What are you trying to say? Can you be more specific?But how many other things have the talked about over that time that haven't come to fruition? You can't point to a single instance of something someone said years ago and use it to suggest it was the plan all along without also looking at all of the things that didn't come true.
But how many other things have the talked about over that time that haven't come to fruition? You can't point to a single instance of something someone said years ago and use it to suggest it was the plan all along without also looking at all of the things that didn't come true.
Well looking at the small diesize and low price they did more than talk about it. Its been planned for several years so this bs about pascal whatever is nonsense. The decisions was made like what 1-2 years ago?
Bringing a GPU to market is quite a lengthy and complicated process. The market could take an unexpected shift along the way, but everything I've seen so far points to Polaris being approximately where RTG was aiming for; small die for mass production. If it was a huge die selling for cheap then I would say something went wrong, arguably like Hawaii. This time however Polaris looks right on the mark. When I'm not traveling I'll link the interview.
Also Im guessing AMD is getting better deal from Globf. So they can get more margins out of each wafer that allows them to have more aggressive pricing.
What are you trying to say? Can you be more specific?
It doesnt make sense in this context. Its irrelevant for this reason:AMD may have talked about this plan of theirs 2 years ago, but they also would have talked about a lot of other things, many of which never happened or didn't come true.
You can't pick just one thing that they said in the past and use it to support the idea that they had planned for something all along. Otherwise you could also pick one thing that they said which didn't come true and state that all of their plans changed or they didn't know what they're talking about.
Consider the case where right now I make 15 predictions about what will happen with some tech company in 2 years time. One of those predictions happens to come true, but that doesn't mean that I actually know what I'm talking about, especially if the other 14 predictions are flat out wrong. You'd want to look at all of the predictions to get a better idea of whether I just got lucky or just making wild claims, one of which happened to be true.
Also Im guessing AMD is getting better deal from Globf. So they can get more margins out of each wafer that allows them to have more aggressive pricing.
Consider the case where right now I make 15 predictions about what will happen with some tech company in 2 years time. One of those predictions happens to come true, but that doesn't mean that I actually know what I'm talking about, especially if the other 14 predictions are flat out wrong. You'd want to look at all of the predictions to get a better idea of whether I just got lucky or just making wild claims, one of which happened to be true.
It doesnt make sense in this context. Its irrelevant for this reason:
85% of the market is below 300 usd. It doesnt change drastically.
Amd made the decision 2 years ago to market p10 p11 and vega and slice the 3 dies the size we know aprox.
Amd doesnt make two cards p10 and vega (2 of 3 sizes) for 15% of the market. They have one size vega for the 15% - its plenty and perhaps even an overshoot.
Saying p10 should in anyway target the 15% market is so far out its crazy.
Not happening. GDDR5X is around 5-8% faster costing 20% more. Not pratical for a mainstream card. :thumbsdown:
That assumes the yields are good though, which I don't think they are at this point. Better pricing could also be a case of GloFo discounting wafer prices because the yields are below agreed levels.
It doesn't really matter either way though as the price is what it is. Whether AMD makes money hand over first or nothing at all doesn't change the cost to me.
My analysis goes this way:That assumes the yields are good though, which I don't think they are at this point. Better pricing could also be a case of GloFo discounting wafer prices because the yields are below agreed levels.
It doesn't really matter either way though as the price is what it is. Whether AMD makes money hand over first or nothing at all doesn't change the cost to me.
That assumes the yields are good though, which I don't think they are at this point. Better pricing could also be a case of GloFo discounting wafer prices because the yields are below agreed levels.
It doesn't really matter either way though as the price is what it is. Whether AMD makes money hand over first or nothing at all doesn't change the cost to me.
Huh? GDDR5X is much faster than that.
GDDR5X is specced up to 14 GHz (effective), which is 75% faster than the 8 GHz GDDR5 AMD is using for the RX 480.
Even if AMD can't get their hands on GDDR5X quite that fast yet, there is still the 10 GHz GDDR5X that Nvidia is using on the 1080, which would still be 25% faster.
GDDR5X may very well be too expensive for a $300 card though.
Why don't you think yields are good? (genuine question)
That's why they only got 10G. The yields according to industry are terrible. They cannot get much more than that. So yea, its only 5 -7%.
Why don't you think yields are good? (genuine question)
Source?
It's natural to assume that yields for a brand new process are limited, but the obvious indication that 480 is yielding well and at large quantities is the very aggressive $199 price.
He clearly stated that even 10GHz GDDR5X is 25% faster than the fastest GDDR5 available, 8GHz.
How is that 5 - 7%? What on earth??
Price isn't just a factor of yields, but also of performance. If they tried selling the 480 for $300 it wouldn't sell no matter how good yields were. $200 is what they think will allow them to sell the most chips.
If yields are good, it means they make more profit on every $200 sale because each chip is less expensive to manufacture. If yields are bad, it means they don't make as much profit on each chip.
If Polaris 10 were the only GPU available they could pick a price to maximize profit, but that isn't the case so they have to choose a price to maximize revenue (sales) without regard to yield. Even if they have to lose money on each chip until yields improve, it's better than not selling any chips at the price required to profit, but one which no one will pay due to having other options.