Every Super launch is better than the OG launch. Because NV needs sales the following year.Public opinion of the 40 series was a more favorable after 18 months because the prices dropped. But I think most people still say it was bad at launch.
Well, you know, I just think we can be level-headed on this one and resist the urge to rush to judgement regarding how this generation actually pans out. These things tend to start off kind of rocky and they usually end up working out well for most gamers. I have no reason to think the 50 series will be any different. People were a little skeptical about the 40 series at first, but after about that 18 month mark or so it seems like opinions of the series started to look more favorable. I think we might be in for a pleasant surprise with the 50 series if we can just have a little patience, and again, that's around that 18-24 month mark when they start to make their way into some of the retail channels, but these things are complex and do take some time so let's just try to be a little more mature and understanding right now.
It's gonna be like less than 12 months till 70 series, might as well wait for those...I’m waiting for the 60 super launch to upgrade, cause I know for a fact NV will bottleneck the vanilla 60 series in some way
70 Super is the one, guys and gals. It’s going to save PC Gaming and in 2032 you can quote me on this.It's gonna be like less than 12 months till 70 series, might as well wait for those...
The attitude is that Nvidia doesn't care about gamers, but wasn't the 40 super series actually a decent upgrade with decent availability? I think a ton of people bought 4070ti supers and especially 4070 supers at msrp or near msrp. I think we won't know how the 50 series will be remembered until about a year or two from now when some cards are available to purchase near msrp. Let's sit tight and wait another 18 months or so to see how things look after some cards make their way onto store shelves.
No because it's a 15% speed shrink (10% realistically for GPU). Area scaling isn't anything too good either.it would have been pretty good as well I’d reckon
no it's a major microarch failure. their 'new' client SM (sm1200) has two major regressions and dats it!They are competing against themselves and milking us gamers like cows.
The best 4080 was the 12GB model. Remember that one? Man Nvidia was really on their game for that one!The 4090 was a huge upgrade, one of their best generational increases ever. If they’d have used 3nm for Blackwell, it would have been pretty good as well I’d reckon. They are competing against themselves and milking us gamers like cows.
And they’re apparently trying to cook us over literal fires with Blackwell cards as well.
If we just wait for the new series to fully shake out and the market to settle, I think these cards will look pretty decent, all things considered. So again, it's likely right around that 18-36 month mark where things should start making more sense for the 50 series.The 4090 was a huge upgrade, one of their best generational increases ever. If they’d have used 3nm for Blackwell, it would have been pretty good as well I’d reckon. They are competing against themselves and milking us gamers like cows.
And they’re apparently trying to cook us over literal fires with Blackwell cards as well.
What? The 4090 was good, but hardly one of the best ever. It was ~50% faster than a 3090 Ti and maybe another 10% over the vanilla 3090. That's the same increase the 3090 Ti had over the 2080 Ti. 1080 Ti was twice as fast as a 980 Ti; 980 Ti was 40-50% faster than big Kepler (780 Ti) on the same node. 780 Ti was twice as fast as Fermi (580), and Fermi was again about twice as fast as Tesla. That's just the CUDA era though, go back further and the big steps were even bigger.The 4090 was a huge upgrade, one of their best generational increases ever. If they’d have used 3nm for Blackwell, it would have been pretty good as well I’d reckon. They are competing against themselves and milking us gamers like cows.
And they’re apparently trying to cook us over literal fires with Blackwell cards as well.
Earth is dying and Mars is our only hope. The entire world participates in a GPU based distributed computing effort to power the essential Mars Colony simulations. And oh, the seats are limited so only a few thousand of the top contributors to the Global Mars Colony Simulation Effort will be able to leave Earth.I'm working on a future news article "Silicon Valley startup founder found bludgeoned to death in alley" and would like to fill in some of the details such as what kind of visionary idea he had.
Given tariffs incoming, lack of PhysX support, little to no performance increase over 4000 Supers, lack of stock, and crazy prices, I'm holding onto my 4060, 4070 Ti Super, and 4090 for a good while.I wasn't going to touch the 4000 series but lack of PhysX on 5000 series, if not fixed soon, is going to force me to get one at some point (maybe two or three years later). Maybe from one of you good folks. Don't run them too hard, please.
