Polaris not so bright [Hardocp]From ATI to AMD back to ATI? A Journey in Futility

[flash-gordon]

Wow, he really went clickbait beyond fud, wccftech and videocardz all at once...

 

[SimianR]

This post from the HardForum cracked me up:

MacLeod said: ↑
Not really what I was getting at. AMD has been having problems for the better part of a decade now. But it's only recently we're getting these editorials about how AMD cut you off but you ain't mad bro......AMD is crap and their mom is so fat that when the [censored] jump up in the air she gets stuck.....but I ain't mad.

He makes a good point

 

[oussama-tn]

If Polaris is a big fail then why it will be powering the next Ps and Xbox...

 

[AnandThenMan]

If Polaris is a big fail then why it will be powering the next Ps and Xbox...

Does not compute, Kyle said Polaris is a failure that's the end of it.

 

[oussama-tn]

This reminds me of last year Kitguru's article/review about AMD DDR4 memory.

 

[techne]

It seems that they have noticed this thread... ^_^

I said: they [H] have noticed this thread (Anand).

 

[MongGrel]

I said: they [H] have noticed this thread (Anand).

I think most people have gotten that all ready.

I posted a link myself after someone had posted one from Overclockers.

Why does it matter that much ?

 

[digitaldurandal]

Kyle B - "As of two weeks ago, I am dead on about Polaris. And I do not think AMD is going to fix the problem by seeding reviewers with cherry picked samples of the cards this weekend. I do not think they can get back on track that quickly with their clocks for production. Polaris was never meant to be a middle of the stack part, but you can spin it however your want."

I do not understand how anyone with any level of knowledge can really believe this about a chip of this size. I wonder if he will explain this at any time, since he really does not have a reason not to and has not I will continue to assume he has finally went full on ragemode.

AMD intended this chip at ~2/3 the size of the 1080 to be their high performer... what?

 

[tential]

Kyle B - "As of two weeks ago, I am dead on about Polaris. And I do not think AMD is going to fix the problem by seeding reviewers with cherry picked samples of the cards this weekend. I do not think they can get back on track that quickly with their clocks for production. Polaris was never meant to be a middle of the stack part, but you can spin it however your want."

I do not understand how anyone with any level of knowledge can really believe this about a chip of this size. I wonder if he will explain this at any time, since he really does not have a reason not to and has not I will continue to assume he has finally went full on ragemode.

AMD intended this chip at ~2/3 the size of the 1080 to be their high performer... what?

Sometimes I question if they really do have the knowledge since they aren't exhibiting it.....

But at least from his standpoint it REALLY does suck. It's like hard to get excited about a midrange product that doesn't push the performance bar.

I'm bored already with Polaris.

I just want Nvidia to get onto the next part of the show and show us the new Titan at this point.

 

[escrow4]

If AMD releases a 480X for $299 or less that is as fast as a 390X with HEVC and VP9 decoding support I can't see how that is bad. I also note that [H]'s forums are disgustingly ugly, red and black makes you want to claw your eyes out.

 

[techne]

I think most people have gotten that all ready.
I posted a link myself after someone had posted one from Overclockers.
Why does it matter that much ?

I never said that.

 

[Azix]

If AMD releases a 480X for $299 or less that is as fast as a 390X with HEVC and VP9 decoding support I can't see how that is bad. I also note that [H]'s forums are disgustingly ugly, red and black makes you want to claw your eyes out.

I think it almost needs to be faster than 390x. approaching fury non-x performance at $300 would be a good spot for a new generation. They should not go off their MSRP for these cards and think $300 for 390x is a great thing. 390 was adequate and cheaper than that already.

 

[airfathaaaaa]

its funny how everyone forgot about this
https://hardforum.com/threads/amd-polaris-gcn-4-0-macau-china-event.1900514/
literally 2 days ago

Honestly, there is likely not much to miss here, as we do not think AMD has any answer for GTX 1080, but the middle of the product stack will surely get revamped if GlobalFoundries and crank out the 14nm parts.

and then we are the bad ones calling out than tech child

 

[mohit9206]

Kyle's loyal forum followers praising him for having the guts to write such an article is hilarious. It's like a hive mind there all agreeing with whatever Kyle wrote without questioning anything. Such fierce loyalty is hard to achieve and i congratulate Kyle for that.

