Polaris 11 could have 8.6 billion transistors.

[DocSwag]

source: http://techfrag.com/2016/02/24/amd-polaris-11-may-have-8-6-billion-transistors-launching-mid-year/

AMD seems to have gone to twitter teasing Polaris again

Quote

https://twitter.com/intent/retweet?tweet_id=700420476070301697

It seems that Chris Hook could be saying Polaris has 8.6 billion transistors.

However, I personally feel the 16 million part might also be a bit of a hint. Perhaps Polaris 10 has 8.6 billion and 11 has 16 billion?

Quote

Internet is currently trying to decrypt the recent Twitter conversation between AMD graphics boss Raja Koduri and Marketing Director Chris Hook. While the tweets from two AMD officials don’t directly mention the transistor count on AMD’s upcoming Polaris 11 chip, they do mention the Capsaicin (chilli pepper’s active component) which could somehow refer to 8.6 billion transistors.

As well, they seem to be talking about hot and spicy stuff, which could either be just confirming that Polaris is really efficient or could mean AMD is planning on releasing an extremely powerful version of Polaris.

I personally am not sure what this all means. Tell me your thoughts!

[kelvindeschutter]

looks like amd is teasing the big guns and getting ready to beat NVIDIA

[werto165]

I'm guessing it's the scoville ratings of those two chillis multiplied together. 

[Humbug]

1 minute ago, kelvindeschutter said:

looks like amd is teasing the big guns and getting ready to beat NVIDIA

I'm betting that regardless of who wins the polaris vs pascal round they will be within 15% of each others general performance... Like they always are. Don't see anybody totally wiping the floor with it, based on recent GPU wars history...

[Starelementpoke]

6 minutes ago, DocSwag said:

source: http://techfrag.com/2016/02/24/amd-polaris-11-may-have-8-6-billion-transistors-launching-mid-year/

AMD seems to have gone to twitter teasing Polaris again

It seems that Chris Hook could be saying Polaris has 8.6 billion transistors.

However, I personally feel the 16 million part might also be a bit of a hint. Perhaps Polaris 10 has 8.6 billion and 11 has 16 billion?

As well, they seem to be talking about hot and spicy stuff, which could either be just confirming that Polaris is really efficient or could mean AMD is planning on releasing an extremely powerful version of Polaris.

I personally am not sure what this all means. Tell me your thoughts!

AMD got a 15 kill streak. Tactical nuke against nivida.

[werto165]

1 minute ago, Starelementpoke said:

AMD got a 15 kill streak. Tactical nuke against nivida.

thought it was 25 kills .

[Starelementpoke]

Just now, werto165 said:

thought it was 25 kills .

I dunno, I don´t play cod.

[werto165]

Just now, Starelementpoke said:

I dunno, I don´t play cod.

neither do I... 

[don_svetlio]

Wasn't Fury X 8,6 Billion? Shouldn't it be more?

[werto165]

1 minute ago, don_svetlio said:

Wasn't Fury X 8,6 Billion? Shouldn't it be more?

over 10bn apparently.  

 

yeah it was. 

[don_svetlio]

Just now, werto165 said:

over 10bn apparently. 

Smells good

[werto165]

Just now, don_svetlio said:

Smells good

I think that was wrong, me not checking with more websites... 

[AlTech]

Just now, don_svetlio said:

Wasn't Fury X 8,6 Billion? Shouldn't it be more?

Fiji was 8.9 Billion.

Tonga was 5 Billion.

Hawaii was 6.2 Billion

 

So Polaris 11 is a let down already. Unless it means that it's the successor to Hawaii. In which case it makes perfect sense.

 

That would mean Vega 10 would be the successor to Fiji.

[don_svetlio]

1 minute ago, AluminiumTech said:

Fiji was 8.9 Billion.

Tonga was 5 Billion.

Hawaii was 6.2 Billion

 

So Polaris 11 is a let down already. Unless it means that it's the successor to Hawaii. In which case it makes perfect sense.

