• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

NVIDIA 4000 Series

I have a LGA 1200 Socket Motherboard with an 11900k CPU. According to benchmark tests a Nivida RTX 4090 would be bottlenecked by a 11900k more then 20% running at stock speed however I was wonder whether overclocking the CPU and member would help in improving the performance of the 4090?
I think it depends on your resolution, at 4k it would be largely ok although it will still bottleneck some titles, if your running 1440p or 4k with DLSS then that would bottleneck a lot more.

If you use the MSI afterburner OSD you can see the percentage of the GPU being utilised so this lets you know how much bottleneck you have by how far under the 99% utilisation you are.

You can use DLDSR to remove bottleneck, this won’t improve FPS but will improve image quality so atleast you can leverage your gpu to the full.
 
Last edited:
"US Commerce Secretary Says Any AI Chips Designed To Circumvent Restrictions On China Will Be Banned The Very Next Day"

"I know there are CEOs of chip companies in this audience who are a little cranky with me because you’re losing revenue. Such is life. Protecting our national security matters more than short-term revenue. It's time to pick a side"


 
"US Commerce Secretary Says Any AI Chips Designed To Circumvent Restrictions On China Will Be Banned The Very Next Day"

"I know there are CEOs of chip companies in this audience who are a little cranky with me because you’re losing revenue. Such is life. Protecting our national security matters more than short-term revenue. It's time to pick a side"


I was wondered what RTX 4090D spec will be.

Maybe same spec as RTX 4090 but tensor core laser disabled?

If RTX 4090D will still have tensor core enabled or some cores reduced or disabled and it will be banned the very next day.
 
"US Commerce Secretary Says Any AI Chips Designed To Circumvent Restrictions On China Will Be Banned The Very Next Day"

"I know there are CEOs of chip companies in this audience who are a little cranky with me because you’re losing revenue. Such is life. Protecting our national security matters more than short-term revenue. It's time to pick a side"



Be interesting to see how this plays out long term.

The US is really stamping down hard. No more 4090D, E, small pepe then :cry:.
 
Introducing the brand new exciting "it just works a bit worse" Nvidia gives you the "D".

upsCZ1S.jpg
 
Last edited:
Nvidia delivers astonishing 500% AI performance increase with just a software update

Nvidia hopper GPUs process large language models 5 times faster than before, pre-emptively crushing any hopes AMD had with its new GPU launch

 
Last edited:
Nvidia delivers astonishing 500% AI performance increase with just a software update

Nvidia hopper GPUs process large language models 5 times faster than before, pre-emptively crushing any hopes AMD had with its new GPU launch


So that one will cost an arm and a leg... and also the other arm and other leg if nVidia are the only player in town at that level of performance.
 

NVIDIA preparing GeForce RTX 3050 with 6GB VRAM, 8GB model to be phased out​




:cry: Nvidia... The more you buy ... the less you get ...

Seems to be hurting their 4060 sales which really is a x50 card... under a new/misleading name...
 
Surprised they don't just axe the 3050.

They seem to sell like crazy, guessing business and new/midrange gamers like them. Seen a lot in mini/small builds and HTPC setups. Now Nvidia has decided they want to cripple them a little and make their 4060 cards look more appealing but at £80-100 more for similar model.

So if anyone is after an 8GB 3050, seems now is the time to grab one before they all end up with 6GB... :rolleyes:
 
So if anyone is after an 8GB 3050, seems now is the time to grab one before they all end up with 6GB... :rolleyes:
The cynic in me thinks this is all just a clever scheme by Nvidia to hike prices of 8GB 3050s and sell more of them.

Since their Chinese 4090 plans are scuppered, now they're looking elsewhere to make a quick buck.
 
Nvidia delivers astonishing 500% AI performance increase with just a software update

Nvidia hopper GPUs process large language models 5 times faster than before, pre-emptively crushing any hopes AMD had with its new GPU launch

Would tie in with the rumour that Nvidia was pulling people off the gaming driver team a while ago to work on AI, Nvidia’s GRD drivers lately have been buggy and issues don’t seem to be getting fixed in a hurry.
 

NVIDIA preparing GeForce RTX 3050 with 6GB VRAM, 8GB model to be phased out​


:cry: Nvidia... The more you buy ... the less you get ...

Seems to be hurting their 4060 sales which really is a x50 card... under a new/misleading name...
Making a graphics card worse, and continuing to sell it under the same name? Classic Nvidia.
 
Nvidia delivers astonishing 500% AI performance increase with just a software update

Nvidia hopper GPUs process large language models 5 times faster than before, pre-emptively crushing any hopes AMD had with its new GPU launch


And AMD will probably sell most of them fine:
  • Meta, OpenAI, and Microsoft said they will use AMD’s newest AI chip, the Instinct MI300X — a sign that tech companies want alternatives to the expensive Nvidia graphics processors that have been essential for artificial intelligence.
  • If the MI300X is good enough and inexpensive enough when it starts shipping early next year, it could lower costs for developing AI models.
  • AMD CEO Lisa Su projected the market for AI chips will amount to $400 billion or more in 2027, and she said she hopes AMD has a sizable part of that market.

All of these companies have huge software teams,so they will be doing their own software development too.

You seem to forget AMD also makes CPUs too,and is probably aiming to sell these as entire packages for supercomputers. The fastest supercomputer in the world,ie,Frontier,uses AMD CPUs and MI250 dGPUs. That is reinforced by the fact the MI300A/MI300X APUs are socket compatible with Epyc so existing customers can slot one in. That is what they are probably aiming to do,so they are not fighting Nvidia directly.
 
Last edited:
  • Like
Reactions: TNA
Back
Top Bottom