Jensen Huang will have a keynote at Siggraph:
Last edited:
Nvidia shows signs in [2023]
- Thread starter pharma
- Start date
- Status
Looks like they might finally have a fix for the nVidia driver DPC latency problems, but it's still being tested and hasn't rolled out yet.
I'd never even heard of this. How does it manifest in the real world?
For the 99.99% of users it doesn't. It may cause issues for those who need real time response - this is mostly people working with audio production.
A recent RTX 4060 Techgage review indicates the card might generate low budget Creator interest in specific applications like Blender.
Wow first cousin’s daughter? That is extremely close. Didn’t realize they were real “family”.
DavidGraham
Veteran
In my opinion, this is what NVIDIA should've named their current cards, this would have made much more sense, with a vastly better general critical/public reception but would've made pricing much more difficult. Imagine a 4070 costing 800$, or a 4050 with a 300$ price tag.
RTX 4090 : RTX 4090
RTX 4080 : RTX 4080
RTX 4070 Ti : RTX 4070
RTX 4070 : RTX 4060 Ti
RTX 4060 Ti : RTX 4060
RTX 4060 : RTX 4050
RTX 4090 : RTX 4090
RTX 4080 : RTX 4080
RTX 4070 Ti : RTX 4070
RTX 4070 : RTX 4060 Ti
RTX 4060 Ti : RTX 4060
RTX 4060 : RTX 4050
arandomguy
Regular
But that usage case places an even higher emphasis on VRAM than gaming and the RTX 4060 comes in at a downgrade against the outgoing 3060 12GB. Anecdotally in creator circles just last gen the preference was for the 3060 12GB over the 3060ti even if they were the same price.
Names don't matter. Even if they would name 4090 a 4050 that wouldn't make any change to perf/price.
Agreed. I think much of the blame is due to not knowing competitor pre-launch price/performance in similar tiers.
Nvidia’s H100: Funny L2, and Tons of Bandwidth
GPUs started out as devices meant purely for graphics rendering, but their highly parallel nature made them attractive for certain compute tasks too. As the GPU compute scene grew over the past cou…
chipsandcheese.com
First time in a while I've agreed with you, but yeah it'd be much less awful if'n they followed your naming scheme. It would have better reflected the cards performance relative to previous generations.
I think I'd have been happier with those names even without a price change (although obviously lowered pricing in line with the names would have been better). At least then we would have a much clearer generational leap at each tier that aligbs better with previous generational uplifts and would just have to get used to the fact that each tier has gone up massively in price. It would seem more... honest, than the current scheme.
The only 'tier' in existence worth talking about is the pricing tier. The rest are just people being tools for marketing depts.
True but there are a whole lot more "tools for marketing" than informed consumers, especially in GPUs.
It's telling that you had to cherry pick one of the worst releases you could find from a very specific architecture release that had bloated die sizes due to still using an old(at the time) 28nm process which allowed them to keep costs down even on the relatively larger dies.
The 960 was widely considered to be a very disappointing part. It still technically sold well thanks to its popular inclusion in like 90% of prebuilts at the time, but it was not a part anyone informed would readily recommend, especially in the face of AMD's superior alternatives.
This might honestly be the only place on the internet where anybody is seriously trying to ignore the reality of what's going on in order to put in place some distorted view that Nvidia's pricing is actually just fine. It's just wild and really makes this place seem like it's not really the place to be to have informed discussion outside any deeper technical areas.
Surely actual performance is also relevant.
The way I see it there are a number of tiers we can categorise GPU's into such as:
Price
Performance (relative to each other, last gen, and the competition)
Power Draw (relative to the rest of the current lineup)
Chip size (relative to the rest of the current lineup)
And then naming scheme should be applied based on the above characteristics, ideally following historical trends as a way of informing the consumer.
Clearly the needle has moved this generation on Price vs the others, and if the genuine reason for that is increased costs then I'd prefer the IHVs were honest about that and just increasd the pricing at each naming tier.
Instead I feel they're trying to obfuscate the pricing increase by rearranging the naming scheme. Essentially trying to fool less informed customers into thinking they're getting a higher end GPU (relative to the current and last gen stacks) than they actually are.
Which is where perf/price comes from naturally.
People should really stop seeing various "tiers" in marketing namings and start looking at what the same price gets you - adjusting this for inflation maybe which is in fact quite sizeable when comparing modern GPUs to something like Pascal.
- Status
Similar threads
- Locked
