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

[ManofGod]

If the 1080 was the best or fastest selling card for Nvidia, how are prices too high? There was and will always be premium for the best performing card, that was true even back in the day. The 8800 ULTRA was $830+, the 8800 GTX was $600+.

Which is stupid and expensive let alone the fact that you lose very quickly when the next "newest" thing comes out. However, as long as people are willing to pay the high, high prices, they will continue charging them. The last time the high end prices were good were the 290/290x and the 970/980. Now, people consider over $600 good? Oh well, to each their own but I stand by my opinion.

 

[Deleted whining member 223597]

You were never the target for high end cards, neither am I. I don't see the point in buying the high end because of exactly what you said, there is always something new around the corner. If anything, I would get the GTX 1070 for ~$400 and get amazing performance at really low power draw. The extra $200 isn't worth it for me.

 

[mesyn191]

If the 1080 was the best or fastest selling card for Nvidia, how are prices too high?

Too high in the sense that your per dollar fps value isn't all that good. Just because that is normal for high end cards to be a poor value doesn't mean that its a good or desirable thing.

 

[ppi]

How fast Radeon RX 480 can be? | VideoCardz.com

Hmm looks like better cooling isn't the only thing that the ref cards need for overclocking, better to get non ref boards for all those people who want to overclock.

From that link, I find the memory overclocking results most interesting.

 

[Pieter3dnow]

Performance per watt and power consumption is pretty important, otherwise why do reviewers always look at it? I'm not saying that lower end cards need to be the most efficient, but it looks like this iteration of Polaris is being badly beaten in performance per watt by this iteration of Pascal. In a case where other factors like initial cost and overall performance are similar or close, be it budget or high end, why go with the thing that's delivering significantly less performance per watt?

It is a product that is at $199 $239 and there is no way around this. You come from a point of view that you would have to have things in order to enjoy RX 480 as a user which is rather silly.
You pretend that power and performance is more important then anything else which is your point of view.

Comparing a product that is in the high end with a low end product and using power as a deciding factor is not important as price. If it was important where are the other cards which use less power and have more performance ?

 

[Rustynuts]

You were never the target for high end cards, neither am I. I don't see the point in buying the high end because of exactly what you said, there is always something new around the corner. If anything, I would get the GTX 1070 for ~$400 and get amazing performance at really low power draw. The extra $200 isn't worth it for me.

Yet not that long ago I had trouble justifying even $300+ for a 970. Now we've lept up in price 33% for a 1070.

 

[Sith'ari]

This card got WAY OVERHYPED. That's the problem. It's a $200 part that's nothing more or less than a $200 part for 2016.

I think i'll have to mention this many times:
It's NOT a 200$ part, it's a 240$ part (all the reviews used the 240$ 8GB version instead of the highly AMD-advertised 199$ 4GB version).
And for those 240$ you get a card with no backplate, fan that is extremely loud ( AMD Radeon RX 480 8 GB Review ), and a possibly dangerous power output ( AMD Radeon RX 480 8GB Power Consumption Results ).
So in order for the aftermarket cards to fix all these issues, i think it's safe to assume that they will cost >240$, quite possibly near 280$ ?

P.S. My personall opinion is that i'm nowhere near impressed with this card. It's like celebrating if Nvidia's 1080s was the same performer as the 980 (* not even 980Ti) but for a lower price and for the same time AMD's RX 480s were outperforming the FuryXs more than 50%!! I'm sorry but in a scenario such as this i wouldn't celebrate at all !!

 

[schizo]

There's no compelling reason to want >4GB of RAM on a card with this level of performance. The 8GB model has its memory clocked higher, so that may matter; we don't know yet.

As for the rest, backplates are cosmetic fluff, the fan is only loud when you overclock (and you should buy an AIB card for that), and the power consumption stuff isn't yet confirmed as a real problem.

