Official AMD Polaris Review Thread: Radeon RX 480, RX 470, and RX 460

Page 4 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

gusfring22

Junior Member
Jun 27, 2016
5
0
0
No one has answered this question yet - how much will performance improve with drivers and game updates lets say a year from now compared to todays reviews?
 

boozzer

Golden Member
Jan 12, 2012
1,549
18
81
No one has answered this question yet - how much will performance improve with drivers and game updates lets say a year from now compared to todays reviews?
that is because no one knows, no one can be absolutely sure even if AMD has been superb in this area for the last 3 years.
 

Piroko

Senior member
Jan 10, 2013
905
79
91
Polaris dont have VP9 decode either in hardware.
Polaris has 4k VP9 decode in hardware. Source: PCGH
Beim Dekodieren schlägt Polaris seine Vorgänger ebenfalls deutlich: HEVC mit 4k60 im Main-10-Profil, was 10 Bit Farbtiefe pro Kanal beinhaltet, VP9 wird bis zur 4K-Auflösung entschlüsselt, MJPEG bis 4K30, H.264 bis 4K120 MP4-P2 sowie VC1 bis 1080p60. Da konnte bisher nur Fiji mithalten und dem ging die HDR-Fähigkeit für HEVC ab.

Also, the PCGH Polaris sample was measured at 3.3 Sone, their 1080 sample was quite a bit louder (4.2 Sone) but their 1070 sample was equally quieter (2.7 Sone).

Seems like the 480 sustains boost clocks in the 1150 MHz range while gaming, which explains the rather mediocre results. At the same time, overclocking is locked to 1.1V on the stock boards, the PCGH sample hit 1350 MHz which gave a solid boost in the 15 to 20% range.

I'd say it is a bit lower performance than I expected, however I also expected the card to be more expensive so it's still a very nice product to me.
 

boozzer

Golden Member
Jan 12, 2012
1,549
18
81
Polaris has 4k VP9 decode in hardware. Source: PCGH


Also, the PCGH Polaris sample was measured at 3.3 Sone, their 1080 sample was quite a bit louder (4.2 Sone) but their 1070 sample was equally quieter (2.7 Sone).

Seems like the 480 sustains boost clocks in the 1150 MHz range while gaming, which explains the rather mediocre results. At the same time, overclocking is locked to 1.1V on the stock boards, the PCGH sample hit 1350 MHz which gave a solid boost in the 15 to 20% range.

I'd say it is a bit lower performance than I expected, however I also expected the card to be more expensive so it's still a very nice product to me.
wow, so basically AIB boards can easily give 20% more performance then?
 

richaron

Golden Member
Mar 27, 2012
1,357
329
136
No one has answered this question yet - how much will performance improve with drivers and game updates lets say a year from now compared to todays reviews?

Well I've got no sources to back it up, I can remember what someone else worked out with a weird spreadsheet somewhere... :sneaky:

That is: from Kepler-Maxwell AMD card have been increasing 6-8% per year.

This seems about right to me, who has followed reviews over the years, but I'm not going to claim it's 100% true. Of course this is not counting dx12 games, which take advantage of AMD's apparently more forward looking architecture. It could easily be (and has already been seen) double digit performance % increase in these as well. But we really can only tell after the fact.
 

Aikouka

Lifer
Nov 27, 2001
30,383
912
126
I've only read Anandtech's preview for the card, and if you're only willing to spend $200 on a card, it seems like a decent catch. I think the biggest bummer is how some people seemed to hope that the 480 in Crossfire would beat the 1070. To be fair, I haven't seen any reviews with Crossfire numbers (if they even exist), but I'd assume scaling to be around 50-60% on average. If you can keep that scaling factor, you may have a competitive setup depending on the game, but you're spending around the same amount, using more power, and getting close to the same performance. In other words, there's just not much of a benefit.

Now, if you really like Hitman: Absolution, then apparently, the 480 is a winner.
 

crisium

Platinum Member
Aug 19, 2001
2,643
615
136
wow, so basically AIB boards can easily give 20% more performance then?

Who knows. Even if AIB gets half that improvement which is nice... AMD already made the mistake of ensuring these blower cards will be the face of 480 for years to come as they are reused in subsequent reviews and people google "Radeon 480 reviews" and bring up launch reviews regardless of how good AIB is.

What a mistake (again).
 

Piroko

Senior member
Jan 10, 2013
905
79
91
wow, so basically AIB boards can easily give 20% more performance then?
Presumably, yes. The specific clock ranges noted by PCGH:
Full HD : 1.190-1.220 MHz
WQHD : 1.120-1.170 MHz
Ultra HD : 1.050-1.100 MHz

These drops in clocks at higher resolution (due to higher shader load) also perfectly explain the lower scaling compared to Hawaii, there's no indication that this card is ROP bottlenecked.

