Vega/Navi Rumors (Updated)

[badb0y]

Based on the preview article on Vega and the Ryzen ES benchmarks I am very excited for how things shape up in the next few months. I think this is the most competitive AMD has been in like 8 years.

 

[AtenRa]

Like the ZEN will be a small core, no R&D for funds etc etc we had the same for VEGA will be a small die because they dont have the funds for R&D etc etc.

 

[ShintaiDK]

You have established that, that doesnt mean its correct though

When Techreport, an who´s owner works for AMD says its 1070-1080 performance for Vega I trust it over your guesses.

 

[AtenRa]

Hummm. Even GP100 isn't a very good GP102 competitor for gaming purposes

Near identical gaming performance I think? GP100 is significantly larger (610mm2 vs 471mm2), more power draw - so I guess slower once you really clock GP102 - no doubt a lot more expensive to make etc.

It really is a non trivial built in handicap for AMD to try and overcome this, on top of everything else. Hopefully vega is least a success in the compute market as that could get them some much needed income.

I mentioned GP100 because i believe AMD will use big Vega for both Gaming and HPC but yes you are correct GP102 is more gaming efficient than GP100

 

[AtenRa]

When Techreport, an who´s owner works for AMD says its 1070-1080 performance for Vega I trust it over your guesses.

We will see in the coming months but unless VEGA has lower perf/mm2 than Polaris , a 500-550mm2 VEGA will not be a 300mm2 Pascal competitor

 

[Qwertilot]

We will see in the coming months but unless VEGA has lower perf/mm2 than Polaris , a 500-550mm2 VEGA will not be a 300mm2 Pascal competitor

Isn't that - it being less efficient in terms of perf/mm2 than Polaris - almost a given though? They've added a whole lot of HPC stuff, and even if they've done that really well there are non trivial costs involved in doing that. (Thus GP100 vs GP102.).

Getting it near GP100 in terms of compute efficiency is likely to be the fairest yard stick, although some risks that it might end up running into GV100 there.

 

[ShintaiDK]

We will see in the coming months but unless VEGA has lower perf/mm2 than Polaris , a 500-550mm2 VEGA will not be a 300mm2 Pascal competitor

If you measure it that way, would it be a surprise? GP104 easily beats GP100 in that type of metric. GP104 also beats the current Titan X as well based on GP102. And just look at Hawaii vs Fiji.

 

[MrTeal]

I'm throwing my hat in with predictions, but given that Fury X is within 3% of a 1070 at 4k in TPU's latest test suite, it would be quite surprising if a 550mm² 14nm GPU didn't exceed the performance of a 596mm² 28nm GPU by a reasonably substantial margin.

 

[AtenRa]

And you need 2x232mm2 5.7 billion transistors to match a GP104 dont you. Or 11.4 billion transistors to match 7.2 billion. And do I have to mention the power usage?

Why i need 2x 232mm2 to compete against 300mm2 ??

232mm2 Polaris competes against 200mm2 Pascal

300mm2 Pascal is 50% larger than 200mm2 Pascal,

that means we need ~50% larger than 232mm2 Polaris to compete against 300mm2 Pascal

So, 232mm2 + 50% = 350mm2

 

[ShintaiDK]

Why i need 2x 232mm2 to compete against 300mm2 ??

232mm2 Polaris competes against 200mm2 Pascal

300mm2 Pascal is 50% larger than 200mm2 Pascal,

that means we need ~50% larger than 232mm2 Polaris to compete against 300mm2 Pascal

So, 232mm2 + 50% = 350mm2

It doesn't work that way.

https://www.techpowerup.com/reviews/NVIDIA/Titan_X_Pascal/24.html

 

[moonbogg]

I think Vega will come in between 1080-Titan level performance and will spur a 1080ti reaction from NVidia. The slow roll Vega release explains Nvidia's silence on 1080ti. Looks like 2017 is the year for slowass, delayed releases. Also, I don't know how I feel about another 8gb card coming on the scene. AMD used to be ahead of the curve in terms of Vram, but lately they are lagging behind with their high end offerings. I'd rather have 10-12gb of DDR5X than 8gb of HBM2 since the gaming performance will be the same anyway. We've seen DDR5X perform just fine despite my early reservations. People trying to drive new 4k/144hz panels with 8gb might encounter some insecurities in the next year or two to say the least.

 

[AtenRa]

It doesn't work that way.

https://www.techpowerup.com/reviews/NVIDIA/Titan_X_Pascal/24.html

It works exactly like that, we have the same architectures at different die sizes.

