Recent content by troyan

Na, Ampere is much better: https://www.hardwareluxx.de/index.php/artikel/hardware/grafikkarten/54610-endlich-wieder-ein-duell-auf-augenhoehe-radeon-rx-6800-und-radeon-rx-6800-xt-im-test.html?start=15 RTX3070 and 6800XT have both 20 TFLOPs. Games are not limited by compute performance.

"Half the compute"? 3.1 TFLOPs is 25% less.

nVidia put the GPU part of PhysX under open source: https://github.com/NVIDIA-Omniverse/PhysX/discussions/384

30mins shorter than the OG Switch. I think this is a great result for the device. And it isnt just a better SoC and more on board memory. The display is much better, the internal speed of the memory is 3x better, better Joy Cons, better sound etc. And power consumption will be more like 1/6 of a...

Metroid is nativ 4K. Sega said that Yakuza 0 will be nativ 4K/60 FPS, too.

The Switch 2 is a great example how all this talk about the process technology is nonsense. With 8nm the SoC delivers performance very close to the PS4 Pro with ~1/2 of the power consumtion.. The PS5 has only ~3x higher compute performance with a 300mm^2 5nm SoC while being less advanced than...

It is not their job. It should be in the specification and mandatory. Reminds me of OLEDs which can burn in. Nothing will ever pretend it. So in the end it is the end user who has to be careful.

They didnt cheap out. It is just a connector. Its like blaming a hair dryer company for not safe guarding when someone is throwing it into a full bathtube... The problem is that this connector needs more space and bending can result in loosening the connection:

"Better"? It run in 480p with 30 FPS on a xbox 1. Here is a video with a Geforce 3 TI 200: Cant find any benchmarks from 20 years ago. Battlefield 5 with Raytracing ultra runs with 60 FPS on a 2080 TI in 1080p.

How do you think Morrowind with DX8 run on a GF3?

Battlefield 5 with Raytracing was released two months after Turing. I cant remember which was the second game with pixelshader support in 2001. Most games had been released after the Xbox 1.

It means alot when a 750mm^2 die is barely faster than a 350mm^2 from the competition despite having >2x the raw performance. For nVidia it is currently a waste of wafers to produce GB202. They can not optimize software like with AI. Engine developers dont care either. Gamers will have less...

The 5090 is only 2.2x faster than the 6900XT. GB202 is a huge waste of wafers when these engines are in such a state. These compute solutions are so inefficient. Instead of doing a proper raytracing implementation it looks like we are back in 2020.

When your AI find this, i think it makes sense to not invest in wafer for gaming chips: https://www.techpowerup.com/review/assassin-s-creed-shadows-performance-benchmark/5.html why build the biggest chips when game developers will find the most inefficient way to render a frame.

nVidia announced new workstation systems with Grace and Blackwell Ultra: https://www.nvidia.com/en-us/products/workstations/dgx-station/ 288GB HBM3e and 500GB from Grace.