The GTX 1650 gets another variant.
GALAX launches TU106-based GeForce GTX 1650
Not all graphics cards are created equal. The fabrication yields of the processors may be high, but even then there is a percentile of the silicon that will not qualify for intended graphics card models.
NVIDIA sells the surplus of graphics processors that would otherwise be used for higher-end models to some partners, but as a lower-tier SKU. This is actually a very common practice for NVIDIA, especially for the mid-range GTX xx60 series.
The GALAX GTX 1650 Ultra is a new model based on TU106 GPU. It’s the same GPU that would be used for the original RTX 2070. However, the TU106-125 variant of the processor has -61.1% cores compared to TU106-400.
The Ultra is a sub-variant of the GTX 1650 GDDR6, as they share almost the same specs. These cards are not advertised by NVIDIA, but can easily be found in the drivers where such models are listed with Device IDs of a different GPUs.
The GTX 1650 Ultra from GALAX is a fourth variant of the GTX 1650 series. It shares the core count with two other 1650 models. In fact, it also has the same clock speed. The difference is in the TDP, which is now 90W. That’s 15W higher than the TU117 model, but also, 10W lower than the TU116-based SUPER variant.
| GALAX GTX 1650 GDDR6 4GB Ultra | |||
|---|---|---|---|
| Specs provided by VideoCardz.NET Database | |||
| GPU : | TU117-300 | Base Clock : | 1410 MHz |
| Cores : | 896 | Boost Clock : | 1590 MHz |
| TMUs : | 56 | Memory Clock : | 12 Gbps |
| ROPs : | 32 | Memory : | 4 GB GDDR6 128b |
| NVIDIA GeForce GTX 1650 Series | ||||
|---|---|---|---|---|
| VideoCardz.com | GTX 1650 SUPER | GTX 1650 GDDR6 TU106 | GTX 1650 GDDR6 | GTX 1650 |
| GPU | TU116-250 | TU106-125 | TU117-300 | TU117-300 |
| CUDA Cores | ||||
| Base Clock | ||||
| Boost Clock | ||||
| Memory Clock | ||||
| Memory | ||||
| Memory Bus | ||||
| TDP | ||||
GALAX website lists 1660 Ultra instead of 1650 Ultra and 6GB instead of 4GB in the specifications tab
Source: Galax via @momomo_us
