Details on the future launch of the Nvidia GeForce GTX 1630indicating that it will be silicon-based Turing TU117-150manufactured by TSMC at 12nm, which will offer a configuration of 512 CUDA coresalmost half compared to the 896 CUDA cores seen in the GeForce GTX 1650 (the simplest variant), which would reach a Turbo frequency of 1800MHz accompanied by 4 GB of GDDR6 memory @ 12 Gbps, faster than the 8 Gbps chips used by the GeForce GTX 1650, but of course it lacks in the memory interface, which is 64 bit (vs 128), which gives a bandwidth of 96 GB/s (vs 128) keeping the same TDP of 75W due to increased frequencies (vs 1590 MHz).

Nvidia GeForce GTX 1630

The information known to Nvidia GeForce GTX 1630 is finished knowing that it would go on sale the next May 31stso it could be announced next week at Computex.

It should be remembered that this GPU was born for the sole purpose of not leaving the AMD Radeon RX6400 in a free will on the entry-level segment, since its rival for the price would be an Nvidia GeForce GTX 1050, which is much less powerful, so it makes sense that this GPU will eventually be replaced, and that’s thanks mining and all things rarity, this GPU has managed to stay on the market despite its debut 6 years ago.

Model GTX-1630 GTX1650
GPUs TU117-150 TU117-300
CUDA Cores 512 896
Freq. turbo 1800MHz 1590MHz
Freq. Memory 12 Gbps 8 Gbit/s / 12 Gbit/s
Memory 4 GB of GDDR6 memory 4 GB of GDDR5 / GDDR6
memory interface 64 bit 128 bit
Bandwidth 96 GB/s 128 GB/s / 192 GB/s
PDT 75W 75W/90W

via: video cardz