What do you think the top end Vega model needs to do to be successul?

For me, the success of Vega means the 1080 gtx performance for 1070 gtx price, that uses less power than 1080, comes with completely redesigned stock cooler, that does not operate on jet engine noise level and does not have throttling issues. I know, that's not gonna happen, but one can hope.
 
Personally i'm much more excited for Ryzen than Vega, and I'm on a 5820k + 1070 combo.

The problem with being this late is that you already lost most of your potential customers to the competitor already, so even if your product is competitive you're still serving a smaller market.
 
Personally i'm much more excited for Ryzen than Vega, and I'm on a 5820k + 1070 combo.

The problem with being this late is that you already lost most of your potential customers to the competitor already, so even if your product is competitive you're still serving a smaller market.


Same here, I don't have confidence in AMD graphics division, and the R&D cut backs they suffered due to their CPU division. I think if Ryzen reaches what has been shown so far in over all benchmarks, Navi will be the best hope for AMD to catch up to nV as money can be put back into R&D.
 
There will be plenty of buyers from previous generations that if Vega is worth while may invest and have fun with. Not everyone just automatically immediately bought Nvidia Pascals. Many has but far more have not.

If AMD has a very compelling product, it will sell well is the short answer. If not then it may just limp along.

Now I hope AMD concentrates somewhat on VR since they are the ones broadcasting the so call premium experience.
 
There will be plenty of buyers from previous generations that if Vega is worth while may invest and have fun with. Not everyone just automatically immediately bought Nvidia Pascals. Many has but far more have not.
If AMD has a very compelling product, it will sell well is the short answer. If not then it may just limp along.
Now I hope AMD concentrates somewhat on VR since they are the ones broadcasting the so call premium experience.

The problem with VR is that it is not mainstream :) . So what little AMD can do even if they owned the VR side of business AMD would still have a hard time generating money. The knee jerk reaction from AMD regarding their VR claims prove that there trying to force brand recognition upon Virtual Reality where everywhere in the stock market VR is still considered a growing market.

AMD would do what they always tend to do a product with a good price performance ratio. The only drawback for Vega would be if the consumer version came with HBM2 where the cost would make it not price/performance competitive.
 
There will be plenty of buyers from previous generations that if Vega is worth while may invest and have fun with. Not everyone just automatically immediately bought Nvidia Pascals. Many has but far more have not.

They have to price it in a completely different price range then, else most of the buyers are gone. And then we are back to the profitability question. Vega better be more than 1080ish performance at 250W, 450mm2 and HBM2.

AMD would do what they always tend to do a product with a good price performance ratio. The only drawback for Vega would be if the consumer version came with HBM2 where the cost would make it not price/performance competitive.

Polaris 10 as the latest is a good example of the bad cost structure. Needs 33% more memory bandwidth, 37% more power and 30% more transistors to compete against GP106. But this unfortunately have been the standard for a long time including CPUs. And AMD have to pay the cost of its own pocket to compete. They cant charge 30% extra for Polaris due to it.
 
Last edited:
They have to price it in a completely different price range then, else most of the buyers are gone. And then we are back to the profitability question. Vega better be more than 1080ish performance at 250W, 450mm2 and HBM2.



Polaris 10 as the latest is a good example of the bad cost structure. Needs 33% more memory bandwidth, 37% more power and 30% more transistors to compete against GP106. But this unfortunately have been the standard for a long time including CPUs. And AMD have to pay the cost of its own pocket to compete. They cant charge 30% extra for Polaris due to it.
Nvidia just went clock Nazi and amazingly kept the power down. Brilliant engineering in the end. 1080 would be nice to beat at a lower price point but the 1070 is where AMD really needs to come ahead. As for the 1080Ti - not sure if AMD will be able to beat or compete unless they have a 16gb full Vega that OC's well. Not sure how a dual GPU card would sell in the end.

