Digitimes Says AMD RX Vega Card Shortages Will Persist

GDDR5X/GDDR6 is as good as HBM2 in power consumption. GP102 also got more bandwidth than any Vega card. (547.7GB/sec).

The only reason why anyone uses HBM2 is ECC and perhaps size. (Vega isn't using it for size). And no, HBCC also works on GDDR cards like Polaris ;)

Using GDDR5 as the compare metric when GDDR5X have been out for 15 month+ is just silly.
I really think sticking with HBM2 has more to do with marketing and desperatly trying to save face. If they abandoned HBM then everyone would declare it a failure.
 
I really think sticking with HBM2 has more to do with marketing and desperatly trying to save face. If they abandoned HBM then everyone would declare it a failure.

Sticking to HBM2 is because they try the HPC segment and require ECC. Gamers have always been secondary.

Where it goes wrong big time is that AMD has no money for R&D and IC designs (14nm is really killing them here). So another chip for gamers using GDDR5X for example was out of the question for the time being. Something as we see with Nvidia with GP/GV100 vs GP/GV102.

Navi seems to be the solution in 2019. 7nm GDDR6 4096SP and essentially just a shrinked Vega 10 with different MC. But before that in 2018, Vega 20 with 4 HBM2 stacks and FP64 for HPC only, never to be a gaming card. Again HPC before gamers.
 
Last edited:
Sticking to HBM2 is because they try the HPC segment and require ECC. Gamers have always been secondary.

Where it goes wrong big time is that AMD has no money for R&D and IC designs (14nm is really killing them here). So another chip for gamers using GDDR5X for example was out of the question for the time being. Something as we see with Nvidia with GP/GV100 vs GP/GV102.

Navi seems to be the solution in 2019. 7nm GDDR6 4096SP and essentially just a shrinked Vega 10 with different MC. But before that in 2018, Vega 20 with 4 HBM2 stacks and FP64 for HPC only, never to be a gaming card. Again HPC before gamers.
I just don't see them competeing well in HPC though; NVIDIA practically owns that market.
 
I just don't see them competeing well in HPC though; NVIDIA practically owns that market.

Its the nemesis of AMD. The segment they always fails in and sacrifice everything else to get there.

By the looks of it they pissed future consoles away and the bread and butter APU is still to be seen.

If you cant compete on performance metrics, you better be able to compete on cost. But they cant do any of that.
 
People need to stop the excuses that Navi will be a Raja design and act like he had nothing to do with everything before and is a single handed god. Also dont expect Navi to be anything than a shrinked and glorified Vega as the HPC roadmap suggest as well. 4096SP and 7nm perhaps with GDDR6. Vega 20 will be the big GPU of that timeframe.

You can bitch all you want. However nothing you say will change the fact that Vega technically is a part of the Arctic Island architecture that was designed before Raja got there.

Yeah. You probably have a vested interest in seeing AMD fail given your history of posting and pro intel/Nvidia bent, but the true facts of the matter are that Navi will be the first technology engineered completely under Raja's eye.

None of your bitching will change that.

That being said, no one is saying Navi will be a success.

Navi puts all the chips in, in terms of Raja and his legacy with AMD.

If it succeeds then we are looking at a Ryzen style Revival. If it fails then he likely will be out of a job and AMD will not have the funds two recoup the millions spent in R&D development and research. The only solution I would see for AMD at that point would be to sell off the RTG, because I don't know if they want to have to keep playing the games they are playing now with the last remaining Arctic Islands cards (Polaris and Vega) to try to sell enough to recoup their research costs & break even.
 
You can bitch all you want. However nothing you say will change the fact that Vega technically is a part of the Arctic Island architecture that was designed before Raja got there.

Yeah. You probably have a vested interest in seeing AMD fail given your history of posting and pro intel/Nvidia bent, but the true facts of the matter are that Navi will be the first technology engineered completely under Raja's eye.

