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

[poofyhairguy]

Curious as to why you want to do crossfire? Are you running 1440p or higher?

Three reasons really:

1. (main reason) I will mine Ethereum in the downtime so I want more power than a single 480 will give me. Might as well use the slot.

2. (vain reason) I have been kinda sad that my 390X at its best can't match some of the benchmark numbers that my OCed 970 could do. I would love to finally feel like I had a significant upgrade over that damn 970 without waiting for Vega.

3. (vain and dumb reason) Quite frankly I could get by on a 280X for gaming because I only play older games (Steam sales) at 1080p. I have just bought into this whole "video cards as a sports team" mentality and in a sad way I will feel better about my tech decisions overall when the 480 rebrand is competing with the 1160 in a few years.

 

[Stuka87]

Oh wow, half? That is crazy.

Thank you for the answers.

Half is not correct. If each card uses around 110W (rumored), that would be 220W with both. Your 390X is a 250W card. So less, but certainly not half.

 

[poofyhairguy]

Half is not correct. If each card uses around 110W (rumored), that would be 220W with both. Your 390X is a 250W card. So less, but certainly not half.

Yeah but that means I could even get by with a 500w PSU right? That would be great, the 500w PSU I have is better than the 600w one.

 

[96Firebird]

Yeah, a decent 500W PSU should be able to power 2x RX 480s.

Edit - Assuming you have 2x 6-pin PCI-E connectors, which any decent 500W PSU should have...

 

[poofyhairguy]

Yeah, a decent 500W PSU should be able to power 2x RX 480s.

Ok then should I avoid non-reference cards that need more than a 6 pin?

Won't that hold back OCing and single card performance (ie how I expect I will play a lot of the time)?

I can put in the 600w if it means I get 20% more on OCing or something.

 

[sze5003]

Three reasons really:

1. (main reason) I will mine Ethereum in the downtime so I want more power than a single 480 will give me. Might as well use the slot.

2. (vain reason) I have been kinda sad that my 390X at its best can't match some of the benchmark numbers that my OCed 970 could do. I would love to finally feel like I had a significant upgrade over that damn 970 without waiting for Vega.

3. (vain and dumb reason) Quite frankly I could get by on a 280X for gaming because I only play older games (Steam sales) at 1080p. I have just bought into this whole "video cards as a sports team" mentality and in a sad way I will feel better about my tech decisions overall when the 480 rebrand is competing with the 1160 in a few years.

Good reasons. I also want to feel like I will get an upgrade. I've never done cross fire so I was thinking about it but I want less hassle for myself when I go to get rid of them once Vega comes out. Currently I'm just giving my card to my little brother. I haven't upgraded in so long and this time is a weird one.

 

[poofyhairguy]

Good reasons. I also want to feel like I will get an upgrade. I've never done cross fire so I was thinking about it but I want less hassle for myself when I go to get rid of them once Vega comes out. Currently I'm just giving my card to my little brother. I haven't upgraded in so long and this time is a weird one.

I feel you, I was going to just temporarily get a 480 and flip for a vega but I am kinda sick of the waiting game and Vega+1080p sounds like a waste.

Once Ethereum mining bottoms out I will have a pile of old GPUs to sell so maybe at that point I sell everything and go Vega, who knows.

 

[PhonakV30]

#TBT @anandtech Reviews the HD 4850 / 4870 in 2008. Something seems very familiar in 2016.. http://www.anandtech.com/show/2556/2?sf29484048=1

https://twitter.com/AMDRadeon/status/745979045419094016

 

[sze5003]

I feel you, I was going to just temporarily get a 480 and flip for a vega but I am kinda sick of the waiting game and Vega+1080p sounds like a waste.

Once Ethereum mining bottoms out I will have a pile of old GPUs to sell so maybe at that point I sell everything and go Vega, who knows.

Yea that's right about 1080 and Vega. I'm telling myself as long as the 480 is at least 50% better than my card, I'll get it for 1080p and by the time the next cards are out I should have a 1440 monitor. That way I'm not spending it all at once on one item.