 

[razor1]

It is a product that is at $199 $239 and there is no way around this. You come from a point of view that you would have to have things in order to enjoy RX 480 as a user which is rather silly.
You pretend that power and performance is more important then anything else which is your point of view.

Comparing a product that is in the high end with a low end product and using power as a deciding factor is not important as price. If it was important where are the other cards which use less power and have more performance ?

What happens with the 1060 comes out with pretty much the same performance and has less power consumption, (better perf/watt)?

Will you say the same thing?

What did nV say, perf/watt is the same thing as performance. It actually is, because the more you can do with less, the more you have

If AMD just matched Maxwell 2 from last gen which is 1 and half nodes back (28nm vs 14nm), that doesn't give you a chill up your back what the 1060 can do, a smaller chip than the rx 480 and the same 1 and half node jump from Maxwell 2?

If everything is equal on the performance and price fronts, then other factors come into play, like perf/watt, like power consumption, like how hot the chip gets. If those other fronts are better, then price can be set higher, if they are lower then price will be set lower.

 

[heatlesssun]

Too high in the sense that your per dollar fps value isn't all that good. Just because that is normal for high end cards to be a poor value doesn't mean that its a good or desirable thing.

You don't buy high end components for value ratios, they're never good. You buy high end components for high end performance.

 

[heatlesssun]

If everything is equal on the performance and price fronts, then other factors come into play, like perf/watt, like power consumption, like how hot the chip gets. If those other fronts are better, then price can be set higher, if they are lower then price will be set lower.

Not really getting the people that think that power draw and performance per watt are meaningless. No, they aren't everything but when other things are equal these things become deciding factors. Plus just looking at the numbers so far Pascal is KILLING Polaris in power efficiency. A 480 at almost half the performance is drawing almost the same amount of power as a 1080. That's simply doesn't bode well for the 480 if the 1060 can retain that kind of power efficiency.

 

[razor1]

Not really getting the people that think that power draw and performance per watt are meaningless. No, they aren't everything but when other things are equal these things become deciding factors. Plus just looking at the numbers so far Pascal is KILLING Polaris in power efficiency. A 480 at almost half the performance is drawing almost the same amount of power as a 1080. That's simply doesn't bode well for the 480 if the 1060 can retain that kind of power efficiency.

Yeah they factor is but as a secondary nature. From an OEM point of view though power consumption is a big thing, it means costs of other components can be higher or lower based on the power consumption of the graphics card.

Pretty much what it tells me is AMD didn't have time to do any major architectural changes to Polaris to address power concerns from last gen, what they got was through FinFet and the lower leakage of Finfet. Its going to be a tough road for them if they don't start changing the way they think about designing their GPU's.

Actually the only two major changes AMD has done thus far from what have seen from synthetics is memory compression is a bit better and polygon throughput but both of those are still far behind nV's numbers.

 

[mesyn191]

You don't buy high end components for value ratios, they're never good. You buy high end components for high end performance.

Of course. But guy I was responding to was posting about the value of high end hardware and seemed confused about how it could possibly be considered bad.

Not really getting the people that think that power draw and performance per watt are meaningless.

They matter if there is a game changing difference but that doesn't seem to be the case here. Yes the RX480 uses more power than it really should for the performance it offers but total power is still quite low. Remember, most sites are showing total system power draw as around 250-260W while gaming with the RX480. When 300W PSU's are more than good enough its not really an issue.

 

[heatlesssun]

Of course. But guy I was responding to was posting about the value of high end hardware and seemed confused about how it could possibly be considered bad.

It could be considered bad that performance doesn't scale proportionally with cost. But that's how the market works hardware.

 

[mesyn191]

It could be considered bad that performance doesn't scale proportionally with cost. But that's how the market works hardware.

Ok but did I say otherwise? I mean what exactly are you disagreeing with in my original post under discussion here? I'd even mentioned in the post you quoted originally that its normal for high end cards to have poor value too.