 

That would mean Vega 10 would be the successor to Fiji.

I think he might be saying the minimum in the line-up will be 8.6 - look at the wording again

nothing less than the 8-methyl-N-vanillyl-6-nonenamide-infused Carolina Reaper will do...

[werto165]

I know absolutely nothing when it comes to making GPU's but isn't it similar to the way you can have IPC gains with CPU's you could possibly have a faster GPU with the same number or a similar number of transistors? 

[AlTech]

1 minute ago, werto165 said:

I know absolutely nothing when it comes to making GPU's but isn't it similar to the way you can have IPC gains with CPU's you could possibly have a faster GPU with the same number or a similar number of transistors? 

SHouldn't be. Only perf comes from more Transistors, better architecture or higher clock speed. Not a new process by itself.

[AlTech]

5 minutes ago, don_svetlio said:

I think he might be saying the minimum in the line-up will be 8.6 - look at the wording again

nothing less than the 8-methyl-N-vanillyl-6-nonenamide-infused Carolina Reaper will do...

What is Polaris 10 supposed to replace though?

 

If it replaces Tonga and Bonaire in one go then yes it makes sense but from 2 ish to 8.6? damn. that would be around the 4x they said originally. So then i'm guessing Vega 10 will have 32 billion, and Polaris 11 will have 24 Billion.

 

If that is even possible. 8.9 was the max that 28nm could handle. AMD built the largest die they could. They can't do that on their first 14nm architecture. It leaves no space for improvement.

[Sauron]

Not very meaningful tbh, until we have performance numbers there's no way of knowing how it will actually perform. Same for nvidia's next architecture.

[TidaLWaveZ]

Pascal has been rumored to have 17 billion transistors.  Titan X already has been out for quite a while with 8 billion transistors and the Fury X with 8.9 billion.

 

Why is this a big deal?

[Prysin]

You guys forget... raja koduri is indian. Indian food uses some of the most insane spice mixes, including the infamous ghost chilli.

Guess they are discussing hos to feed the rumor-mill (and wccf tech)

[Prysin]

8 minutes ago, AluminiumTech said:

What is Polaris 10 supposed to replace though?

 

If it replaces Tonga and Bonaire in one go then yes it makes sense but from 2 ish to 8.6? damn. that would be around the 4x they said originally. So then i'm guessing Vega 10 will have 32 billion, and Polaris 11 will have 24 Billion.

 

If that is even possible. 8.9 was the max that 28nm could handle. AMD built the largest die they could. They can't do that on their first 14nm architecture. It leaves no space for improvement.

They can, if they plan to pull another rebrand party next year or the one after. Simply pull another 300 series. Hue hue hue.

 

Or, they know samsungs 10nm is going to arrive fast enough that they can jump down yet another node very soon.

[Citadelen]

36 minutes ago, DocSwag said:

-snip-

Polaris 10 has up to 19 billion transistors., as opposed to Pascals 16, because of the process being 2nm smaller.

 

[DocSwag]

2 minutes ago, Citadelen said:

Polaris 10 has up to 19 billion transistors., as opposed to Pascals 16, because of the process being 2nm smaller.

 

Processes can't be compared like this. There are many other factors that are in play. AMD also tends to create smaller dies... Fiji was an exception as if they didn't go so big Nvidia would blow them out at the high end.

Sure, the 14nm transistors will be smaller, but also remember about transistor density. Intel has extremely high transistor density and (I believe) Intel can fit more transistors in the same space as, say GloFo or Samsung even though they are both on 14nm.

[Citadelen]

1 minute ago, DocSwag said:

-snip-

I drew the basic dots, whether that's the reason or not, it's still got more transistors.

[DocSwag]

6 minutes ago, Citadelen said:

I drew the basic dots, whether that's the reason or not, it's still got more transistors.

Nothing will be sure until they release it.

Goddamnit AMD just release it already!