AMD's next GPU uarch is called "Polaris"

[Mondozei]

So there's that

The 2.5 times is higher than the previous 2X energy efficiency claim. Let's hope that turns out to be true.

Polaris sounds damn cool, as well.
Certainly a hell of a lot better than "Tonga" or the like.

 

[Headfoot]

oh god the speculation on an off hand comment about the brightness of the Polaris star is literally cringe inducing. Pretty typical for WCCF tech.

Read like a high school english paper "literary analysis" stretching for everything to possibly fill the 5 page assignment requirement.

Maybe they should change the font and increase the size to 13 and see if the teacher notices

 

[ShintaiDK]

You have to give them that, 2 articles unready on an undocumented vague leak. Only took them 2 weeks to discover it.

 

[KaRLiToS]

Misread, sorry

 

[DooKey]

Can we outlaw wccftech "articles" from now on?

 

[myocardia]

Can we outlaw wccftech "articles" from now on?

Seconded, they make the other bad sites look excellent in comparison.

 

[mysticjbyrd]

Can we outlaw wccftech "articles" from now on?

Then your bs detection skills would dull.

 

[MrTeal]

Did you know Polaris Aa is 4.5 times more massive than the sun? And what island is located on a planet orbiting our sun? Fiji.

It seems likely at this point that that Polaris will be a large 2682mm^2 die, four and a half times larger than AMD's existing Fiji GPU. It's not known yet how they managed to get around the reticle limit, but we expect performance of this new part to be massive.

 

[frozentundra123456]

A new low even for WCCF Tech.

Now don't get your shorts in a bunch, AMD fans. Polaris could be a wonderful product, but that is probably the most cringe-worthy "technical" article I have ever read.

BTW, Polaris isnt even that bright. It is only the 46 brightest star, which is something I guess, but nothing compared to what that cringe inducing picture implies.

 

[frozentundra123456]

Did you know Polaris Aa is 4.5 times more massive than the sun? And what island is located on a planet orbiting our sun? Fiji.

It seems likely at this point that that Polaris will be a large 2682mm^2 die, four and a half times larger than AMD's existing Fiji GPU. It's not known yet how they managed to get around the reticle limit, but we expect performance of this new part to be massive.

Naw, it is going to go supernova and swallow up nVidia. Seriously, I am winding down my work day, just sitting here stunned that anybody except maybe the Onion would print such a ridiculous article.

 

[waffleironhead]

Can we outlaw wccftech "articles" from now on?

Thats why the source article is supposed to be in the title. So you can skip the thread if you so choose.

 

[RussianSensation]

So there's that

The 2.5 times is higher than the previous 2X energy efficiency claim. Let's hope that turns out to be true.

Breaking the laws of physics now, huh? :hmm:...and this is EXACTLY how BS rumors start on the Internet when literal translations are misconstrued by sites like WCCFtech. Next thing you know "AMD promised us 2.5X perf/watt but they only managed to provide 80%. AMD lies, again, here is the quote from Raja...."

------

The Casual Sky Observer's Guide: Stargazing with Binoculars and Small Telescopes - pages 28-29

"Star Brightness

A star's brightness is measured in magnitudes (mag). It's an old method, introduced by the Ancient Greeks. The stars were divided into six categories. The brightest stars belonged to the class of first magnitude, the faintest visible to the naked eye were called sixth magnitude stars. The Greeks didn't have telescopes in their era, so there were not aware of the fact that there were fainter stars than the naked eye can see. A first magnitude star is 2.5 times brighter than a second magnitude star. A second magnitude star is 2.5 times brighter than a 3rd magnitude star, and so on. Thus a first magnitude star is 2.5 x 2.5 x 2.5 times brighter than a 4th magnitude star."

^

"Research reported in Science suggests that Polaris is 2.5 times brighter today than when Ptolemy observed it, changing from third to second magnitude."
https://en.wikipedia.org/wiki/Polaris#cite_note-16

So for example, this star's magnitude could have changed 2.5 x 2.5 times instead of just 2.5 times. WCCFtech is correlating things to mean what they want them to mean. IMO, Raja is not stating anything about changing perf/watt efficiency from 2X to 2.5X. He is stating a literal observation about this star.

