• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

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

Page 162 - 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.
Let's not try to pussyfoot around what that witcher result is - the card is unfortunately throttling at 1266 @ 82c and that explains why overclocking on the reference is bad.

Have both companies seriously left the 300 USD price point without a good product?

Everyone and their mother has been waiting for this new node to buy the $300 gtx 970 / 390 equivalent and the fact is that everyone wants a card with a bit more juice than this one, but doesn't want to pay for the 1070.

I'll probably buy it anyway, but if I cab find a 1070 at 400 USD it's mighty tempting
 
http://www.overclock.net/t/1603915/coming-soon-unofficial-polaris-owners-thread/40#post_25298305
I do not have steam. 3D Mark GPU Score 12508

my temp looks good even running the 3D Mark Fire Strike ... hovers around 45-50
2816498
Click to expand...
So, around 980 level performance on old drivers (not release day drivers).
This test was done with a i3.
 
Last edited:
Armsdealer said:
Let's not try to pussyfoot around what that witcher result is - the card is unfortunately throttling at 1266 @ 82c and that explains why overclocking on the reference is bad.

Have both companies seriously left the 300 USD price point without a good product?

Everyone and their mother has been waiting for this new node to buy the $300 gtx 970 / 390 equivalent and the fact is that everyone wants a card with a bit more juice than this one, but doesn't want to pay for the 1070.

I'll probably buy it anyway, but if I cab find a 1070 at 400 USD it's mighty tempting
Click to expand...

Where do you see that happening? If its running @ 1266 it is not throttling, as that is the clock it should be running at. One of the other leaked videos shows it running witcher 3 @ 75c...
 
Bacon1 said:
Where do you see that happening? If its running @ 1266 it is not throttling, as that is the clock it should be running at. One of the other leaked videos shows it running witcher 3 @ 75c...
Click to expand...

https://www.youtube.com/watch?v=eUiaJXLoKnE

That's the one I'm looking at. Maybe it's a unique scenario and won't be the case in general.
 
Armsdealer said:
https://www.youtube.com/watch?v=eUiaJXLoKnE

That's the one I'm looking at. Maybe it's a unique scenario and won't be the case in general.
Click to expand...

Thats the video showing witcher running @ 75-78c not 82c

They also test Ashes in DX11 mode?

Furmark runs cooler than gaming?

None of that makes any sense

I can't wait for actual reviews tomorrow 🙂
 
Don't forget AMD drivers mature and improve over time. Unlike Nvidia's, which cripple instead. And since 1080p is still one of the main resolutions the 480 looks promising.
 
A lot of changes..If rx480 is slower than 390x then it is only because only 32rops.
390x-2816SP at 1050Mhz
rx480 2304SP at 1266Mhz

390x have 22% more SP
RX480 have 22% more clock
It should perform in worst case at 390x level with 64rops.

if its slower then we can blame only AMD for bad design with 32rops.
 
Last edited:
The BiST technology is very interesting to me. I did some work on this early in grad school but in RFICs. I hope we get some more info on this.
 
Cooler sucks? Probably. It doesn't look too good. I'm glad they didn't solder the heatsink to the VRM and Memory front plate. That makes it very easy to slap on a 3rd party cooler with the memory and VRM front plate intact. I have spare coolers around, so I'm not too worried. But, yeah, blower style coolers are the worst. Stay the heck away if you want to overclock.

AIB or slap a custom cooler on there.
 
Head1985 said:
Nope pixel fillrate is low.
Edit..Yep it looks like 380x is faster than 390.But that makes zero sense..
Click to expand...

You can't pick one spec in isolation and expect it to explain 100% of the performance scaling of a video card.

Thus the oft-repeated: wait for NDA to expire and real benchmarks/reviews 🙂
 
Hard to say if it's the ROPs or the lower memory bandwidth compared to the 390X. ROPs seem more likely since Polaris has memory compression whereas Hawaii didn't, although it remains to be seen exactly how much of the performance delta the memory compression accounts for.
 
So it looks like we get improved tessellation through Primitive Discard Accelerator?

Also, based on that slide, Polaris 10's memory compression is much improved. Theoretically equivalent to 345GB/s of uncompressed memory bandwidth. I don't think memory bandwidth will be problem.
 
rx480 have 256GB/s + 2 generations delta color compresion(first tonga and then polaris)
390x have 384GB/s with no delta color compresion.

rx480 should have same or better bandwidth(tonga memory compresion brings +20% avg bandwidth + polaris also have new compresion)
 
The ROP count should only be an issue as resolution scales up. At 1080P, it should be a non issue. At 1440P and up, yes, it will have an impact. But I highly doubt very many people looking at this card want it for 1440P.
 
Head1985 said:
rx480 have 256GB/s + 2 generations delta color compresion(first tonga and then polaris)
390x have 384GB/s with no delta color compresion.

rx480 should have same or better bandwidth(tonga memory compresion brings +20% avg bandwidth + polaris also have new compresion)
Click to expand...

Yep looks to be ~18-19% with Fury vs Hawaii and ~35-37% with Polaris, or almost the same gain with Polaris vs Fury that Fury had over Hawaii
 
Status
Not open for further replies.

TRENDING THREADS

Back
Top