Please note that this post is tagged as a rumor.
NVIDIA RTX 4090 Ti back in the rumor mill
NVIDIA is allegedly putting aside the best AD102 chips for the new flagship model.
With today’s release of RTX 4090, the focus will now shift towards the upgraded Ti model, which had been rumored for many months already. NVIDIA has indeed left a lot of headroom on the AD102 GPU for additional cores, and even kept the board design and cooling solution ready for something even more powerful.
The RTX 4090 Ti is rumored to feature 18176 CUDA cores and 24GB of GDDR6X memory, except the memory would be faster at 24 Gbps. This means that the maximum bandwidth would again increase from 1 TB/s to 1.1 TB/s.
According to XpeaGPU, who has been teasing information on the flagship Ada GPU in the past, NVIDIA is already saving chips for the RTX 4090 Ti variant which would offer 10 to 20% better performance:
Best dies are already saved for a 2.75Ghz boost (2.95 typ. gaming) 475W 18176 CUDA cores 96MB L2 cache 24Gb beast. Performance 10~20% above 4090 pic.twitter.com/z5XO8R9PyJ
— AGF (@XpeaGPU) October 12, 2022
Furthermore, is it claimed that NVIDIA AD102 GPUs for this model would offer a boost at 2.75 GHz, so at least 200 MHz higher than 4090. The actual in-game frequency could reach 2.95 GHz, and this basically means 107 TFLOPS of single-precision compute performance out of the box. This is a similar frequency to RTX 4090, except when it is overclocked.
The rumored RTX 4090 Ti GPU would obviously not launch very soon. NVIDIA is most likely waiting for a response to its new Ada GPUs from AMD, who are now expected to launch their RDNA3 GPUs next month.
NVIDIA GeForce RTX 40 Series Specs | ||||
---|---|---|---|---|
VideoCardz.com | GeForce RTX 4090 Ti | GeForce RTX 4090 | GeForce RTX 4080 16GB | GeForce RTX 4080 12GB |
Board-SKU | PG136-TBC | PG136-SKU330 | PG136-SKU360 | PG141-SKU331 |
Architecture | Ada (TSMC 4N) | Ada (TSMC 4N) | Ada (TSMC 4N) | Ada (TSMC 4N) |
GPU | AD102 | AD102-300 | AD103-300 | AD104-400 |
SMs | ||||
CUDA Cores | ||||
Boost Clock | ||||
Memory | ||||
Memory Bus | ||||
Memory Speed | ||||
Bandwidth | ||||
Default TGP | ||||
MSRP | 💰 |
Source: @XpeaGPU