Speculation and Rumors: Nvidia Blackwell ...

Seanspeed said:
We all know and have seen how Nvidia architectures more than a single generation old tend to fall backwards heavily in performance in newer games.
Click to expand...
The 3080/3090 vs 6800XT/6900XT comparisons don't agree with this assessment. The 2070/2070 Super vs 5700XT comparisons don't agree with this assessment, nor did the 1080 vs Vega 64 (now out of driver support) comparisons, nor did the 980Ti vs FuryX (now out of driver support) comparisons. All cards remain at their relative positions compared to the competition.
 
Last edited:
Seanspeed said:
Arguing that we should be paying based on level of performance improvement is basically the kind of thing you'd expect a senior exec at Nvidia to argue. To see consumers arguing this is insanely depressing.
Click to expand...

What a luxury product costs to manufacture is literally none of our business. The only thing that matters to a paying customer is the value of the delivered product. If the market thinks it’s poor value it won’t sell.
 
Seanspeed said:
That's literally how things have ALWAYS worked in the processor world up until super recently. And still how it actually works with CPU's.
Click to expand...
Patently false, and belies your complete inability to understand pricing of products versus their component cost in this or apparently any related field. The price you've paid for any CPU, literally ANY CPU in the last 30 years, has been wholly unrelated to the physical cost of the silicon or substrate of that part. Full stop, this is inarguable for anyone who actually works in this field.

Despite your overtures to the contrary, prices are set to what the market will bear. That's how ALL of this works, and when I mean ALL, I'm not just talking about compute hardware.
 
trinibwoy said:
What a luxury product costs to manufacture is literally none of our business. The only thing that matters to a paying customer is the value of the delivered product. If the market thinks it’s poor value it won’t sell.
Click to expand...
It absolutely should matter, and used to matter.

Pascal generation gave us massively improved performance without some giant exploitation of consumers in the process. This will get ignored in all arguments made going forwards.

This idea of "Well if people are willing to pay it, we should totally be ok with it" philosophy is insane to me. This is a literal pro-corporation stance.
 
Albuquerque said:
Patently false, and belies your complete inability to understand anything in this segment. The price you've paid for any CPU, literally ANY CPU in the last 30 years, has been wholly unrelated to the physical cost of the silicon or substrate of that part. Full stop, this is inarguable for anyone who actually works in this field.
Click to expand...
You heard it here first folks - there's literally no reason that a 7900X would cost more than a 7700X.
 
The cost of the physical bill of materials has never mattered; this is why boutique brands and halo products have existed for the duration of the lives of the folks posting in this thread.

Every one of us is a consumer, and all of us would love to see better, faster, AND cheaper in our next products. That's not how life actually works, whether any of us like it or not. Rather than tilting at windmills, please consider NOT telling us how a perfect world might operate since none of us have ever lived there and will not be living there for the reasonable future.
 
Mod mode: Nope, I've had it. For anyone else who I didn't just give a two-week timeout to, don't act purposefully obtuse to rational conversation. If you can't chill out, then find yourself unable to post for a few days or more. I've had it with the five year old retorts.
 
Seanspeed said:
You heard it here first folks - there's literally no reason that a 7900X would cost more than a 7700X.
Click to expand...
Now do 7700X and 8700G. The latter is more expensive to make, yet AMD still prices it cheaper, due to lower CPU performance and a general expectation of APUs as a budget lineup.

By a different argument - even when only considering rasterisation performance, AMD needs more silicon to equal nvidia's performance. Should they both be pricing according to their bill of materials, and AMD accordingly just have completely unviable products?
 
Last edited:
I think most of us in this thread understand a bill of materials isn't causally linked to the MSRP of any consumer product. Automobiles, homes, food stuffs, electronics, entertainment, essentially our entire lives are priced as a function of a consumer's perception of worth.

If someone thinks your handbag is worth $1000, then it doesn't matter if the bill of materials is $25 or $925. We all know this instinctively, and we've known it since most of us were adolescent.