AMD can compete with folks looking at SFF with Nano type small cards that pack a punch - this time they also need to incorporate the Pro models as well with the series. Since Nvidia does not have a big chip small package like the Nano. So in other words a Vega that performs 1080 or better in a card same or smaller than a Nano. Would also help if they kept the power down to 175w or less since for cases that the Nano barely fit into don't need a lot of heat being dumped into them. So having power options for low power and options to really push it would be ideal.
 
I see it as AMD needs a product that gives 1080 performance with a 1070 price in a R9 Nano size package with HBM2…

And I would LOVE to see a Ryzen-based APU with a minimum of quad CPU cores (with multi-threading) & dedicated on-die HBM2 for the GPU cores, that offers 1060 or better gaming performance…!!!
 
I see it as AMD needs a product that gives 1080 performance with a 1070 price in a R9 Nano size package with HBM2…

And I would LOVE to see a Ryzen-based APU with a minimum of quad CPU cores (with multi-threading) & dedicated on-die HBM2 for the GPU cores, that offers 1060 or better gaming performance…!!!

I really want to see Vega being faster than the 1080.

As for Zen, it won't have GPU performance like the 1060. It would pretty much kill off the whole Polaris range in one go. AMD don't need that.
 
Nvidia just went clock Nazi and amazingly kept the power down. Brilliant engineering in the end. 1080 would be nice to beat at a lower price point but the 1070 is where AMD really needs to come ahead. As for the 1080Ti - not sure if AMD will be able to beat or compete unless they have a 16gb full Vega that OC's well. Not sure how a dual GPU card would sell in the end.

AMD can compete with folks looking at SFF with Nano type small cards that pack a punch - this time they also need to incorporate the Pro models as well with the series. Since Nvidia does not have a big chip small package like the Nano. So in other words a Vega that performs 1080 or better in a card same or smaller than a Nano. Would also help if they kept the power down to 175w or less since for cases that the Nano barely fit into don't need a lot of heat being dumped into them. So having power options for low power and options to really push it would be ideal.

There isn't really any sales in SFF cards, its a niche of ITX. ITX setups tends to either feature full cards or none at all.
 
I'm looking forward to Zen, I pretty much gave up AMD's graphic lines. I mean, I feel spoiled getting regular driver updates and support now, sometimes I even feel like dammit not again, where as with the 290 I had, crickets, or a beta driver once and awhile.
 
There isn't really any sales in SFF cards, its a niche of ITX. ITX setups tends to either feature full cards or none at all.
Your right but having a professional version (higher price tag as well) would also help. Apple could bite as well.

I do not think AMD will give up on small potent cards when HBM2 allows it effectively. Marketing wise AMD could do much better for the smaller is bigger better angle. Also to start building that market segment where you can own the high end on it is not a bad place to start.
 
I'm looking forward to Zen, I pretty much gave up AMD's graphic lines. I mean, I feel spoiled getting regular driver updates and support now, sometimes I even feel like dammit not again, where as with the 290 I had, crickets, or a beta driver once and awhile.
Personally I am impressed with AMD software/driver side for their graphics cards. OCing built in (with volt modification and states plus graphing with monitoring). Relive for capturing and streaming except I am having a hard time getting it to work in Doom. The UI looks slick, modern and easy to navigate - Nvidia drivers looks like from the stone age plus GeForce Experience is mostly advertising spam which I hate. Updates now are way frequent - almost too frequent. Now I hope Vega will surprise us in a pleasant way and not really be 2nd quarter of next year - That would suck.
 
Your right but having a professional version (higher price tag as well) would also help. Apple could bite as well.

I do not think AMD will give up on small potent cards when HBM2 allows it effectively. Marketing wise AMD could do much better for the smaller is bigger better angle. Also to start building that market segment where you can own the high end on it is not a bad place to start.

Smaller card=more noise due to smaller cooler without using premium components. Unless power consumption goes down significantly.
 