Also worth noting: PCGH mentions that there's a bug in the current driver that stops Polaris from entering the lowest power mode. Expect idle power consumption to drop by ~30% with newer drivers.
 

MrTeal

Diamond Member
Dec 7, 2003
3,587
1,748
136
I've only read Anandtech's preview for the card, and if you're only willing to spend $200 on a card, it seems like a decent catch. I think the biggest bummer is how some people seemed to hope that the 480 in Crossfire would beat the 1070. To be fair, I haven't seen any reviews with Crossfire numbers (if they even exist), but I'd assume scaling to be around 50-60% on average. If you can keep that scaling factor, you may have a competitive setup depending on the game, but you're spending around the same amount, using more power, and getting close to the same performance. In other words, there's just not much of a benefit.

Now, if you really like Hitman: Absolution, then apparently, the 480 is a winner.

Actually, 480CF does beat 1070 when scaling works.



On average 480CF is about as fast as GTX1070, and just including games that scale with CF it's 20% or so faster. The downsides are that 480CF costs more than a single GTX1070, consumes more power than a GTX1070, and in those games where it doesn't scale you're going to have to turn down some settings if you're playing at higher resolutions.
 

gusfring22

Junior Member
Jun 27, 2016
5
0
0
Well I've got no sources to back it up, I can remember what someone else worked out with a weird spreadsheet somewhere... :sneaky:

That is: from Kepler-Maxwell AMD card have been increasing 6-8% per year.

This seems about right to me, who has followed reviews over the years, but I'm not going to claim it's 100% true. Of course this is not counting dx12 games, which take advantage of AMD's apparently more forward looking architecture. It could easily be (and has already been seen) double digit performance % increase in these as well. But we really can only tell after the fact.

Thank you very much. So it sounds like a buyer of a new card can be confident that the benchmarks/performance we see today is the "floor" and it will only improve for the next ~3 years of ownership
 

Azix

Golden Member
Apr 18, 2014
1,438
67
91
I've only read Anandtech's preview for the card, and if you're only willing to spend $200 on a card, it seems like a decent catch. I think the biggest bummer is how some people seemed to hope that the 480 in Crossfire would beat the 1070. To be fair, I haven't seen any reviews with Crossfire numbers (if they even exist), but I'd assume scaling to be around 50-60% on average. If you can keep that scaling factor, you may have a competitive setup depending on the game, but you're spending around the same amount, using more power, and getting close to the same performance. In other words, there's just not much of a benefit.

Now, if you really like Hitman: Absolution, then apparently, the 480 is a winner.

it really should beat it. But dx11 multiGPU is hit/miss. I haven't seen the crossfire benches yet but I doubt they are doing their testing with the better optimized multiGPU games. We all know it can be broken, so just test where it actually works.

edit: mrteal post the crossfire summary. Definitely as expected when its actually working. Crossfire that's not working is not worth testing/including. Pretty close to 1080 perf.
 

Cares

Senior member
Mar 8, 2005
868
0
76
My question is...should I return my $180 MSI R9 380 4GB for this 8GB card that will cost about $270 after taxes.? Is it worth the 50% price increase? Is it worth considering the 4GB version of this card at $215?
 

Elixer

Lifer
May 7, 2002
10,376
762
126
This was in my in-box this morning...
XFX Radeon RX 480 DirectX 12 RX-480M8BBA6 256-Bit GDDR5 PCI Express 3.0 Video Card
Over Clocked to 1328MHz, Backplate and DP to DVI Adapter Included. Limit 2 per Customer.
$269.99

As for the 480 itself, pretty much what I was expecting, though, I was hoping for a bit more oomph. It still is a good card for the target audience, I am just not in that target.
It is also too bad that AIBs won't have their superior cooling systems available day 1.
AMD is in a quiet period as well now, so all these live-streams can't really talk about much.
 
Last edited:

Thinker_145

Senior member
Apr 19, 2016
609
58
91
So disappointed. All this time I didn't buy the 390 just so I could save $60? Well I doubt it's even gonna be $60 since launch prices are always inflated where I am.
 

beginner99

Diamond Member
Jun 2, 2009
5,224
1,598
136
Someone please tell me Vega is not at Global foundry because this could get uglier. This is assuming 14nm finfet is crap and not the Polaris design itself.

Agree. if Vega is GF too, I don't see a bright future for AMD. It really looks like the process sucks pretty bad. Assuming 480 equals 390 in performance on average (in fact it's a bit slower) and using the power numbers on anandtech review (eg. Load power crysis - idle power), performance per watt increased by only 1.5x or 50% on the 390 which is setting the bar pretty low already and barley beats the GTX 970 on a node shrink...

Reference model running at it's limits. Looks like it will be memory bandwidth limited in some games as the 4 GB version with slower RAM clearly loses in some titles.

wow, so basically AIB boards can easily give 20% more performance then?

Probably, but they will need at least 1 8-pin power connector. They will also run cooler and make less noise but more expensive obviously and more power usage.
 