 

[HurleyBird]

I'd like to point out is that it's theoretically possible for Polaris to be efficient as an architecture, while at the same time having SKUs with lower performance and higher power consumption than the current Pascal products. All it takes is AMD (in their infinite wisdom) pushing the chips far past sane limits in an attempt to compete with dies that are substantially larger. Such a scenario wouldn't necessarily spell doom for this generation of AMD cards the same way an architectural deficiency might, although either way Kyle doesn't seem the kind of person to take any sort of nuanced view on anything.

That said, the part about RTG rebelling against the rest of AMD in an attempt to split the company and get swallowed up by Intel... on the face of it sounds like a conspiracy theory dreamt up in the mind of a twelve year old.

Time will tell, I suppose. But I'm more skeptical of the business side of the article than the GPU side.

 

[jpiniero]

I think it almost needs to be faster than 390x. approaching fury non-x performance at $300 would be a good spot for a new generation. They should not go off their MSRP for these cards and think $300 for 390x is a great thing. 390 was adequate and cheaper than that already.

Seems like AMD had wanted to sell Polaris 10 for $399 though. It would not shock me if they are only selling cutdown models since GloFo can't produce enough to even paper launch something like nVidia did.

 

[railven]

The bipolarism for [H]ardOCP here baffles me. Beside that, RTG going to Intel would be AWESOME!

I've been saying Intel should just buy those GPU patents and curb stop Nvidia. If I got to give someone $200 or more for their margins, I got no qualms giving it to Intel

On the rumor of Polaris, if anymore leaks come about, might be time to worry.

Polaris bursting into flames is probable to me, and the only reason I'd entertain the thought is: GloFo.

They've being hindering AMD for years it seems. If they botch this, AMD better GTFO out of there ASAP. How many more products can GloFo ruins for AMD before they end that abusive relationship. Woof.

 

[airfathaaaaa]

The bipolarism for [H]ardOCP here baffles me. Beside that, RTG going to Intel would be AWESOME!

I've been saying Intel should just buy those GPU patents and curb stop Nvidia. If I got to give someone $200 or more for their margins, I got no qualms giving it to Intel

On the rumor of Polaris, if anymore leaks come about, might be time to worry.

Polaris bursting into flames is probable to me, and the only reason I'd entertain the thought is: GloFo.

They've being hindering AMD for years it seems. If they botch this, AMD better GTFO out of there ASAP. How many more products can GloFo ruins for AMD before they end that abusive relationship. Woof.

getting quite a lot of billions from the saudis if glofo fails isnt really something amd wants second time..

 

[krumme]

Temper passion and balls. Kyle is always best and fun like that. I dont think its clickbait or pr. Kyle is absolutely not like that.

Unfortunately i dont go there anymore and dont even care to read it and would probably just feel sad about it. Liked their real world testing very much. But the reviews simply became weird like 5 years ago. Remarks conclusions and data didnt met.

The old man need to take a break do something else and learn to use his ears and give himself some slack. Its not an easy balance to drive a business like that i assume.

Gf 14nm might possible be a trainwreck given their track record but that is no excuse for whining and the idiotic remarks he have done in forum about p10 positioning.

 

[linkgoron]

AMD intended this chip at ~2/3 the size of the 1080 to be their high performer... what?

After mentioning to Kyle that the P10 is rumored to be 230mm^2 and GP104 is 315mm^2 he replied:

230mm^2 x 1.3 give you 300mm^2 approximately. The jump between 14nm to 16nm is considered to be ~30%, so you tell me what AMD could have imagined.

 

[Arachnotronic]

because according to Kyle:

Kyle is wrong there. Difference is pretty minimal, same metal pitches for 14/16nm but 14nm has slightly tighter gate pitch.

 

[airfathaaaaa]

i can understand him
http://www.24liveblog.com/live/1331452

its quite the trip those people had

 

[AnandThenMan]

After mentioning to Kyle that the P10 is rumored to be 230mm^2 and GP104 is 315mm^2 he replied:

Wait, he said this?

230mm^2 x 1.3 give you 300mm^2 approximately. The jump between 14nm to 16nm is considered to be ~30%, so you tell me what AMD could have imagined.

 

[jpiniero]

Yeah the difference in density between the A9 models is like 9%. But if the full P10 does have the same core configuration as GP104 it would be 35% denser, so maybe Kyle is confused.

 

[railven]

i can understand him
http://www.24liveblog.com/live/1331452

its quite the trip those people had

So any comments from people that participated are now moot since, you know, gravy train?