LaptopMedia apparently has a sample of the laptop-bound GTX 1050 Ti on-site, and provided the specs and benchmarks that compare the chip to other GeForce models for laptops. This isn’t unusual, as Nvidia has been known to hand out samples early for testing purposes. Either there’s no NDA involved (for obvious promotional reasons), or LaptopMedia is breaking the rules.
Regardless, here are the specs for the GTX 1050 Ti for laptops compared to the one provided for desktops:
| Laptop | Desktop | |
| Graphics core: | GP107/Pascal | GP107/Pascal |
| Base clock: | 1,490MHz | 1,290MHz |
| Boost clock: | 1,624MHz | 1,392MHz |
| CUDA cores: | 768 | 768 |
| Memory: | 4GB GDDR5 | 4GB GDDR5 |
| Memory bandwidth: | 112.1GB/s | 112.1GB/s |
| Memory interface: | 128-bit | 128-bit |
| Render output units: | 32 | 32 |
| Texture mapping units: | 64 | 64 |
| TDP: | Around 60 watts | 75 watts |
Did you catch the difference? The laptop version is clocked higher than the desktop model using the same number of cores. At first, that doesn’t make sense, and could indicate that the CUDA count number for the laptop model is incorrect. Either Nvidia may have injected the GTX 1050 TI for laptops with 640 CUDA cores, or the clock speeds are off.
For example, just take a look at the differences between laptop and desktop versions of the GTX 1060. The two 6GB models are nearly identical save for the lower clock speeds of the laptop unit to keep the thermals low. And even though the 3GB desktop model has higher clock speeds than the two 6GB units, it has a lower core count.
| Laptop | Desktop 3GB | Desktop 6GB | |
| Graphics core: | GP106/Pascal | GP106/Pascal | GP106/Pascal |
| Base clock: | 1,404MHz | 1,518MHz | 1,506MHz |
| Boost clock: | 1,670MHz | 1,733MHz | 1,708MHz |
| CUDA cores: | 1,280 | 1,152 | 1,280 |
| Memory: | 6GB GDDR5 | 3GB GDDR5 | 6GB GDDR5 |
| Memory bandwidth: | 192GB/s | 192GB/s | 192GB/s |
| Memory interface: | 192-bit | 192-bit | 192-bit |
| Render output units: | 48 | 48 | 48 |
| Texture mapping units: | 80 | 80 | 80 |
However, the GTX 1050 Ti may be clocked higher in the laptop due to the Mobile PCI Express Module (MXM). This handles the software interface, the thermal, electrical, and mechanical aspects of the connection between the main system and the graphics chip. In other words, the laptop version has an extra power boost whereas the desktop version draws power only through the PCIe bus, limiting the clock speeds.
In the case of the GTX 1060, the desktop version has a second 6-pin power connector providing a combined maximum power draw of 120 watts, enabling faster clock speeds than its laptop-based twin. The model supplied to notebook vendors uses around 80 watts to keep the overall thermals inside the device down to a “safe” limit.
Moving past the questionable clock speed / core count numbers, the site ran the GTX 1050 Ti laptop sample through three different tests to conclude that the GTX 1050 Ti aimed at laptops is around 10 percent faster than the GTX 970M, and more than 60 percent faster than the GTX 960M. This should be good news for PC gamers on a tight budget looking for a decent laptop to lose themselves in Fallout 4 with while seated on a plane.
Here are the numbers:
| 3DMark Cloud Gate Graphics |
3DMark Fire Strike Graphics |
Unigine Heaven 4 | |
| GTX 1050 Ti Laptop (sample) |
49976 (+10%) | 7757 (+7%) | 1836 (+9%) |
| GTX 970M (Asus GL502) |
45541 | 7271 | 1691 |
| GTX 965M (HP Omen 2016) |
38125 | 6063 | 1260 |
| GTX 960M (Lenovo Y700) |
31097 | 4451 | 989 |
As for a benchmark we can actually see, the site posted a video of Doom running at 1,920 x 1,080 using the Ultra settings and OpenGL (shown above). The framerate averages between 50 to 60 frames per second, which is great for a laptop. Then again, we don’t know the other laptop specs that help push the game’s framerate. We’ll find out the real numbers once this GPU rolls out in laptop solutions possibly in the beginning of 2017.