An interesting point to consider along this topic: for all the folks who are employed (sorry, not me at the moment), it's worth noting your payrate is similarly a function of perceived value. Be careful wishing for a world where valuation is purely a function of BOM, because a LOT of people might end up taking a pretty severe pay cut. It's easy to cheer for the obvious losers, executives and other "senior leadership" members. But that perception of value payscale applies to everyone, and while a BOM-based system of remuneration might be the great equalizer and might perhaps be best for a utopian society, it's going to cause severe mayhem for a LOT of folks who "only" make $80k-200k and feel like they're middle class.
 
It's way more complex than just consumer perception. Prices are a result of competition for consumer interest which means that sometimes we get lower than we'd expect prices while sometimes the prices aren't going anywhere - or increasing even. The relation between pricing and demand is elastic which means that there is a rather huge range in where some product can be priced - depending on the manufacturer considerations about the size of the addressable market and general consumer willingness to spend as much on something which they don't actually need (generally speaking PC h/w isn't something anyone need to survive).

Then there's inflation which already made $500 from 2000 into $900+ now which also plays into the "perception" when someone starts comparing "x-something-series" as if nothing's changed between GF256 and RTX4090.
 
Wouldn't competition affect a consumer's perception of value? If someone considers a lesser-priced product B to be of equivalant value to a higher-priced product A, doesn't that adjust the consumer's perceived value of both products? Creating an artificial vacuum of demand from decreased or limited production runs means fewer people will be able to get those items, thus continuing to influence the halo perception of the product. Going back to my old car analogy, no small part of the reason Ferraris are so expensive isn't just the BOM or the artistry or the performance, it's also the extremely limited production runs of any of their cars. Even if you could afford the SF90, there's a high probability you will NOT be able to buy one because they're all spoken for. Of course, you can wait to buy one used in a few years, for a considerable markup.

Not specifically trying to say you're wrong, just thinking "out loud."

Inflation and other economic factors certainly play important roles in pricing, unrelated to consumer perception.
 
Price sometimes influences our perception of value but I don’t think I it works for graphics cards. It’s too easy to compare the most expensive version to cheaper variants using quantifiable metrics. So then it comes down to how much premium you’re willing to pay to have the best. There’s no illusion that a 4090 is 60% “better” than a 4080 due to the price difference when benchmarks say otherwise.
 
Right, I agree with that stance. I think it can go both ways -- something which is catastrophically expensive to "normal" people could be seen as either a status symbol of wealth, or just flagrant disregard for anything but capitalistic greed.

A contrived example might be a $2500 T-shirt from some name brand haute couture (or however it's spelled) shop on Hollywood Blvd. It's a frickin' T-shirt, why would anyone pay $2500 for it when I can buy a six pack at Costco for $9.99? But hey, if you're really into high fashion, that $2500 T-shirt might mean something to you. It's lost on me...

I had an old CIO who liked to say "perception is reality" and I really hated when he quipped it. My inner monologue always wanted to retort back that reality was reality, and percpetion is some bullshit people tell themselves to delude themselves about their place in life. It didn't dawn on me until years later how right he ended up being, and wasn't really commenting on pricing of electronics during those conversations.
 
Last edited:
If the 5080 is around 35% faster than a 4080 then that’ll make it probably 80% faster than my 3080 which is a nice upgrade depending on the price. Don’t care what the bus width is. If it’s faster it’s faster.

I just tried to watch this hardware unboxed video and off the rip it’s so dumb. Why does anyone care about tiers if the don’t work for the company? You have a product. Look at the alternatives and check if they’re better and if it’s worth the cost. If you don’t like it don’t buy it. Skip a generation if there’s nothing in your price range that’s worth it. Who cares about a product in some price tier you wouldn’t buy anyway. The only relevant comparison is to what you already own.
 