None of your bitching will change that.

That being said, no one is saying Navi will be a success.

Navi puts all the chips in, in terms of Raja and his legacy with AMD.

If it succeeds then we are looking at a Ryzen style Revival. If it fails then he likely will be out of a job and AMD will not have the funds two recoup the millions spent in R&D development and research. The only solution I would see for AMD at that point would be to sell off the RTG, because I don't know if they want to have to keep playing the games they are playing now with the last remaining Arctic Islands cards (Polaris and Vega) to try to sell enough to recoup their research costs & break even.

What happens when Navi is just a shrinked Vega with possible GDDR6 and still GCN?

And Raja joined AMD again in 2013. Now 4 years later we are still waiting for his "creation" that will first come out in 2019 if not 2020. He was also CTO when something like the dreaded R600 was launched. There is no such thing as something magic he personally engineered. Everything they launched was engineered under his eyes and watch. Then you can make up whatever excuses you want to believe some fairy tale is around the corner.
 
What happens when Navi is just a shrinked Vega with possible GDDR6 and still GCN?

And Raja joined AMD again in 2013. Now 4 years later we are still waiting for his "creation" that will first come out in 2019 if not 2020. He was also CTO when something like the dreaded R600 was launched. There is no such thing as something magic he personally engineered. Everything they launched was engineered under his eyes and watch. Then you can make up whatever excuses you want to believe some fairy tale is around the corner.

You can keep your "what happens when". I already covered the possibilities.

When Raja joined, the Arctic Islands architecture was already developed. It was a poorly designed architecture, however AMD did not have the capital to Simply Scuttle everything and start over. If you had any concept of where AMD was then you would know this. In fact, all you had to do was read Kyle's take on AMD and the potential for a partnership with Intel from Rajas point of view to get a greater scope of what was going on.

You don't have to like it and you don't have to understand it. But the truth is that until they got out from this bad architecture they will not be able to move forward. And they couldn't move forward unless they made money off of this failed architecture. That is why they aren't doing a damn thing about miners buying up these cards. The mining craze was a godsend for AMD.

They can now break even on all the costs.

And this will free them up to pursue something new.

You see, Vega is actually Greenland from the old Arctic Islands architecture. You can piss and moan all you want about how long Raja has been there but the fact of the matter is the Greenland architecture and the Arctic Islands architecture as a whole was already designed.

When you consider that Polaris was originally intended to compete with nvidia's high-end and where it fell, the fact that Vega/Greenland comes within spitting distance of nvidia's 1070 and 1080 while utilizing similar architecture may indeed be a window into what changes Raja was able to implement on an otherwise poorly designed architecture, and bode well for the RTGs future.

But we simply will not know until Navi.

Navi, is not a part of the Arctic Islands architecture. It is something brand new.

That doesn't mean it won't utilize some useful pieces of architecture from the past but what it does mean is that from the ground up it is something entirely conceived and brought forth by Raja.

Therefore, we will finally be able to get a true evaluation.

Based on the tone of your posts, the part you seem most confused by is that no one is saying Navi will be great. No one is saying Navi will be poor. All we are saying is Navi will be the true look into how good or bad the RTG can be. And there is no more " wait until the next card" after it.

Because just as Raja has he had to get out from the Arctic Islands architecture in order to put forth his own vision, so too will Raja's architecture for Navi decide the path for not just the next video card but the next 4-5, which if they release every year or every other year will put AMD behind by another 4 to 10 years.

If that happens, I can see Raja getting his wish and AMD selling the RTG off to Intel, who has an inferior integrated graphics to begin with and For Whom the RTG would make sense as a purchase, even if they had to absorb some of the financial loss for R&D Which Intel is in a position to do so.

The great irony of it is if that happens Raja won't be around with Intel once the sale is made.

Everything hinges on Navi when it comes to the RTG. Up until this point it was just them trying to make lemonade from lemons.
 
Back
Top