It’s been almost two years since NVIDIA launched the RTX 3000 series of graphics cards. This is the year they launch the sequel to one of the most influential and controversial gaming technologies.
Now the company is almost ready to unveil its next generation of GPUs, the RTX 4000 series.
NVIDIA confirmed during a conference call that it will unveil its new architecture for consumer graphics cards next month at the company’s GTC 2022 keynote on September 20, 2022. We might not see new graphics cards ready to buy that day, but we will take a look at the general architecture that will power upcoming GeForce GPUs.
NVIDIA also announced on the earnings call that demand for gaming hardware was down 33% year-over-year and 44% sequentially. The company said this was due to “challenging market conditions”. The global economy is experiencing some inflation right now, so probably fewer people are willing to spend hundreds of dollars on gaming hardware. However, demand has also dropped – crypto mining is currently less profitable (if at all), so crypto farms are not buying up all of the available supply like they were a few months ago.
Nvidia’s next-gen RTX 40-series graphics cards were originally slated for release in late 2022 – and given the two-year gap between the RTX 20-series and RTX 30-series, that shouldn’t come as a huge surprise.
In April 2022, it was hinted that the AD102 GPU, the GPU expected to power the top-end RTX 4090, had entered testing. If true, it’s a good sign that development of the next-gen lineup is progressing well, and barring unforeseen testing issues, the design of the GPU is likely complete. Interestingly, other leakers also seem to think that the RTX 4090 will launch the first and possibly only 4000 series in 2022. According to analysts, the reason is “too many shares”.
This is in reference to the rumor that Nvidia still has plenty of RTX 30 series GPU inventory due to overproduction and saturation of GPUs in the used market amid the cryptocurrency’s recent crash.
Rumor has it that the RTX 4090 will feature 126 streaming multiprocessors with a total of 16,128 CUDA cores and 24GB of GDDR6X memory clocked at 21Gbps
The RTX 4080 could offer up to 10,240 CUDA cores, 16GB of GDDR6X VRAM clocked at 21Gbps, a 4GB increase compared to the 12GB GDDRX-capable RTX 3080 Ti with an increased TDP of 420W.
The RTX 4070 could feature 7,168 CUDA cores, 10 or 12GB of slightly slower GDDR5 VRAM clocked at 18Gbps and a TDP of 300W.
This article was previously published on Source link