Early performance benchmarks for NVIDIA’s GeForce RTX 4070 Ti graphics card have leaked into the Geekbench 5 benchmark database.
NVIDIA GeForce RTX 4070 Ti settles on par with AMD RX 7900 XTX in leaked Geekbench 5 benchmark
The NVIDIA GeForce RTX 4070 Ti graphics card will be launched in early January and will feature the same specifications as the “Unlaunched” 12GB GeForce RTX 4080 graphics cards. Rumors suggest that the graphics card may feature the same US$899 price tag as the previous generation card but that is not clear yet.
Performance leaks on the Geekbench 5 OpenCL benchmark test showing the NVIDIA GeForce RTX 4070 Ti boost up to 2730MHz and based on the performance numbers, the card appears to be on par with the AMD Radeon RX 7900 XTX which costs $999 US. The card is about 5% faster than the RX 7900 XT in the same benchmark and if NVIDIA can deliver the same performance increase in AAA games, it would explain why NVIDIA wants to keep prices at $899. Compared to its predecessor, the RTX 4070 Ti is about 45-50% faster than the 3070 Ti which is a huge jump but again, we’ll have to wait and see if games also show a similar performance boost.
65980
131,960
197,940
263920
329,900
395,880
“Official” NVIDIA GeForce RTX 4070 Ti 12GB Specifications
The NVIDIA GeForce RTX 4070 Ti will use an “Ada Lovelace” AD104-400 GPU configuration with 7680 cores or 60 SMs enabled which is a full fat SKU. The GPU will pack 48MB of L2 cache and feature a 192-bit bus interface, which is why we get up to 12GB of GDDR6X capacities. The memory is said to run at 21Gbps for a memory bandwidth of 504GB/s. The card is said to be clocked at 2610MHz but the peak frequency will be higher.
- “Official” NVIDIA GeForce RTX 4070 Ti 12GB Graphics Card TBP – 285 watts
- NVIDIA GeForce RTX 3070 Ti 8GB TBP “Official” – 290 watts
In terms of power, the TBP is now set to be rated at 285W which is a 35W drop versus the RTX 3080 10GB model and a 5W drop from the RTX 3070 Ti. The new graphics card is expected to perform better than the RTX 3080, but looking at specs beyond the 4080 16GB, there will be a significant gap between both models of about 30-40%.
In terms of power, the TBP is now set to be rated at 285W which is a 35W drop versus the RTX 3080 10GB model and a 5W drop from the RTX 3070 Ti. The new graphics card is expected to perform better than the RTX 3080, but looking at specs beyond the 4080 16GB, there will be a significant gap between both models of about 30-40%.
NVIDIA GeForce RTX 40 Series Official Specifications:
| Graphics card name | NVIDIA GeForce RTX 4090 | NVIDIA GeForce RTX 4080 | NVIDIA GeForce RTX 4070 Ti graphics card |
|---|---|---|---|
| GPU name | Ada Lovelace AD102-300 | Ada Lovelace AD103-300 | Ada Lovelace AD104-400 |
| process node | TSMC4N | TSMC4N | TSMC4N |
| die size | 608 mm2 | 378.6 mm2 | 294.5 mm2 |
| transistors | 76 billion | 45.9 billion | 35.8 billion |
| CUDA cores | 16384 | 9728 | 7680 |
| TMUs/ROPs | 512/176 | 320/112 | 240/80 |
| Strained/RT cores | 512/128 | 304/76 | 240/60 |
| base clock | 2230 MHz | 2210 MHz | 2310 MHz |
| clock increase | 2520 MHz | 2510 MHz | 2610 MHz |
| FP32 account | 83 TFLOPs | 49 TFLOPs | 40 TFLOPs |
| RTFLOPs | 191 TFLOPs | 113 TFLOPs | 82 TFLOPs |
| Tensor-TOPs | 1321 Top | 780 peaks | 641 top |
| memory capacity | 24 GB GDDR6X | 16 GB GDDR6X | 12 GB GDDR6X |
| memory bus | 384 bits | 256 bits | 192 bits |
| memory speed | 21.0 Gbps | 23.0 Gbps | 21.0 Gbps |
| bandwidth | 1008 GB / s | 736 GB / s | 504 GB / s |
| TBP | 450 watts | 320 watts | 285 watts |
| Price (MSRP/FE) | $1,599 USD / 1,949 EU | $1,199 / $1,469 | to be announced later on |
| price (current) | $1,599 USD / 1,859 EC | $1199 / $1399 AU | to be announced later on |
| launch (availability) | October 12, 2022 | November 16, 2022 | January 5, 2023 |
news source: Seating