AMD Polaris Thread: Radeon RX 480, RX 470 & RX 460 launching June 29th

Page 164 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Status
Not open for further replies.

Bacon1

Diamond Member
Feb 14, 2016
3,430
1,018
91
Micro center pricing (US):



Edit: 390x down to $299 with a nice XFX Warranty and DD cooler too.
 

Mopetar

Diamond Member
Jan 31, 2011
8,005
6,451
136
$40 for the extra 4 GB seems to be a bit much unless those cards are offering something else on top of that. I suppose that assumes that the 4 GB card goes for $200 and there isn't any price gouging there which remains to be seen.
 

crisium

Platinum Member
Aug 19, 2001
2,643
615
136
What? If it ties a 390X, at it's perf/w, that's efficiency gained.

More performance at the same power, or the same performance at less power is "efficiency" improvements.

Are you defining shader efficiency as wattage? Is that how AMD is defining it? Wouldn't it be "IPC"?
 

JDG1980

Golden Member
Jul 18, 2013
1,663
570
136
So, you know more than the chip designers and the limitations that they had to work with?
Nothing is free in silicon, and you have to manage what you got.

Be that as it may, the fact remains that Fiji in particular was shit. With 4096 shaders, it should have been over 45% stronger than Hawaii, but instead it was only about 17% better. That proves it was being bottlenecked by way too few ROPs. And Tahiti was bottlenecked, too (not surprising considering the shader to ROP proportion was the same as Fiji). R9 280X was less than 30% better than R9 270X even though the shader count indicates it should have been 60% better.
 

Saylick

Diamond Member
Sep 10, 2012
3,385
7,151
136
/snip
Shaders:

/snip

Hm, Polaris does not appear to use that shader patent that was discussed a month or two back. Looks like it's minor updates to the CUs, mostly improving the memory subsystems for extracting extra perf not unlike how CPU IPC has been slowly increasing through prefetching, larger reorder buffers, caches, etc.
 

JDG1980

Golden Member
Jul 18, 2013
1,663
570
136
There is no reason to increase the cost of the chip too add more ROPs. This is a 1080P card, 32 ROPs is the correct number for this resolution.

Polaris 10 is almost certainly going to be used for 5K resolution on the next Retina iMac.
 
Feb 19, 2009
10,457
10
76
Are you defining shader efficiency as wattage? Is that how AMD is defining it? Wouldn't it be "IPC"?

I don't know how AMD defines it, the slides leaked are missing all the footnotes.

But if you want to compare IPC, it's # of shaders at a clockspeed ratio comparison.

2304 SP @ 1.26ghz vs (390X) 2816 SP @ 1.05ghz.
 

crisium

Platinum Member
Aug 19, 2001
2,643
615
136
I don't know how AMD defines it, the slides leaked are missing all the footnotes.

But if you want to compare IPC, it's # of shaders at a clockspeed ratio comparison.

2304 SP @ 1.26ghz vs (390X) 2816 SP @ 1.05ghz.

Yup, and with only 1% disadvantage for 480, if it was improved it would make up for that and, ideally, then some.
 

Stuka87

Diamond Member
Dec 10, 2010
6,240
2,559
136
Polaris 10 is almost certainly going to be used for 5K resolution on the next Retina iMac.

Yes, it will. But the iMac is not a gaming machine. They never have been. A Tonga powers the current top end model. P10 should be a huge jump over the current models. But I do not think using the iMac as an example for video performance is realistic.
 

Stuka87

Diamond Member
Dec 10, 2010
6,240
2,559
136
Last edited:
Feb 19, 2009
10,457
10
76
Yup, and with only 1% disadvantage for 480, if it was improved it would make up for that and, ideally, then some.

Yes, if you measure perf per shader per clock, RX 480 should be onpar with a 390X. If there's architectural gains and no other bottlenecks, it should be faster. If it's not faster, either there's no uarch gains or there's bottlenecks. Your point is sound logic.
 

iiiankiii

Senior member
Apr 4, 2008
759
47
91

AstuteCobra

Junior Member
May 5, 2015
2
0
66
Hm, Polaris does not appear to use that shader patent that was discussed a month or two back. Looks like it's minor updates to the CUs, mostly improving the memory subsystems for extracting extra perf not unlike how CPU IPC has been slowly increasing through prefetching, larger reorder buffers, caches, etc.

