile based on the size of the gas tank. But anyway, this is pointless. Thank you for confirming that you have no evidence that the 1060, at this time, will be faster than the RX 480.

The 1060 will be price and performance competitive with the 480 while drawing substantially less power. I think that's all but certain.
 
Back
Top