 

[MTDEW]

AMD Radeon RX 480 3DMark Results (R9 390/X - GTX 980 etc)
Be warned the music may be offensive to some, so be prepared to mute it if you know you fall in that category.

 

[antihelten]

I believe NCIX had them listed at $350CAD+, and I'm assuming prices listed on their website are before taxes?

I honestly have no idea if NCIX (or any other specific site) includes taxes in their listed prices or whether they are applied at checkout. Either way it doesn't really matter, if the $350 price on NCIX doesn't include taxes, then we are looking at a price roughly a 15% above MSRP, which again is nothing unusual for pre-launch listings.

 

[96Firebird]

Ok then should I avoid non-reference cards that need more than a 6 pin?

Won't that hold back OCing and single card performance (ie how I expect I will play a lot of the time)?

I can put in the 600w if it means I get 20% more on OCing or something.

No, the PSU wattage rating is just a number. You might have a good 500W PSU that is better than a crappy 600W PSU. Your 500W might even have 2x 8-pin, if the manufacturer is confident in the ability to supply that power. What are the model numbers of the PSUs in question?

 

[frozentundra123456]

AMD Radeon RX 480 3DMark Results (R9 390/X - GTX 980 etc)
Be warned the music may be offensive to some, so be prepared to mute it if you know you fall in that category.

Those results look pretty reasonable, in contrast to some of the speculation.

 

[Sweepr]

Guys, I had some free time this morning so I went ahead and did a comparison video. While our specs are not the same, I wanted to see the difference.

Overclocked GTX 970 1522/8000 vs. his RX 480 @ 1266/8000 ? in GTA V.

The settings are as close as I could get them. Since my card is limited to
4GB/3.5GB :\ I couldn't max out the Extended distance scaling. I'm not sure how much difference it makes?

Anyway, here is a video of the comparison.

https://youtu.be/T0b-lBEkfjg

Thank you! Impressive showing for OCed Geforce GTX 970.

Comparing the two benchmarks the 480 performed around ~82% of the r9 390 system in GTA 5. R9 390 system also has motion blur enabled and AMD CHS soft shadows, advanced graphics might have some differences also. I know AMD CHS has around a 10% performance hit and motion blur is practically free. So the 480 is probably around ~75% of the 390 in GTA 5 on the same settings. Systems difference will also cause some differences, not sure how much.

RX 480 8GB
i5-6400
2x4GB 2133MHz DDR4
1TB 7200RPM HDD

vs

i5-4690
msi r9 390 gaming 8g
asus h81-gamer
kingston 8gb(2x4gb) 1600mhz cl9 ddr3
samsung 850 evo
corsair cx750

Could be drivers, but it would be very disappointing if it can't match Hawaii in DX11 games. 32 ROPs (1/2 Hawaii) got me a bit worried.

 

[boozzer]

by seeing the videos. I do not dream like AMD and their fanboys did with Fury X OC.

what videos? what did those videos show you?

and please don't use the fanboi label in this thread. lets keep it civil. you can easily be accused of it just by using it.

 

[poofyhairguy]

No, the PSU wattage rating is just a number. You might have a good 500W PSU that is better than a crappy 600W PSU. Your 500W might even have 2x 8-pin, if the manufacturer is confident in the ability to supply that power. What are the model numbers of the PSUs in question?

I don't remember the exact model numbers. One is a EVGA 600w Plus supply, the other is a 500w EVGA gold supply (500B I want to say).

 

[Thala]

Could be drivers, but it will be very disappointing if it can't match Hawaii in DX11 games. 32 ROPs (1/2 Hawaii) got me a bit worried.

ROPs essentially affect fill-rate. Doubt modern games are fill-rate limited instead of shader limited. For simple games with lots of overdraw...possibly.

 

[Shivansps]

Mmm he gets 60-70 with hairworks off, i Just tested a gtx970 everything max with hairworks off and it gets 50-60, im petty sure the 970 can get better fps than that if i matched the detail level.

 

[Shivansps]

If this end up to be true ill bet its memory the problem. We know from carrizo that color compresión cant do magic.

