NVIDIA GeForce GTX 1050 Ti

The NVIDIA GeForce GTX 1050 Ti is a Pascal-generation part built on the GP107 chip at Samsung's 14 nm process, packing 3,300 million transistors into a 132 mm² die. It ships with 768 shading units, 48 texture mapping units, and 32 ROPs, running at a base clock of 1291 MHz and a boost clock of 1392 MHz. Memory is 4 GB of GDDR5 across a 128-bit bus, yielding 112.1 GB/s of bandwidth. The card was released on 2016-10-24, is now end-of-life, and had a launch MSRP of 139 USD. Its average benchmark score is 5316, which places it at the 30th percentile of all GPUs in the database—firmly in entry-level territory.

Benchmark Performance

The GTX 1050 Ti’s average benchmark score of 5316 puts it just above the 30th percentile, meaning it outperforms only about a third of the GPUs tracked in the database. That is a low standing, but the card’s individual test results show a mixed profile across different workloads. In 3DMark Steel Nomad DX12, it scores 305, a modest figure that reflects limited modern API performance. Geekbench results are more robust: 7834 in Metal, 20638 in OpenCL, and 19851 in Vulkan. These compute-oriented scores indicate that the card can handle general-purpose workloads reasonably well, though they are still far from competitive with even mid-range parts.

PassMark tests reveal a telling pattern. The DirectX 9 score of 104 is by far the highest among the PassMark DirectX tests, while DirectX 11 drops to 45 and DirectX 12 to 26. This suggests that the card’s architecture is better suited to legacy APIs, and that its efficiency degrades sharply as API complexity increases. The PassMark G3D score of 6340 and G2D score of 651 are moderate, while the GPU compute score of 2652 indicates limited parallel compute capability. The FP32 throughput is 2.138 TFLOPS, and FP16 is just 33.41 GFLOPS (1:64), so half-precision compute is extremely weak—a factor that will hinder any workload relying on FP16 acceleration.

When placed against its nearest rivals, the 1050 Ti is essentially tied with all of them. Its average score of 5316 is within 0.8% of every competitor listed. The deltaPct values are 0% versus the GeForce 930A, 0.4% faster than the GeForce 940M, 0.7% faster than the Radeon R7 240, and 0.8% slower than the Radeon R7 M445. These differences are negligible in real-world terms, meaning the 1050 Ti does not offer any meaningful performance advantage over its closest peers. It simply sits in a cluster of very similar low-end parts, with the percentile rank confirming its place near the bottom of the overall performance ladder.

How It Compares

NVIDIA GeForce 930A – The 930A has an average score of 5317, which is a 0% delta from the 1050 Ti. The two cards are effectively identical in performance, with a difference of just one point in the aggregate benchmark. This is a direct parity situation; neither card holds an edge over the other.

NVIDIA GeForce 940M – The 940M scores 5297 on average, putting the 1050 Ti 0.4% ahead. That is a hair-thin margin, far below the threshold where any game would show a perceptible difference. The 940M is a mobile part, while the 1050 Ti is desktop, but the benchmark scores indicate they perform at the same level.

AMD Radeon R7 240 – The R7 240 averages 5280, and the 1050 Ti is 0.7% faster. Again, this is a negligible advantage. The R7 240 is an older, lower-end AMD part, and the 1050 Ti barely edges it out in aggregate performance. In practice, users would see no difference in frame rates.

AMD Radeon R7 M445 – The R7 M445 is the only rival that beats the 1050 Ti, with an average score of 5361. That represents a 0.8% deficit for the 1050 Ti. While technically slower, the gap is so small that it would not affect gameplay. The R7 M445 is a mobile GPU, so this comparison is between a desktop card and a laptop part, but the performance parity is clear.

Memory Subsystem

The GTX 1050 Ti comes with 4 GB of GDDR5 memory on a 128-bit bus, providing 112.1 GB/s of bandwidth. The memory clock is 1752 MHz, which translates to 7 Gbps effective. This configuration is adequate for 1080p gaming at low-to-medium settings in most titles, but it becomes a bottleneck at higher resolutions. At 1440p or 4K, the 4 GB capacity and 112.1 GB/s bandwidth are insufficient for modern textures and large scenes. The 128-bit bus is narrow, and the bandwidth is low by today’s standards, so memory-intensive workloads will quickly saturate the interface. For 1080p, however, the 4 GB pool is enough for many games, especially those from the era when the card was current. The pixel rate of 44.54 GPixel/s and texture rate of 66.82 GTexel/s further indicate that the card is designed for lower resolutions rather than high-fidelity rendering.

Who Should Consider It

Given its 30th percentile ranking and the benchmark scores, the GTX 1050 Ti is best suited for 1080p gaming with modest settings. The PassMark DirectX 9 score of 104 is comparatively strong, suggesting that older titles—particularly those using DX9—will run smoothly. DirectX 11 games at 1080p with medium settings are also feasible, as the DX11 score of 45 is moderate. However, the DirectX 12 score of 26 is low, meaning modern DX12 titles will require reduced resolution or significant graphical compromises. The 3DMark Steel Nomad DX12 score of 305 reinforces this: the card struggles with current-generation APIs.

For users who primarily play esports titles, older games, or indie releases, the 1050 Ti can deliver playable frame rates at 1080p. Those wanting to play newer AAA games will need to lower settings to low or medium and possibly drop to 720p to maintain smooth performance. The 4 GB VRAM is sufficient for 1080p textures, but not for high-resolution texture packs or 1440p. The card is also a candidate for budget HTPCs or secondary machines, given its 75 W TDP and lack of power connectors, which simplifies installation. But anyone expecting modern high-fidelity gaming should look elsewhere.

Ray Tracing and Feature Set

The GTX 1050 Ti has no ray tracing cores and no tensor cores, as indicated by the null values in its specifications. This means it cannot perform hardware-accelerated ray tracing or leverage deep learning super sampling (DLSS). Any ray-traced effects would have to be handled by the shader units, which are already limited in compute power. The card does support DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4, giving it access to modern API features, but without the dedicated hardware for advanced effects. The FP16 throughput of 33.41 GFLOPS is minuscule, so mixed-precision workloads are not viable. Display outputs include one DVI, one HDMI 2.0, and one DisplayPort 1.4a, which covers standard multi-monitor setups. The card’s feature set is strictly baseline—no ray tracing, no AI acceleration, and only modest API support. For users who prioritize ray tracing or future-proofing, this card is not a consideration.