The adaptive aging compensation might be apart of that patent.

 

Bacon1

Diamond Member
Feb 14, 2016
3,430
1,018
91
Amazon needs to stop letting idiots write "reviews".

For the MC ones, I am wondering if the ones that cost more have dongles/adapters in them than the ones that are lower cost?

Lol @ the reviews, even the 5 star are terrible, flagged.

Yeah the XFX one shown before did include an adapter, not sure if that was on the OC'd version or regular one, or if all regular boxes will include the adapter. They might since there is no DVI port and many monitors are DVI. I'm guessing they'll get a lot of bad reviews from people if they don't include an adapter because people won't realize they need it until after they've ordered.
 

RussianSensation

Elite Member
Sep 5, 2003
19,458
765
126
I wonder if AMD's ROPs are really ginormous, with tons of transistors, and this is why AMD is so reluctant to provide enough. We've seen this happen recurrently: Tahiti (despite being a flagship card at the time) had only 32 ROPs, which was badly inadequate, and Tonga didn't fix this. Fiji was also badly shorted on ROPs, with a mere 64 when it should have had 128.

Or by far the MOST obvious scenario: RX 480 is an R9 380 replacement, not an R9 390/390X replacement. But why then does RX 480 perform so similarly to the R9 390? That's because next generation's lower tiers start matching previous generation's mid-range tiers. Subsequently, next generation's mid-range tiers will come close to last generation's flagship tiers (similar to have GP104 competes and outperforms Maxwell's 980Ti). It's generally been this way for decades.

You were also completely wrong about AMD's "lack of architectural improvements". For AMD to pull off near R9 390 level of performance with a 256-bit bus card with only 32 ROPs is one of the biggest architectural improvements they've made in the last 5 years.

For those among you who still do not understand AMD's desktop discrete GPU product lines, AMD has 4 distinct desktop dGPU tiers.

4th tier = x60/x70 = low end
3rd tier = x80 = mainstream/performance
2nd tier = x90 = high end (although many of us would call the R9 390 $329 tier mid-range and R9 390X $429 upper mid-range)
1st tier = Fury (fancy name line) = enthusiast



It appears that after AMD changed the naming convention and introduced the Fury line, RX480 is AMD's 3rd tier product line.

This is how the current gen is shaping up:

Polaris 11/10 = RX 470/480 will compete against GTX1050/1060/1060Ti
Vega 10 = RX 490/490X (?) will compete against GTX 1070/1080
Vega 11 = Fury / Fury X successors should compete against Big Pascal

This means AMD should have 2 more tiers (i.e., 2 completely different segments) above the RX 480. This of course will mean $329->$499 tier with the rumoured Vega 10 (this will be RX 490/490X style cards, direct replacements for R9 390) and then there will be Fury/Fury X replacements with the rumoured Vega 11.

The confirmation of 32 ROPs is by far the biggest egg on the face of HardOCP (essentially Kyle's entire credibility is lost completely if he thinks AMD replaced $329/429 Hawaii with a $199-229 32 ROP videocard); and literally everyone who maintains that RX 480 is an R9 390/390X successor. RX 470/480 are R7 370/R9 380 successors just as many of us have been saying for months when we heard the rumours of RX 480 ~ R9 390/X.

As I said earlier, if RX 480 ~ R9 390/970 and come close to 980/390X with only 32 ROPs, the Vega 10 and 11 HBM2 parts with 64 (or possibly 96 ROPs) will be very fast.

Another hint is the RX 480's die size of only 232mm2. That leaves AMD room for a chip with a 300-375mm2 die size and another chip with a 425-500mm2 die size.
 
Last edited:

boozzer

Golden Member
Jan 12, 2012
1,549
18
81
Read the review. That sounds totally like someone from this very forum. No Joke. Man, it's like some kinda of marketing campaign AGAINST the RX480. It's surreal.
you mean the 3 star review? that was 100% copy pasta of the some posts in this very thread
 
Feb 19, 2009
10,457
10
76
As I said earlier, if RX 480 ~ R9 390/970 and come close to 980/390X with only 32 ROPs, the Vega 10 and 11 HBM2 parts with 64 (or possibly 96 ROPs) will be very fast.

All the recent leaks point towards 390X class performance. I'm going to buy it anyway but it would be nice if it can overclock well!
 
Status
Not open for further replies.
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/    |