Thinking I should have banked a couple of those $730 MSI 4070 Ti Supers. Which is ludicrous.Given tariffs incoming, lack of PhysX support, little to no performance increase over 4000 Supers, lack of stock, and crazy prices, I'm holding onto my 4060, 4070 Ti Super, and 4090 for a good while.
As an owner with a 4090, agreed!The 4090 was a huge upgrade, one of their best generational increases ever. If they’d have used 3nm for Blackwell, it would have been pretty good as well I’d reckon. They are competing against themselves and milking us gamers like cows.
And they’re apparently trying to cook us over literal fires with Blackwell cards as well.
I am sorry, but I have to hard disagree with you. My 4090FE was a huge leap over my 3090 in every metric. It is the fastest (by far), quietest, and most efficient card I've ever owned. Maybe AIB cards aren't so great? This is the first time I've bought an NVIDIA made card. I primarily bought from EVGA, and once, a long time ago, a ZOTAC card or two (GTX 275 and something else I think) Outside of that I used to be a Radeon guy before they fell behind.What? The 4090 was good, but hardly one of the best ever. It was ~50% faster than a 3090 Ti and maybe another 10% over the vanilla 3090. That's the same increase the 3090 Ti had over the 2080 Ti. 1080 Ti was twice as fast as a 980 Ti; 980 Ti was 40-50% faster than big Kepler (780 Ti) on the same node. 780 Ti was twice as fast as Fermi (580), and Fermi was again about twice as fast as Tesla. That's just the CUDA era though, go back further and the big steps were even bigger.
The only Ada and Blackwell are giant leaps forward are if you take the gaming applications of all the AI specific transistors they're filling the die with at face value and run with FG and MFG.
From what I understand, the 5090FE is even better. I've yet to play with one, but the fact it is a 2 slot card? Nobody else has managed that.
Fair enough. It's based on Techspot (HUB) and Techpowerup 4k launch results, both are in the 50-60% range. It hasn't changed a tonne though, even in the 5080 reviews with an updated 9800X3D platform it's listed as an average of 53% faster. I'm not saying it's a terrible upgrade, just that it's average or even a bit lower than average for a high end generational increase. Your 1080 Ti was a much larger step up from a 980 Ti than the 4090 over the 3090 Ti, and that's even before you look at the (almost) doubling of VRAM making that difference grow over time.I am sorry, but I have to hard disagree with you. My 4090FE was a huge leap over my 3090 in every metric. It is the fastest (by far), quietest, and most efficient card I've ever owned. Maybe AIB cards aren't so great? This is the first time I've bought an NVIDIA made card. I primarily bought from EVGA, and once, a long time ago, a ZOTAC card or two (GTX 275 and something else I think) Outside of that I used to be a Radeon guy before they fell behind.
From what I understand, the 5090FE is even better. I've yet to play with one, but the fact it is a 2 slot card? Nobody else has managed that.
I suspect you watch very specific reviews or something. FWIW I play a mix of old and new, indie games, AAA games, love setting things to the max, and live for Path Tracing/Ray Tracing because it looks amazing on my OLED monitor or my OLED TV. I'm also not usually concerned about budget because gaming is my hobby. Oh and I'll gladly pay for a 5090FE when the time comes.
Ada is the new Pascal. All it misses out is crappy MFG.Given tariffs incoming, lack of PhysX support, little to no performance increase over 4000 Supers, lack of stock, and crazy prices, I'm holding onto my 4060, 4070 Ti Super, and 4090 for a good while.
No way!!!This affects games like Borderlands 2.
My 4090FE was a huge leap over my 3090 in every metric
Thread with at least a partial list of PhysX titles affected by the shuttering. I wonder if this might be a case for getting a CPU with more cores to offload? Alternatively one might keep an older card in their system to run these older games.
RTX 50-Series GPUs have dropped support for 32-bit PhysX; many older PC games are impacted (Mirror's Edge, Borderlands, etc) PC
https://www.reddit.com/r/nvidia/comments/1irs8xk/rtx_50_series_silently_removed_32bit_physx_support/ A user on Reddit tested Borderlands 2 with a 9800X3D and RTX 5090 and they said this...www.resetera.com