Smaller card=more noise due to smaller cooler without using premium components. Unless power consumption goes down significantly.
Rubber mounted fans/blowers, curved technology directing sound to isolation chambers, sound absorbing materials (porous), sound isolation enclosures have not been implemented on graphics cards yet. Specially designed SFF cases would also help. Many possibilities do exist for noise reduction and also opportunities as well for multi-workstation in close quarters. I think we are just scratching the surface for reducing noise - simplest trick for home users is just donning a high quality sound isolating headphone ;).

Anyways I have money burning in my pockets, pillow case, cat box for a powerful Nano size Vega.
 
Rubber mounted fans/blowers, curved technology directing sound to isolation chambers, sound absorbing materials (porous), sound isolation enclosures have not been implemented on graphics cards yet. Specially designed SFF cases would also help. Many possibilities do exist for noise reduction and also opportunities as well for multi-workstation in close quarters. I think we are just scratching the surface for reducing noise - simplest trick for home users is just donning a high quality sound isolating headphone ;).

Anyways I have money burning in my pockets, pillow case, cat box for a powerful Nano size Vega.

Oh I agree as such. The future will be tiny quiet PCs. But it will also come from a much lower power consumption usage. Despite extra cost spend on quality to make them more attractive. I wouldn't be surprised if 10 years down the road, ITX will be considered big.

But power consumption has to come down.
 
Oh I agree as such. The future will be tiny quiet PCs. But it will also come from a much lower power consumption usage. Despite extra cost spend on quality to make them more attractive. I wouldn't be surprised if 10 years down the road, ITX will be considered big.

But power consumption has to come down.
I do agree with you, AMD has been doing this but Nvidia so far has been better with GPU's with power efficiency. Polaris obviously fell short of AMD goals, does not mean Vega will not be closer to them. Still probably too far for an ideal small system but many of the issues can be mitigated until real solutions can be achieved. I expect Nvidia Volta with HBM2 will really take advantage of the small once that project reaches maturity and becomes available. I am sure AMD at least woken up Nvidia for new possibilities for GPU use - systems that replaces consoles and PC's or an option is my first thought. Folks are not as interested overall in the higher end PC's, they may be if it is better integrated with their TV's, VR space etc. Plus options for effectively streaming transparently anywhere any device without costing a fortune.
 
I do agree with you, AMD has been doing this but Nvidia so far has been better with GPU's with power efficiency. Polaris obviously fell short of AMD goals, does not mean Vega will not be closer to them. Still probably too far for an ideal small system but many of the issues can be mitigated until real solutions can be achieved. I expect Nvidia Volta with HBM2 will really take advantage of the small once that project reaches maturity and becomes available. I am sure AMD at least woken up Nvidia for new possibilities for GPU use - systems that replaces consoles and PC's or an option is my first thought. Folks are not as interested overall in the higher end PC's, they may be if it is better integrated with their TV's, VR space etc. Plus options for effectively streaming transparently anywhere any device without costing a fortune.

You dont need HBM to get a small card. Also I doubt Volta will use HBM outside the 100 and 102 series. GDDR5X/GDDR6 for rest. Again cost and need.

AMD haven't gone anywhere with power efficiency besides nodes. And that's a big problem, as such 28nm Maxwell can compete with 14LPP GCN 1.3.

People go up in SKUs, make no mistake. Higher end cards sell better. Its the lower end that suffers.
 
For me to consider buying it, it has to be faster. If AMD offers a 1080 competitors, then it better be faster. That card is 6 months old. The same with 1070. When it comes to Titan XP then I would accept same performance for less money. And I don't care if their cards are future proof or not. I buy new card(s) every 2 years anyway.
 
It's coming so late, it better be legendary. I would've been happy with 1080 performance if it released with zen.

If it releases in the summer, It needs 20% more performance than a 1080 in both DX11 and DX12 to be worth the wait (for me)
 
Back
Top