AMD: Speculation, Rumors, and Discussion (Archive)

Status
Not open for further replies.
nV hasn't had an operating loss in a long time, And I think that was the last time they reported a loss. But Revenue also dropped though in those Q's (q4 2009 and q1 2010)

Hence, there is no way in hell that Nvidia would price a GP 104 based card anywhere near 199 USD. That would generate a significant hit on their operating margins which they aren't keen on doing. IE - they don't want to go into another price war like what happened with the GTX 280/260 versus 4870/4850.

Regards,
SB
 
Well RX 480 at 150 watts vs Nvidia 1070 GTX at 150 watts with the former with a performance similar to a 980 and the later to a 980TI tells us than Polaris architecture with respect to Pascal has almost not moved a pinch from where it stood at 28 nm with respect to Nvidia Maxwell architecture efficiency.Of course they have to play the price card, at least until 1060 is out.
We don't actually know the actual power consumption yet, just that it's max 150W.
 
That old 2008 price war had NV selling >550mm^2 512-bit GPUs for around the same prices as AMD 260mm^2 256-bit GPUs. And NV still did pretty well, and still gained marketshare.
 
Lol, you want a source for a hypothetical scenario? The source is me.

You made the ridiculous claim that a 315mm^2 lower binned chip + 8GB of the most expensive non-X GDDR5 would make nVidia "a crap ton more money" than AMD's 232mm^2 chip with 4GB GDDR5 if both were sold at $199. Because something something operating margins something something production costs.

And the source is you.

Okay. I know where these claims are coming from.
 
A 199 USD GTX 1070 would have significantly less profit margins (assuming it had a profit margin at all at that price) than a 199 USD Rx 480.

Notice I said nothing about profit. You can't make that statement unless you know AMD's and nVidia's cost structures. Hint - it's a lot more complicated than comparing die area.

I'm repeating myself now but the original point was that consumer pricing has little to do with cost and even less to do with die sizes. The 480 isn't $199 because of its die size.
 
I am curious if AMD is able to move as much product as the price suggests.

At least until a GP106 comes out, the typical inverse relationship between price and volume might insulate Polaris from GP104.
Nvidia currently can't satisfy the volume demands of the price tiers it is selling at, and it's high enough that I wonder if even if there were a further cut-down SKU below the 1070, whether it would hit a pricing tier (and its volume) that would absorb product and Nvidia's attention before getting to the 480.

Shouldn't Nvidia's yields be decent by now for GP104? Is there enough for a salvage SKU left over after selling in the higher tiers, if they aren't cutting them out of the wafer with a Dremel?
 
Self-awareness much?


He is talking about net profit. gross profits don't mean much to company health.

And yeah nV can cut down if they want to, but is there a need to.

AMD just replaced a midrange card with another midrange card. Tonga is 200 bucks and the new midrange card is 200 bucks, performance went up from a midrange segment card *tonga, to a performance segment card 390x, this is a fairly typical increase.

nV has two options, either make a 1060 with similar performance characteristics of a 970 and 980 or go higher matching over clocked versions. I think they will go with matching overclocked versions since they did that with their performance segment. If they do, they don't need to drop prices to compete. No price war needed.

The reason why I think they will stick with the match over clocked performance of the 970/980 is because they need something that offers more than what current people have. Otherwise people looking to upgrade can get 970's/980 left over stock or used for cheap.
 
Last edited:
Hence, there is no way in hell that Nvidia would price a GP 104 based card anywhere near 199 USD. That would generate a significant hit on their operating margins which they aren't keen on doing. IE - they don't want to go into another price war like what happened with the GTX 280/260 versus 4870/4850.

Regards,
SB


more so than that, nV was forced to, they would have lost a lot more if they didn't go into a price war, the 4xxx series was perf.watt better, close to the same performance *within 10% and priced at half the cost. There was no way any OEM or upgrader could ignore any of that.
 
$200 for r9 290 performance or better is a great deal. I may put this in my gf's pc for civ 6 at the end of the year. Also zen if its performance and price is good. Her 6 core phenom isn't holding up that well anymore
 
this is good, and i have nothing to add, but off topic can anyone explain why the word uplift is being used of recent times in corporate tech jargon instead of increase.

its stupid, its as stupid as saying: "this graphics chip will provide a 20% jumpboost in videocarding."

ive seen it thrown around a few times and i just couldnt take it anymore, theres no reason to invent a compound word when an actual word exists that means what the made up one is supposed to mean.

anyway continue.
 
Is it TDP or AMD's Typical Board Power, which Fury X has shown has a margin of error roughly equivalent to one Polaris 10?
At least with Polaris and unlike other AMD boards, the 6-pin connector gives them less wiggle room on "typically" not blowing out the marketing number for board consumption. It's not a perfect limit, but it's more definitive than what we've gotten for a while.

