Red Hat Announces RHEL AI

[mSparks]

lets say its 15fps then you activate FSR4.0 and you have playable 30-40fps...

so its basically a RTX4060 but $1000
not feeling that tbh

https://m.youtube.com/watch?si=KatJN5Ip3URLDdBd&v=roFZeSxmDN8&feature=youtu.be

[qarium]

so its basically a RTX4060 but $1000
not feeling that tbh
https://m.youtube.com/watch?si=KatJN5Ip3URLDdBd&v=roFZeSxmDN8&feature=yo utu.be

the target performance is a 4070 not the 4060...

the old steam deck is 420-679,00€ so its not a surprise that a AMD Strix Halo version of the steam deck costs like 1000€

[mSparks]

the target performance is a 4070 not the 4060...

current 4080 perf will be no more than $300 in a few months, blackwell is already in production. I really wouldn't go for either right now, they are just shipping the last of the 4000 series now they are obsolete..
Maybe you missed the press release?

https://nvidianews.nvidia.com/news/nvidia-blackwell-platform-arrives-to-power-a-new-era-of-computing

[qarium]

current 4080 perf will be no more than $300 in a few months, blackwell is already in production. I really wouldn't go for either right now, they are just shipping the last of the 4000 series now they are obsolete..
Maybe you missed the press release?
https://nvidianews.nvidia.com/news/n...a-of-computing

Nvidia blackwell is an HPC AI accelerator ?... clearly not the gpu what gives you 4080 perf for 300 dollars...

really man whats wrong with you ?...

[mSparks]

Nvidia blackwell is an HPC AI accelerator ?... clearly not the gpu what gives you 4080 perf for 300 dollars...

really man whats wrong with you ?...

blackwell is the architecture that will be in the 5000 series launching end of the year.
claiming 250% increase in IPC.

ever since the 2000 series they step up all the way each series, so
5060 > 4080, 4060 > 3080, 3060 > 2080

[mrg666]

really man whats wrong with you ?...

That is a broken chatbot, I think.

[qarium]

blackwell is the architecture that will be in the 5000 series launching end of the year.
claiming 250% increase in IPC.
ever since the 2000 series they step up all the way each series, so
5060 > 4080, 4060 > 3080, 3060 > 2080

they claim this 250% for AI workload only... for 64bit floading point the performance of blackwell/5000 series is lower than the 4090 had...

keep in mind some games run in 64bit floading point and not in 32bit floading point means they will lose big in these games.

see Nvidia Hopper had 67 FP64 Teraflops and blackwell only has 45 FP64 Teraflops

according to news like this:

https://www-heise-de.translate.goog/news/Nvidias-neue-KI-Chips-Blackwell-GB200-und-schnelles-NVLink-9658475.html?_x_tr_sl=de&_x_tr_tl=en&_x_tr_hl=de&_x_tr_pto=wapp

the main feature of blackwell/5000 series is AI features like 4 bit floading point and 6bit floading point
something what is not used in games at all.

with these 4bit AI units you maybe can speed up AI upscalling like DLSS but not the game itself.

right now blackwell has zero gaming features and only AI features...

[qarium]

That is a broken chatbot, I think.

something is really wrong... blackwell sounds to me like a AI architecture it has many AI features like 4bit floading point and 6bit floading point...

but all articles i read shows blackwell has zero gaming features.

maybe you can speed up DLSS with the 4bit/6bit floading point units.

but the games itself do not use it... also many games use 64bit floading point and blackwell has lower 64bit floading point performance.

[mrg666]

something is really wrong... blackwell sounds to me like a AI architecture it has many AI features like 4bit floading point and 6bit floading point...

but all articles i read shows blackwell has zero gaming features.

maybe you can speed up DLSS with the 4bit/6bit floading point units.

but the games itself do not use it... also many games use 64bit floading point and blackwell has lower 64bit floading point performance.

Gaming market is saturating and It is getting harder to make much improvement in gaming performance without pushing the cost and power requirements. On the other hand, AI is an exponentially expanding market that offers big profits. I will not be surprised if both AMD and Nvidia come up with next generation chips that offer underwhelming gaming improvements but focused more on the AI capabilities.

[mSparks]

Gaming market is saturating and It is getting harder to make much improvement in gaming performance without pushing the cost and power requirements. On the other hand, AI is an exponentially expanding market that offers big profits. I will not be surprised if both AMD and Nvidia come up with next generation chips that offer underwhelming gaming improvements but focused more on the AI capabilities.

For gaming AI gives you generative upscaling, so 320x640 perf for a 4K display, the power and performance benefit claimed there 100% translates into gaming devices.
RT cores need RTX games, currently there are ~ 2 or 3 of them, only one so far built to make use of RTX as its core development focus (Alan Wake 2), that proved you can build a AAA blockbuster on a $50mil budget that takes $500mil without RTX.

RTX4080 "only" has 76 RT cores, all they have announced on blackwell so far is one Datacenter chip (GB200), chip specs for workstations and gaming will be released within the next 6 to 12 months once they know production yields and mobile gaming chips will drip out over the next 4 or 5 years.