A new low even for WCCF Tech.

Polaris stars, counting 1, 2, 3 = HL3 confirmed.

2.5X brighter, 3 stars => Arctic Islands 2.5X faster with HBM3 confirmed.

It seems likely at this point that that Polaris will be a large 2682mm^2 die, four and a half times larger than AMD's existing Fiji GPU. It's not known yet how they managed to get around the reticle limit, but we expect performance of this new part to be massive.

Star's age: Age 7×10^7 years

7nm node confirmed!

 

[Arachnotronic]

Can we outlaw wccftech "articles" from now on?

WCCFTech's analysis needs work.

 

[RussianSensation]

WCCFTech's analysis needs work.

That's one beastly new rig you got! What are you doing with TXs and 64GB of DDR4? Video work? I would have thought MSI Lightning 980Tis max OC > TXs max overclocked, unless TXs are custom bios modded + on water.

 

[DrMrLordX]

Can we outlaw wccftech "articles" from now on?

They're alright so long as they are linking some other source, in which case you can usually skip their analysis (well, most of it) and go to the source. They occasionally turn up stuff that often goes unnoticed by others.

So, in the case of this article, you can ignore the nonsense regarding Twitter activity and go read HWBattle for yourself.

 

[csbin]

 

[DrMrLordX]

Yeah that stuff. Ignore that bit.

 

[Magee_MC]

So there's that

The 2.5 times is higher than the previous 2X energy efficiency claim. Let's hope that turns out to be true.

Polaris sounds damn cool, as well.
Certainly a hell of a lot better than "Tonga" or the like.

Polaris, also known as the North Star. The name goes well with Arctic Islands.

 

[Goatsecks]

'Tea light', will probably be a more appropriate name... :awe:


Boy am I going to get a paddlin' for this one.

 

[Borealis7]

a more credible slide here:
http://hexus.net/tech/news/graphics/89219-amd-radeon-400-series-based-upon-polaris-microarchitecture

I thought the R7/9-4XX (Arctic Islands) series was supposed to be based on GCN still...by the time of its' release, GCN would be between 5 and 6 years old so we are definitely due for a new architecture.

 

[monstercameron]

a more credible slide here:
http://hexus.net/tech/news/graphics/89219-amd-radeon-400-series-based-upon-polaris-microarchitecture

I thought the R7/9-4XX (Arctic Islands) series was supposed to be based on GCN still...by the time of its' release, GCN would be between 5 and 6 years old so we are definitely due for a new architecture.

why would amd need a new gpu uarch? gcn is a relatively futureproof uarch and imo only needs tweaking. the next step should be raytracing cards imo...

 

[AtenRa]

the next step should be raytracing cards imo...

Industry is going for higher resolution (4K, VR) instead of higher Image Quality. You can forget about Raytracing for the next 10-20 years.

 

[Tuna-Fish]

Industry is going for higher resolution (4K, VR) instead of higher Image Quality. You can forget about Raytracing for the next 10-20 years.

The one opportunity for RT is if foveated rendering becomes possible for VR -- RT is more efficient at it, and would allow very smoothly varying rendering unit size and speed.

 

[Tuna-Fish]

why would amd need a new gpu uarch? gcn is a relatively futureproof uarch and imo only needs tweaking.

It's not like they just trash their old designs and start over from scratch. Once they tweak GCN enough, it makes sense to give it a new name, just for the marketing boost.

Personal wishlist: more registers per work unit, dismantling the existing geometry/tesselation and doing it all distributed on the CUs (with necessary tweaks to make that fast).

 

[DrMrLordX]

a more credible slide here:
http://hexus.net/tech/news/graphics/89219-amd-radeon-400-series-based-upon-polaris-microarchitecture

I thought the R7/9-4XX (Arctic Islands) series was supposed to be based on GCN still...by the time of its' release, GCN would be between 5 and 6 years old so we are definitely due for a new architecture.

That's weird, and the first that I've heard of Arctic Islands being non-GCN. Unless Polaris is just GCN 2.0 .