The dropper launch of the NVIDIA Lovelace architecture graphics card generation, the RTX 40, will have a new representative in May and a new chip. We talk about the RTX4060Ti which will equip the AD106 chip, the most modest of the new generation in its complete configuration. Can we expect a worthy successor to the RTX 3060 Ti?
While the RTX 4060 seems to have just disappeared from the market and the RTX 4070 has its release date in a few weeks, April 13 this year. As for the RTX 4060 Ti, we will see it in stores at some point in May and we can also talk about its technical specifications.
Will it be the RTX 4060 Ti, finally an NVIDIA Lovelace within everyone’s reach?
If there’s anything that’s characterized the new RTX 40 line of GPUs it’s its high price compared to previous generations, it’s because a good chunk of RTX 30 inventory is still in store and NVIDIA decided to mark their price based on the performance they offer compared to their predecessors. In any case, what is striking with the RTX 4060 Ti compared to its predecessor is that while the RTX 3060 Ti was a cut version of the RTX 3070, here we find a chip that is a generation below .
Regarding its technical specifications, they are as follows:
- Chip AD106-350:
- Base clock speed: 2310MHz
- Boost: Depends on the TDP of the graphics card, varies between 2535MHzy los 2745MHz.
- 34 SM cores with 4352 total CUDA disks.
- 34 RT cores third generation
- 136 tensor cores
- GDDR6 memory with a 128-bit bus at 288 GB/s
- TDP close to 160W
- makes use of 8 PCI Express lanes fourth generation,
How does it compare to the RTX 3060 Ti?
It’s a step down from the RTX 3060 Ti in some ways, especially in terms of memory bandwidth, although NVIDIA has increased the L2 cache to 32MB. We’re talking about going from a 256-bit bus with a band bandwidth of 448 GB/s at a bus that is almost half that. The other issue is the core count, the RTX 4060 Ti has 4 cores less than its rival, but like the rest of the RTX 40 line it makes up for it with a higher clock speed.
RTX-3060Ti | RTX4060Ti | |
---|---|---|
filling rate | 133.2 Gigapixels/s | 121.7 Gigapixels/s |
texturing rate | 253.1 Gigatexeles/s | 324.5 gigatexeles/s |
TFLOPS (FP32) | 16.2 | 22.06 |
As you can see the case repeats itself with the other graphics cards in the range compared to its predecessors, its higher clock speed compensates for the lower core count. However, most affected is its fill rate, which is limited by the lower video memory bandwidth.
As for its final price, at the moment we do not know it, although if we had to make a bet it could be around 500 dollars, given its performance compared to the previous generation of NVIDIA graphics cards. Too expensive? Well, we’re talking about RTX 4050 with AD107 chip and 6 GB of memory as the lowest point of the RTX 40. As we already said at the time, it seems that we will not see graphics cards below 300 dollars. Or not, who knows?