I think you got the context of the discussion mixed up here or misread my post.

 

[Pieter3dnow]

What happens with the 1060 comes out with pretty much the same performance and has less power consumption, (better perf/watt)?
Will you say the same thing?

Nothing but it won't happen for under $200. You know it I know it and Nvidia knows it. Nvidia doesn't do low margins (ooh god how many times I have heard this being repeated when the console deal was discussed).
Let's hope they can fix their DX12 performance by then that would be such an embarrassment.

 

[razor1]

Nothing but it won't happen for under $200. You know it I know it and Nvidia knows it. Nvidia doesn't do low margins (ooh god how many time I have heard this being repeated when the console deal was discussed).
Let's hope they can fix their DX12 performance by then that would be such an embarrassment.

2 games that are both AMD sponsored is not something I would write home about it nor draw conclusions on. Not to mention Async done right on Hitman when first released really showed it wasn't done right, and the entire Oxide fiasco of jumping on the worthless async train.

Yeah they do love their margins and they can still keep their margins if they price their chips the same as AMD's rx480, since the 1060 GPU is smaller , also the power consumption being lower they can use lower cost components.

 

[Presbytier]

Nothing but it won't happen for under $200. You know it I know it and Nvidia knows it. Nvidia doesn't do low margins (ooh god how many times I have heard this being repeated when the console deal was discussed).
Let's hope they can fix their DX12 performance by then that would be such an embarrassment.

If their dx12 is so bad how come they outperform every amd card in dx12. If you try to say the gains aren't as good compared to amd I'd simply reply of course not, but when amd dx11 performance is so bad of course their gains are going to be larger.

 

[jbc029]

My personall opinion is that i'm nowhere near impressed with this card. It's like celebrating if Nvidia's 1080s was the same performer as the 980 (* not even 980Ti) but for a lower price and for the same time AMD's RX 480s were outperforming the FuryXs more than 50%!! I'm sorry but in a scenario such as this i wouldn't celebrate at all !!

To put that in proper context, if roles were switched and NV was offering what was a 1080 that gave 980 performance for $300 (same performance, 33% lower price) and AMD producing a 480 that was 50% faster than a Fury X but cost $800+ (price of product replaced + 33%) would it still be celebrated? Probably so.

 

[Pyroja]

Where are people getting that this is a loud card? I've watched like 8 different video reviews, and every one of them concluded the card was about as quiet as you could expect from a blower. I've heard no reviewer come out and say it's exceptionally loud... But oh man, every negative forum post likes to say it.

Heck, the [H] review itself says it's quiet under load: "Even during long sessions of gaming there is not an audible increase in fan noise or air moving while gaming with the RX 480."

 

[schizo]

The fan is purportedly extremely loud, approaching the infamous 290x, at 100%. Much louder than the 1070's blower at 100%. Also people saying it's a "grindy" noise.

 

[pcgeekesq]

Nothing but it won't happen for under $200. You know it I know it and Nvidia knows it. Nvidia doesn't do low margins (ooh god how many times I have heard this being repeated when the console deal was discussed).

Nvidia seems like Intel (AMD's other dominating competitor) in that regard. If your project at Intel couldn't show a path to 50% or better gross margins, it was doomed unless it was strategic and tied to an x86 CPU (like a chipset).

Note that AMD's company-wide gross margins are 27% - way to low for a high technology company IMHO. Intel's and NVidia's company-wide gross margins are over 55%.

 

[gordon152]

Which is stupid and expensive let alone the fact that you lose very quickly when the next "newest" thing comes out. However, as long as people are willing to pay the high, high prices, they will continue charging them. The last time the high end prices were good were the 290/290x and the 970/980. Now, people consider over $600 good? Oh well, to each their own but I stand by my opinion.