Furmark makes Fury X power consumption go up for rougly another FuryX(250W ACP from my calculations). Could it be that the program can stress GCN to a point that the whole chip goes fully utilized? Anyway Fury X gaming power consumption is around 390x power, is not that much on gaming loads. Furmark/Bitming power matters, but gaming power consumption matters more.
 
this is good, and i have nothing to add, but off topic can anyone explain why the word uplift is being used of recent times in corporate tech jargon instead of increase.
Because "increase" is overused. People tune it out. When you're trying to sell products, language matters. You need phrases and messaging that will stand out in the buyer's mind.
 
That old 2008 price war had NV selling >550mm^2 512-bit GPUs for around the same prices as AMD 260mm^2 256-bit GPUs. And NV still did pretty well, and still gained marketshare.

Actually not only did they lose market share (2%) they also operated at a loss. Which wasn't healthy for the company.

Of course, they didn't have a lot of choice, at the prospective price points, GTX 280/260 didn't compare well with 4870/4850. It was significantly larger and more power hungry without much of a performance lead. Hence they were able to limit the market share loss to only 2% but at the cost of operating at a net loss. In hindsight I think Nvidia would have preferred to lose 5% or more market share if they had been able to maintain operating profits instead of operating at a net loss.

They managed to regain market share and then some once they got out smaller and more competitive chips (GF 9xxx series). You'll notice that later when they lost a lot of marketshare during 5870/5850 due to being late to market, they didn't go into a price war. They lost market share which certainly hurt but by kept their ASP high. This allowed them to keep their operating margins high, meant they could operate at a net profit. Lessons learned from the detrimental price wars from the previous generation.

GP 104 versus Polaris 10 is a completely different situation. Both chips are not only in different price bracket (like those chips from the price wars), but they are also in completely different performance tiers (completely unlike those previously mentioned chips). Nvidia has no reason to go into a price war and damage their operating margins or operating profit.

So do we see a return of the price wars from 4870/4850 days where Nvidia suffered net operating losses by engaging in a price war? Especially as Polaris 10 isn't a performance threat to GP 104, unlike Rv770. They'd do this by aggressively pricing GP 104 versus Polaris 10.

Or do we see the Nvidia from the 5870/5850 era where they lost market share but kept their high ASP and operating margins, thus maintaining net operating profits? They'd do this by continuing to sell through all of their inventory of GP 104 at the higher ASP.

In short, Nvidia would either have to be desperate (Polaris 10 eating significantly into sales of GP 104 despite the large performance discrepancy) or their yields on GP 104 would have to be absolutely horrible that they have a bunch of salvage dies that can't even go into GTX 1070. Neither of which I believe is likely to happen.

Regards,
SB
 
Actually not only did they lose market share (2%) they also operated at a loss. Which wasn't healthy for the company.

Of course, they didn't have a lot of choice, at the prospective price points, GTX 280/260 didn't compare well with 4870/4850. It was significantly larger and more power hungry without much of a performance lead. Hence they were able to limit the market share loss to only 2% but at the cost of operating at a net loss. In hindsight I think Nvidia would have preferred to lose 5% or more market share if they had been able to maintain operating profits instead of operating at a net loss.

They managed to regain market share and then some once they got out smaller and more competitive chips (GF 9xxx series). You'll notice that later when they lost a lot of marketshare during 5870/5850 due to being late to market, they didn't go into a price war. They lost market share which certainly hurt but by kept their ASP high. This allowed them to keep their operating margins high, meant they could operate at a net profit. Lessons learned from the detrimental price wars from the previous generation.

GP 104 versus Polaris 10 is a completely different situation. Both chips are not only in different price bracket (like those chips from the price wars), but they are also in completely different performance tiers (completely unlike those previously mentioned chips). Nvidia has no reason to go into a price war and damage their operating margins or operating profit.

So do we see a return of the price wars from 4870/4850 days where Nvidia suffered net operating losses by engaging in a price war? Especially as Polaris 10 isn't a performance threat to GP 104, unlike Rv770. They'd do this by aggressively pricing GP 104 versus Polaris 10.

Or do we see the Nvidia from the 5870/5850 era where they lost market share but kept their high ASP and operating margins, thus maintaining net operating profits? They'd do this by continuing to sell through all of their inventory of GP 104 at the higher ASP.

In short, Nvidia would either have to be desperate (Polaris 10 eating significantly into sales of GP 104 despite the large performance discrepancy) or their yields on GP 104 would have to be absolutely horrible that they have a bunch of salvage dies that can't even go into GTX 1070. Neither of which I believe is likely to happen.

Regards,
SB

Looking at this market share chart, the numbers are a bit bizarre for the GT200 RV770 era. NV reached a record level during it, but then tanked before 5870 launched.
https://forum.beyond3d.com/posts/1916435/

Yeah I don't see Polaris inciting a price war unless it has spectacular performance for $200 and that seems unlikely.
 
Last edited:
Status
Not open for further replies.
Back
Top