 

[digitaldurandal]

Curious as to why you want to do crossfire? Are you running 1440p or higher?

Don't suggest crossfiring. Especially not with Vega six months away.

Get one 480 if you need it now. Put 230 dollars in your pocket. Vega releases, sell the 480 and take the 230 dollars out of your pocket and buy a single card in the 400 range (hopefully.)

 

[sze5003]

Don't suggest crossfiring. Especially not with Vega six months away.

Get one 480 if you need it now. Put 230 dollars in your pocket. Vega releases, sell the 480 and take the 230 dollars out of your pocket and buy a single card in the 400 range (hopefully.)

Yea that's my plan. Although I don't know if Vega will cost only $400. Shouldn't be an issue as I'll be keeping it like I've kept my 7970 for a long time.

 

[DownTheSky]

If it can't beat or match 390x/980 then.. what's the point? The closer it comes to the lauch, the more leaks, the worse it looks.

 

[Mercennarius]

If it can't beat or match 390x/980 then.. what's the point? The closer it comes to the lauch, the more leaks, the worse it looks.

For a lot of people...it's the fact that for what a single 390X cost new you can buy TWO 480s and demolish a 390X and 980 in performance. And not only that...probably still use less power consumption than a single 390X.

 

[crisium]

32 ROPs is the new rumour?

ROPs and bandwidth are very important. You cannot downplay this unless AMD has made very, very large strides over Tonga.

Let me demonstrate how to cripple a card with ROPs and Bandwidth.

Tonga claims to have improved ROPs over earlier GCN as well as memory compression. Despite these GCN 1.2 improvements, the older GCN 1.1 Hawaii seems better than it is relative to the 380X precisely because of ROPs and Bandwidth.

ASUS 380X 1030MHz 32 CU's
290 reference 947MHz 40 CU's

This is only a 15% TFLOP advantage for the 290 over the factory OC 380X.

Yet the 290 is 27% faster:
https://tpucdn.com/reviews/ASUS/R9_380X_Strix/images/perfrel_2560_1440.png

Twice the ROPs and 75% more bandwidth matters. A lot. It really, really matters. This is clear demonstration that it allows those Compute Units the room to work without bottlenecks. Remember, Tonga already has improvements over Hawaii.

So, I post this to ponder if the improvements with Polaris 10 are enough. It has more bandwidth now, and the ROPs are higher clocked, but could these limitations be holding back the potential of AMD's new arch?

 

[antihelten]

32 ROPs is the new rumour?

ROPs and bandwidth are very important. You cannot downplay this unless AMD has made very, very large strides over Tonga.

Let me demonstrate how to cripple a card with ROPs and Bandwidth.

Tonga claims to have improved ROPs over earlier GCN as well as memory compression. Despite these GCN 1.2 improvements, the older GCN 1.1 Hawaii seems better than it is relative to the 380X precisely because of ROPs and Bandwidth.

ASUS 380X 1030MHz 32 CU's
290 reference 947MHz 40 CU's

This is only a 15% TFLOP advantage for the 290 over the factory OC 380X.

Yet the 290 is 27% faster:
https://tpucdn.com/reviews/ASUS/R9_380X_Strix/images/perfrel_2560_1440.png

Twice the ROPs and 75% more bandwidth matters. A lot. It really, really matters. This is clear demonstration that it allows those Compute Units the room to work without bottlenecks. Remember, Tonga already has improvements over Hawaii.

So, I post this to ponder if the improvements with Polaris 10 are enough. It has more bandwidth now, and the ROPs are higher clocked, but could these limitations be holding back the potential of AMD's new arch?

There's really no way of knowing whether those results were due to ROPs or not, since the Tonga drivers were quite immature at the time (and arguably still are), and thus results tend to be a bit all over the place.

For instance in your graph the 290 is 55% faster than the 285, but today it's only 45% faster. Assuming the 380X had a similar boost in performance with the new drivers (TPU doesn't include it in their newest tests sadly), then the 290 would only be about 19% faster, not 27%.

19% faster is only 4% more than we would expect by simply looking at the TFLOP numbers, which is well within the margin of error.