That's been my main concern with the thinking of some people and it borderlines brainwashing over the years with incremental price increases. You used to be able to still get a really good performance around the $200 to $250 price range relative to the high end devices, but that changed over time. People lament a RX480 only being faster than a 970, but how comparatively faster is a 380 or 960 from a 7950 from 2011? I know there is inflation, but the general buying power of the general populace has only gone down since the days of the 9700 Pro and $299 being the top end card.

Even this card really doesn't alleviate that, but its been the biggest jump in the $200 space in a considerably amount of time.

 

[razor1]

To put that in proper context, if roles were switched and NV was offering what was a 1080 that gave 980 performance for $300 (same performance, 33% lower price) and AMD producing a 480 that was 50% faster than a Fury X but cost $800+ (price of product replaced + 33%) would it still be celebrated? Probably so.

well if the 1080 at 980 performance had considerably less power usage like 100 watts yes it would be good, that is probably what is going to happen with the 1060, although I think it will be higher more like 100 to 110 watts.

Compared to AMD's and nV's last gen midrange the rx480 is a breath of fresh air, but compared to last gen nV performance cards, not so good, and since prices have dropped on those last gen performance products from their launch prices.....

 

[razor1]

That's been my main concern with the thinking of some people and it borderlines brainwashing over the years with incremental price increases. You used to be able to still get a really good performance around the $200 to $250 price range relative to the high end devices, but that changed over time. People lament a RX480 only being faster than a 970, but how comparatively faster is a 380 or 960 from a 7950 from 2011? I know there is inflation, but the general buying power of the general populace has only gone down since the days of the 9700 Pro and $299 being the top end card.

Even this card really doesn't alleviate that, but its been the biggest jump in the $200 space in a considerably amount of time.

yeah buying power has gone down to some degree but that is because the process node transition and R&D costs have gone up so there is no way to get the same performance benefits at the same costs as before if these companies are trying to maintain margins 45% and higher its even harder they have to give enough bang for the buck for the general consumer to purchase a higher end product.

 

[Pieter3dnow]

If their dx12 is so bad how come they outperform every amd card in dx12. If you try to say the gains aren't as good compared to amd I'd simply reply of course not, but when amd dx11 performance is so bad of course their gains are going to be larger.

So show me the GTX 1060 DX12 benchmarks and well talk ....

 

[Michaelius]

The Good - AMD no longer expects people to buy mid range 300W gpus, price in America (because in Poland they are more expensive than gtx 970)
The Bad - near zero overclocking potential of reference gpu, performance in DX11
The Ugly - throttling at stock speeds, power consumption above norms, zero improvement to driver overhead in dx11 now even more painful than before since $200 gpu will be matched more often with low cpu like i3 or whole FX family.

 

[Pieter3dnow]

2 games that are both AMD sponsored is not something I would write home about it nor draw conclusions on. Not to mention Async done right on Hitman when first released really showed it wasn't done right, and the entire Oxide fiasco of jumping on the worthless async train.

Yeah they do love their margins and they can still keep their margins if they price their chips the same as AMD's rx480, since the 1060 GPU is smaller , also the power consumption being lower they can use lower cost components.

You somewhat miss the point needless to say I'll remind you of that when the GTX 1060 gets benched...

 

[Presbytier]

So show me the GTX 1060 DX12 benchmarks and well talk ....

It doesn't exist and you know it, so pretending like it will massive perform worse in DX12 is bogus there is little evidence that Paschal architecture performs poorly in DX12 outside of games that are badly optimized to begin with, and then everyone's cards perform off.

 

[Pieter3dnow]

It doesn't exist and you know it, so pretending like it will massive perform worse in DX12 is bogus there is little evidence that Paschal architecture performs poorly in DX12 outside of games that are badly optimized to begin wit.

No I will state that AMD will do better.
And that argument might as well flow against Nvidia so all the games that are DX11 are not optimized for AMD ?

 

[razor1]

So show me the GTX 1060 DX12 benchmarks and well talk ....

