NVIDIA GeForce GTX 1650

The NVIDIA GeForce GTX 1650 is a dual-slot, end-of-life graphics card built on the Turing architecture using the TU117 chip, fabricated on TSMC's 12 nm process. It carries 4,700 million transistors on a 200 mm² die, resulting in a transistor density of 23.5M per mm². The card features a base clock of 1485 MHz and a boost clock of 1665 MHz, with 896 shading units, 56 texture mapping units, and 32 render output units. Its production status is listed as end-of-life, and it was released on April 22, 2019, with a launch MSRP of 149 USD.

Memory Subsystem

The GTX 1650 is equipped with 4 GB of GDDR5 memory on a 128-bit bus, delivering a bandwidth of 128.1 GB/s. The memory operates at 2001 MHz, translating to 8 Gbps effective. This configuration is modest by modern standards, but the data shows it is sufficient for 1080p gaming at medium to high settings in older titles. At higher resolutions, such as 1440p or 4K, the 4 GB capacity becomes a limiting factor, as texture-heavy scenes can exceed this allocation, causing stuttering or reduced texture detail. The 128.1 GB/s bandwidth is adequate for the card's compute capabilities, but it is not designed to feed a high-resolution, high-refresh-rate workload. The 128-bit bus width inherently caps peak throughput, meaning that the card will struggle with large data transfers in modern games at elevated resolutions. For users targeting 1080p, the memory subsystem is balanced with the GPU's raw compute power, but it does not offer headroom for future titles that demand more VRAM.

Ray Tracing and Feature Set

This card does not include dedicated ray tracing cores or tensor cores, as those are absent from the fact pack. Consequently, it lacks hardware-accelerated ray tracing and AI-based features like DLSS. The API support includes DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4, which ensures compatibility with modern game engines and graphics APIs. The absence of RT and tensor cores means that any ray tracing workload would fall back to compute shaders on the 896 shading units, which would severely impact performance. The card's feature set is therefore oriented toward traditional rasterization, with no path for hardware-accelerated ray tracing. The display outputs are 1x DVI, 1x HDMI 2.0, and 1x DisplayPort 1.4a, which covers standard monitor connectivity. The DirectX 12_1 support indicates it can handle feature level 12_1, which includes some advanced rasterization features, but not the ray tracing tier found in higher-end Turing cards.

Power and Cooling

The GTX 1650 has a thermal design power (TDP) of 75 W, which is low enough to be powered solely by the PCIe 3.0 x16 slot, as it has no power connectors. The suggested power supply unit is rated at 250 W, making it compatible with a wide range of entry-level systems. The card is a dual-slot design, with dimensions of 229 mm in length (9 inches), 111 mm in height (4.4 inches), and 35 mm in width (1.4 inches). Cooling is handled by a dual-slot cooler, which is more than sufficient for the 75 W TDP. The low power draw means that heat output is minimal, and the card should run quietly under typical gaming loads. The lack of external power connectors simplifies installation, but it also caps overclocking potential, as the power delivery is limited to the slot's 75 W supply. For users with a 250 W PSU or higher, this card is a drop-in solution that requires no additional cabling.

How It Compares

The nearest rival is the NVIDIA GeForce RTX 3050 A Mobile, which has an average benchmark score of 8746, while the GTX 1650 has an average score of 8713. The delta is -0.4%, indicating that the GTX 1650 is essentially tied with this mobile part. The RTX 3050 A Mobile is a laptop GPU, but the performance parity suggests that the GTX 1650's desktop implementation has similar computational throughput in synthetic tests. The 0.4% gap is negligible in real-world usage, meaning the two cards are interchangeable in terms of raw score.

The AMD Radeon 550X is another rival, with an average score of 8749 and a delta of -0.4% relative to the GTX 1650. Again, the performance difference is less than half a percent, making the two cards statistically equivalent in aggregate benchmarks. The Radeon 550X is an older, low-end part, and the data shows it matches the GTX 1650 in this specific benchmark suite. This indicates that the GTX 1650's Turing architecture does not provide a significant advantage over older AMD designs in compute-oriented tests.

The NVIDIA Quadro P2200 is a workstation card with an average score of 8671, which is 0.5% lower than the GTX 1650. The professional card is designed for stability and driver certification, but the raw performance is slightly below the consumer GTX 1650. This suggests that the GTX 1650 offers marginally better compute performance for general tasks, though the Quadro may have optimizations for specific professional applications that are not reflected in these scores.

The AMD Radeon Pro WX 5100 is the final rival, with an average score of 8761 and a delta of -0.5% relative to the GTX 1650. The Pro WX 5100 is a professional graphics card, and it edges out the GTX 1650 by half a percent in aggregate score. This is a marginal difference, but it indicates that the Radeon Pro WX 5100 has a slight computational edge in these tests. The GTX 1650's position is therefore tightly clustered with all four rivals, with deltas ranging from -0.5% to +0.5%.

Benchmark Performance

The GTX 1650's average benchmark score is 8713, placing it at the 43rd percentile among all GPUs. This means that it outperforms 43% of the graphics cards in the database, which is a mid-to-low position. The individual benchmark scores show a wide variance across different test types. In 3DMark Steel Nomad DX12, the card scores 305, which is a low absolute number but reflects the demanding nature of that test. The Geekbench OpenCL score is 39112, while the Geekbench Vulkan score is 35966, indicating that OpenCL performance is about 8.7% higher than Vulkan in this test suite. The PassMark DirectX 9 score is 124, which is high, but the DirectX 10 score drops to 39, and the DirectX 11 score is 58. The DirectX 12 score is 35, which is the lowest among the DirectX variants. This pattern suggests that the card's performance degrades as the API complexity increases, with DirectX 9 being the most favorable and DirectX 12 being the most punishing. The PassMark G2D score is 561, which measures 2D graphics performance, while the G3D score is 7880, which is the primary 3D metric. The PassMark GPU Compute score is 3048, indicating moderate compute throughput.

Compared to its nearest rivals, the GTX 1650 is effectively tied with all of them, with deltas of -0.4%, -0.4%, +0.5%, and -0.5% against the RTX 3050 A Mobile, Radeon 550X, Quadro P2200, and Radeon Pro WX 5100, respectively. The largest gap is 0.5%, which is within the margin of error for synthetic benchmarks. The data shows that the GTX 1650 occupies a performance tier that is indistinguishable from its immediate competition in aggregate scores. The 43rd percentile ranking reinforces this, as it is a mid-pack position with no clear advantage over the listed rivals. The card's compute performance, as measured by PassMark GPU Compute at 3048, is consistent with its FP32 throughput of 2.984 TFLOPS. The FP16 performance is 5.967 TFLOPS, which is exactly double the FP32 rate, indicating a 2:1 ratio. The pixel rate is 53.28 GPixel/s, and the texture rate is 93.24 GTexel/s, which are derived from the core clocks and ROP/TMU counts. These figures position the GTX 1650 as a basic entry-level card for 1080p gaming, but the benchmark data reveals it is not competitive with even the modest rivals listed, as all deltas are within half a percent. The 3DMark Steel Nomad score of 305 is particularly low, suggesting that modern DX12 workloads stress the card heavily, and the PassMark DirectX 10/11/12 scores (39, 58, 35) confirm that the card's performance varies significantly by API generation.