![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Hearing rumors that RDNA4 might not have a halo/high end card in the stack (think similar to RDNA1). If that ends up being the case, Nvidia could very well just get by with a 5090, maybe a 5080, and then Lovelace continues to occupy the lower stack.
Yes, but that is where die shrinks come in, and blah blah blah, N3 is weird you depending on the type of pattern you are laying down you could get a huge 1.6x density improvement, or none at all... So pure transistor density is something like 56% smaller, but things like SRAM might only be 5% smaller. TSMC was working to counter that with increased layer counts going from 19 layers to 26 layers, but each layer you add increases the chance you have a defect in that chip and it slows down production time which results in a cost increase.
I think that’s kind of the point. If saving money is at all important to your decisions or maybe better said “bang for the buck” I’d just say skip a generation between purchases.
Sure, but lets put that all into perspective.
Top end ignores the rest of the product stack.
And every time it has been a reaction to the rest of the market.
The implication there being they are bad with money then?
Oh, so it is about competition then. You're contradicting yourself.
Actually both of those companies try and release the best TV's they can right now. The competition is incredibly fierce at the top end TV market. But it's totally different as well because the biggest process improvements they can muster year to year is 5%, which is very different from silicon which can gain 30%+ swings in a generation.
Again contradicting yourself. They are having to release updated models because of bad value. And it's only bad value because of relative cost vs competition. "Bad value" cannot exist in a void. As annoying as it is to say, Intel dominance during AMD's foibles pre-Ryzen and during Intel's 14nm++++ could be "argued" was bad value. Except it wasn't because there was no alternative.
$3000 top end….
Depends on how high you want to push.
Plus, didn't the crypto craze happen shortly after the Ampere release - especially around the time the upper tier 30 series were released? That's how I remember it, anyway. The 'normal prices/pricing' was pretty brief.
When do they lower their prices at all? Yet, there's supposedly 'improvements' or 'enhancements' to their current gen - 4070 Ti -> 4070 Ti Super etc. and these are getting 'discounts' or 'price reductions' supposedly?!?
I think we will see differently powerful GPU’s as they focus on the render methods being used by developers.
Yes and no. It is technically bringing the AD103 silicon which was previously only in the $1200 4080 to the $800 price point and spec wise is only a little worse than the 4080, but definitely an improvement on the AD104 based 4070 Ti.
Also the 4080 Super's MSRP is $1000 from 4080 $1200.
I'm going to wait for 5 series to upgrade. And I might even upgrade to AMD, who knows?
How do you know that?
It's fairly well known in general. There has been a ton of reportage directly and indirectly.
My synopsis basically is: nVidia still makes the best cards, but not by the margins nVidia fanboys say they do and they also run the company like a bunch of anti-consumer (and anti-business too, when considering how much vendor lock-in they do) jerks.
Sometimes if you lower money, you make more total revenue or profit. The economists call that elastic demand. See this web page.https://www.investopedia.com/terms/e/elasticity.asp Note that electronics are considered elastic. You are assuming that demand is inelastic.
That would suggest that you're agreeing with me. (You're telling me "no, but" and then posting an economic principle that agrees with my statements).
I've made no such assumption. Quite the opposite. You took a single sentence that I wrote entirely out of context - they were rhetorical questions to polonyc2's statements which make absolutely no economic sense.
Sorry if I took what you wrote out of context. I don[t like it when other people do that to me, so I have to apologize.
I would like to see
Intel had barely been advancing their mainstream CPU's though. Nvidia is not in that kind of position
.
Intel had barely been advancing their mainstream CPU's though. Nvidia is not in that kind of position
How much was because they could not get their 10nm node stable and had to scrap 3 or 4 generations of CPU cores as a result and how much was "we're the top and nobody can touch us" is up to debate.Intel had barely been advancing their mainstream CPU's though. Nvidia is not in that kind of position
I mean...4090 is hardly "mainstream" for consumer parts either. I'd say 4060 Ti and 4060 haven't advanced much at all either over 3060 Ti and 3060.Intel had barely been advancing their mainstream CPU's though. Nvidia is not in that kind of position
The 4060 and 7600 exist in a strange price point, the problem is what consumers want to spend or even have to spend at that entry point hasn’t really changed in 6+ years. But costs have increased all around them so that is sadly one of those deflation points where you get much less for the money.