Look AOTS benchmarks are all over the board, reviewers that use Perfmon and test in game results, nV is where its supposed to be,

Hitman someplaces nV does better, some places AMD does better in game, so its dependent on where the test is done

Then we have tomb raider, which yeah thats obvious.

Then you have QB, its just broken, what else is there right now for DX12?

You somewhat miss the point needless to say I'll remind you of that when the GTX 1060 gets benched...

You think nV is just sitting on its laurels and going to watch the rx480 do what it does? You think nV is like AMD and willing to concede an entire segment of the market to AMD?

Common sense after seeing what the rx480 does, unless nV drops the ball big time on itself, like the golden state warriors did being up 3-1 and losing the NBA championship after winning 73 games in the regular season, yeah I don't see the 1060 having any problems mopping the floor with the rx480 in performance or power usage, or perf/watt.

 

[Presbytier]

No I will state that AMD will do better.
And that argument might as well flow against Nvidia so all the games that are DX11 are not optimized for AMD ?

Except they don't do better as of now, and AMD issue in DX11 are historically noted. Paschal does not have an issue in DX12 the performance is either better or on par with DX11. You can not say the same with AMD, and even in the benches where AMD sees performance gain in DX12 it's still considerably lower than Nvidia's cards performance. It's as if you are wearing blinders.

 

[razor1]

You somewhat miss the point needless to say I'll remind you of that when the GTX 1060 gets benched...

We will see in a couple weeks time maybe. But to think the 1060 isn't going to be competitive with the rx480 is not even in question at this point, AMD made it easy for nV to answer back.

 

[pcgeekesq]

We will see in a couple weeks time maybe.

I suspect we'll see on July 7. Kyle's been making noises that are consistent with [H]ardOCP already having a GTX 1060 in the lab for testing already (e.g. "not a phantom product"). And NVidia may allow GTX 1060 reviews to be posted a week ahead of availability, as they did for the GTX 1080, just to stick the knife into AMD and drink their milkshake that much sooner.

 

[Algrim]

It doesn't exist and you know it, so pretending like it will massive perform worse in DX12 is bogus there is little evidence that Paschal architecture performs poorly in DX12 outside of games that are badly optimized to begin with, and then everyone's cards perform off.

I think it does exist. Kyle possibly very subtly hinted that they have one, if I read the shades of the fonts a certain way on one of his posts).

EDIT: Damn it! Beaten to the punch by half a minute.

 

[Presbytier]

I think it does exist. Kyle possibly very subtly hinted that they have one, if I read the shades of the fonts a certain way on one of his posts).

EDIT: Damn it! Beaten to the punch by half a minute.

I meant the benchmarks don't exist yet, at least not in a public manner.

 

[Algrim]

Gotcha...

 

[raghu78]

Good - DX12 performance and perf/$.
Bad - Power consumption and perf/watt.
Ugly - crappy reference cooler, loud noise even at stock clocks, PCI-E power overdraw, significant variance among even review samples with few unable to hit advertised 1266 Mhz boost clock in games, power usage rocketing with overclocks.

Rx 480 is basically running out of its optimum clock range and the main reason is the GF 14LPP implementation. the GF 14LPP process sucks and has significant variation among even review samples. If Vega is built at 14LPP then AMD is dead.

 

[Rizen]

The fan is purportedly extremely loud, approaching the infamous 290x, at 100%. Much louder than the 1070's blower at 100%. Also people saying it's a "grindy" noise.

I don't understand why this is a problem. Who runs their fans @ 100%? I have a 980 Ti with a very quiet EVGA ACX 2.0 cooler and I still max it out at 70%. And I game with headphones.

100% might make sense for benchmarking but you're talking about a significant noise increase on any cooler for what is typical a very marginal performance increase.

 

[Rizen]

I think regardless of cooler, anyone who wants to overclock is going to wait for the AIB custom cards, as a single 6 pin is clearly not enough power for meaningful overclocks. That's probably what AMD banked on as well.