StrangerGuy

Diamond Member
May 9, 2004
8,443
124
106
I've only read Anandtech's preview for the card, and if you're only willing to spend $200 on a card, it seems like a decent catch. I think the biggest bummer is how some people seemed to hope that the 480 in Crossfire would beat the 1070. To be fair, I haven't seen any reviews with Crossfire numbers (if they even exist), but I'd assume scaling to be around 50-60% on average. If you can keep that scaling factor, you may have a competitive setup depending on the game, but you're spending around the same amount, using more power, and getting close to the same performance. In other words, there's just not much of a benefit.

Now, if you really like Hitman: Absolution, then apparently, the 480 is a winner.

The biggest problem with the RX480 as a package is the 1060 next month. IMO it's almost certain the 1060 would at least match a stock 970 and guaranteed better perf/W.
 

Elixer

Lifer
May 7, 2002
10,376
762
126
Has anyone found a review of the 4k@60fps testing for the encoder?
That is what I am most curious about.
 

xpea

Senior member
Feb 14, 2014
451
152
116
I'm speechless regarding Polaris efficiency. What's the problem ?




typical gaming exceeding 150W TDP !!!


and 39W to read a bluray ? Are you kidding me ?



wow much worst than the worst nightmare... Pascal has nothing to fear.
AMD want to take back market share in laptops ? good luck...
 

Madpacket

Platinum Member
Nov 15, 2005
2,068
326
126
Well I'm a little disappointed with the [lack of] overclocking capabilities but stock performance is where I expected it to be. Power consumption is great compared to Hawaii but should be better at 14nm given the competition. Great for anyone wanting to upgrade from a lower end card without breaking the bank (if you live in the US anyway).

My main concern is with the overclocking. I don't see how aftermarket cards will overcome the limitations of the silicon which looks to be the limiting factor.

Once Fury Nano sized versions are released I'll probably pick one up for large screen HTPC/1080P gaming duties. I figure the 4GB models should be sufficient for 1080P given the capabilities of the card so hopefully we'll be able to get better form factor versions and lower priced cards to choose from.

The 1060 will be an option but I'm more interested in where gaming is headed (Async/DX12) and Polaris seems better suited for longevity.
 
Last edited:

poofyhairguy

Lifer
Nov 20, 2005
14,612
318
126
Wow it's worse than my 390x on almost every benchmark I see at 1080p.

I don't know what to do now. Newegg has two models in stock but I almost feel like waiting for AIB versions to hopefully get an upgrade via OCing.
 
sale-70-410-exam    | Exam-200-125-pdf    | we-sale-70-410-exam    | hot-sale-70-410-exam    | Latest-exam-700-603-Dumps    | Dumps-98-363-exams-date    | Certs-200-125-date    | Dumps-300-075-exams-date    | hot-sale-book-C8010-726-book    | Hot-Sale-200-310-Exam    | Exam-Description-200-310-dumps?    | hot-sale-book-200-125-book    | Latest-Updated-300-209-Exam    | Dumps-210-260-exams-date    | Download-200-125-Exam-PDF    | Exam-Description-300-101-dumps    | Certs-300-101-date    | Hot-Sale-300-075-Exam    | Latest-exam-200-125-Dumps    | Exam-Description-200-125-dumps    | Latest-Updated-300-075-Exam    | hot-sale-book-210-260-book    | Dumps-200-901-exams-date    | Certs-200-901-date    | Latest-exam-1Z0-062-Dumps    | Hot-Sale-1Z0-062-Exam    | Certs-CSSLP-date    | 100%-Pass-70-383-Exams    | Latest-JN0-360-real-exam-questions    | 100%-Pass-4A0-100-Real-Exam-Questions    | Dumps-300-135-exams-date    | Passed-200-105-Tech-Exams    | Latest-Updated-200-310-Exam    | Download-300-070-Exam-PDF    | Hot-Sale-JN0-360-Exam    | 100%-Pass-JN0-360-Exams    | 100%-Pass-JN0-360-Real-Exam-Questions    | Dumps-JN0-360-exams-date    | Exam-Description-1Z0-876-dumps    | Latest-exam-1Z0-876-Dumps    | Dumps-HPE0-Y53-exams-date    | 2017-Latest-HPE0-Y53-Exam    | 100%-Pass-HPE0-Y53-Real-Exam-Questions    | Pass-4A0-100-Exam    | Latest-4A0-100-Questions    | Dumps-98-365-exams-date    | 2017-Latest-98-365-Exam    | 100%-Pass-VCS-254-Exams    | 2017-Latest-VCS-273-Exam    | Dumps-200-355-exams-date    | 2017-Latest-300-320-Exam    | Pass-300-101-Exam    | 100%-Pass-300-115-Exams    |
http://www.portvapes.co.uk/    | http://www.portvapes.co.uk/    |