Last edited:
DegustatoR said:
www.tweaktown.com

NVIDIA's next-gen GeForce RTX 5090 Laptop GPU specs leak: 24GB GDDR7 on a 256-bit memory bus

NVIDIA's next-generation GeForce RTX 5090 Laptop GPU rumored specs: 24GB of ultra-fast GDDR7 memory on a 256-bit memory bus, launching in Q1 2025.
www.tweaktown.com www.tweaktown.com
The die size there is about the same on AD103 which means that the clocks would be the only way to reach 4090's performance.
MLID though so doubt and all.
Click to expand...

That reads less like a rumor and more like random speculation.
 
Scott_Arm said:
If the 5080 is around 35% faster than a 4080 then that’ll make it probably 80% faster than my 3080 which is a nice upgrade depending on the price. Don’t care what the bus width is. If it’s faster it’s faster.

I just tried to watch this hardware unboxed video and off the rip it’s so dumb. Why does anyone care about tiers if the don’t work for the company? You have a product. Look at the alternatives and check if they’re better and if it’s worth the cost. If you don’t like it don’t buy it. Skip a generation if there’s nothing in your price range that’s worth it. Who about a product in some price tier you wouldn’t buy anyway. The only relevant comparison is to what you already own.
Click to expand...
I love this take and agree completely.

Customers don't care if the bus width is 64 bits or 1024 bits, just like they don't care if the engine in their car is 200HP or 10,000HP. What they care about is how it performs, and does that performance meet their expectation in-line with the price they're being asked to pay. If the bus width is 64 bit and can play CP2077 at 4k native with pathtracing enabled at a fluid minimum of 120Hz? Well damn, 64 bits is looking really damned fine to me!!!

And the whole generation gap thing is just silly. If you're using an iPhone 15 Pro Max Quad Damage Megakill today, the iPhone 16 Pro Max Quad Damage Ultrakill might not be worth the >$1000 you're gonna spend on it. Buuuuuuut if you're coming from an iPhone 13 Sus Damage Moderate Injury model, then the modern version is gonna look a LOT better for the same price tag. I fully understand why new models are directly compared to the outgoing models, but that's also not (IMO) representative of the target audience.

So basically, again, agreeing with everything @Scott_Arm just said :D
 
trinibwoy said:
That reads less like a rumor and more like random speculation.
Click to expand...
The only interesting part there is about GB203's die size, the rest are just speculation and somewhat expected developments (I mean of course Nvidia will use 3Gb G7 modules once they'll become available; why would they be even made if not for Nvidia to use them? RDNA4 seem to omit G7 support).
 
Albuquerque said:
I love this take and agree completely.

Customers don't care if the bus width is 64 bits or 1024 bits, just like they don't care if the engine in their car is 200HP or 10,000HP. What they care about is how it performs, and does that performance meet their expectation in-line with the price they're being asked to pay. If the bus width is 64 bit and can play CP2077 at 4k native with pathtracing enabled at a fluid minimum of 120Hz? Well damn, 64 bits is looking really damned fine to me!!!

And the whole generation gap thing is just silly. If you're using an iPhone 15 Pro Max Quad Damage Megakill today, the iPhone 16 Pro Max Quad Damage Ultrakill might not be worth the >$1000 you're gonna spend on it. Buuuuuuut if you're coming from an iPhone 13 Sus Damage Moderate Injury model, then the modern version is gonna look a LOT better for the same price tag. I fully understand why new models are directly compared to the outgoing models, but that's also not (IMO) representative of the target audience.

So basically, again, agreeing with everything @Scott_Arm just said :D
Click to expand...

I honestly think this stuff is largely driven by FOMO or PCs as status symbols. Say I waited until 60x0 series to upgrade from my 3080. Well, maybe I look at the price and performance of a 6070 and it fits better into my budget and I get a performance boost that I think is acceptable. Why not? I don't care what model it is. I don't know why anyone does, other than personal feelings about not having the best. I'm currently looking at buying new speakers and there's no reality where I'd ever go back and do a historical analysis of price tiers and models to try to figure out if the model number of the speaker has changed relative to other models in a stack. It's just a completely absurd thing to do.
 
Last edited:
You must log in or register to reply here.
Back
Top