If architecture A at 232mm2 competes against architecture B at 200mm2

then

architecture A at 350mm2 competes against architecture B at 300mm2

 

[krumme]

I'm throwing my hat in with predictions, but given that Fury X is within 3% of a 1070 at 4k in TPU's latest test suite, it would be quite surprising if a 550mm² 14nm GPU didn't exceed the performance of a 596mm² 28nm GPU by a reasonably substantial margin.

1070 is cut down 1080. I bet 1080 plus 50% at 4k tpu suite at that time. There is a lot of weak parts of gcn fixed 2x geometry hbm2 whatnot and polaris was probably a 20nm ps4pro designed derivative beeing ported to 14nm because 20nm crashed. But at say may when we see it we are also close to volta and that will be a serious change to so it will be an interesting year. I think what we have is just silence before the storm...

 

[Phynaz]

They are using HBM2 so i believe a new Vega Nano at 175W TDP will be faster than GTX 1080 and could reach close to 1080Ti performance in DX-12 games. Fiji Nano is already neck and neck with GTX 980Ti in the majority of latest 2016 games.

Tech Report got their hands on one. They said performance was between a 1070 and 1080.

After how far off you were with your Polaris predictions, perhaps you should temper your expectations from Vega.

 

[PhonakV30]

Tech Report got their hands on one. They said performance was between a 1070 and 1080.

After how far off you were with your Polaris predictions, perhaps you should temper your expectations from Vega.

Link ? i can't find it.

 

[MrTeal]

Link ? i can't find it.

http://techreport.com/review/31224/the-curtain-comes-up-on-amd-vega-architecture/4

It seems to be based on the demo that was provided at the event last month.

 

[AtenRa]

Tech Report got their hands on one. They said performance was between a 1070 and 1080.

Got their hands on what ??? they had a closed case that AMD said it had a Vega GPU but not description of what Vega GPU.

The one we are talking TODAY is a big VEGA chip at 500-600mm2. If anyone still believes this huge die compete against 300mm2 Pascal GP104 they better start watching some cooking Youtube videos and stop talking in technical forums like AT

After how far off you were with your Polaris predictions, perhaps you should temper your expectations from Vega.

My predictions were very close, Polaris 10 is faster than GTX 980 today at latest games at lower TDP

 

[Azix]

nice tech. interested. Lots of new things in this one that should add significantly to performance and power consumption reduction. Should be more than 4GB VRAM. Worthy buy at the right price.

 

[PhonakV30]

http://techreport.com/review/31224/the-curtain-comes-up-on-amd-vega-architecture/4

It seems to be based on the demo that was provided at the event last month.

so You are telling that Big Vega Is slower than GTX1080 ?

 

[happy medium]

[

http://techreport.com/review/31224/the-curtain-comes-up-on-amd-vega-architecture/4

It seems to be based on the demo that was provided at the event last month.

but but but. that's little Vega, incoming in 3, 2, 1

It might be....

 

[Glo.]

They are using HBM2 so i believe a new Vega Nano at 175W TDP will be faster than GTX 1080 and could reach close to 1080Ti performance in DX-12 games. Fiji Nano is already neck and neck with GTX 980Ti in the majority of latest 2016 games.

3072 GCN core chip with 8 GB HBM2, and 1.5 GHz - yes it will compete with GTX 1080 at lower thermal envelope.

 

[Phynaz]

Got their hands on what ??? they had a closed case that AMD said it had a Vega GPU but not what Vega GPU.

The one we are talking TODAY is a big VEGA chip at 500-600mm2. If anyone still believes this huge die compete against 300mm2 Pascal GP104 they better start watching some cooking Youtube videos and stop talking in technical forums like AT

So you are saying AMD is so incompetent that they would promote their top of the line GPU by giving hands-on demonstations of a different GPU. I don't think so.

My predictions were very close, Polaris 10 is faster than GTX 980 today at latest games at lower TDP

You know, you can sell your GTX 980Ti at a nice price and buy a brand new Polaris 10 in Summer that will be faster and save money in the process

That's just one. Do I really need to post more? Just admit you over-hyped Polaris.

 

[AtenRa]

[

but but but. that's little Vega, incoming in 3, 2, 1

Are you really suggesting Vega has lower perf/mm2 than Polaris 10 ??

 

[happy medium]

so You are telling that Big Vega Is slower than GTX1080 ?

They said the Vega was 10% faster in Doom, so yes it would be faster than a gtx1080 on that chart but slower in every other game.

 

[AtenRa]

So you are saying AMD is so incompetent that they would promote their top of the line GPU by giving hands-on demonstations of a different GPU. I don't think so.
.

There are two (2) Vega chips, which one was on that PC case ??? AMD didnt say which one they used.

And let me ask you the same question,

Are you really suggesting that Vega has lower perf/mm2 than Polaris 10 ??