NVIDIA GeForce 945M vs NVIDIA GeForce GTX 1650 Comparison

The NVIDIA GeForce 945M and the NVIDIA GeForce GTX 1650 are separated by more than just a generation gap. The data shows a decisive performance chasm, with the GTX 1650 delivering roughly 3.7 times the raw compute throughput of the older 945M in the single shared benchmark. While the 945M represents the tail end of the Maxwell mobile era, the GTX 1650 is an entry-level Turing desktop part, and the benchmark results reflect a complete architectural and performance overhaul rather than a simple incremental step.

Head-to-Head Benchmarks

The only directly comparable data point between the two GPUs is the Geekbench OpenCL test, and the results are lopsided. The GeForce 945M scores 8,099, while the GeForce GTX 1650 scores 29,629. This represents a delta of -72.7% for the 945M, meaning the GTX 1650 is approximately 266% faster in this compute-oriented workload. This is not a marginal victory; it is a generational obliteration.

Contextualizing the 945M’s score, it sits at the 42nd percentile of all GPUs, placing it in the lower-mid range of the database. Its nearest rivals include the NVIDIA GRID K2 (8,080, 0.2% slower), the AMD Radeon R9 M360 (8,129, 0.4% faster), and the NVIDIA GeForce GTX 950M (8,135, 0.4% faster). The 945M is essentially locked in a dead heat with these older mobile and workstation parts, confirming its position as a modest entry-level mobile solution from its era.

On the other side, the GTX 1650’s 29,629 OpenCL score is a massive outlier compared to its average benchmark score of 7,472. This discrepancy is explained by the fact that the GTX 1650 has a broader suite of tests, including DirectX and Vulkan workloads, which lower its average. Its nearest rivals, based on the average score, are the AMD Radeon HD 8850M (7,447, 0.3% slower), the Intel UHD Graphics 750 (7,441, 0.4% slower), and the Intel Arc A310 (7,550, 1% faster). The GTX 1650’s 40th percentile ranking is dragged down by these legacy and integrated graphics comparisons, but its raw OpenCL compute power is in a completely different league than the 945M.

In other benchmark suites where the GTX 1650 has data, it shows strength in varying workloads. It scores 7,880 in Passmark G3D, 3,048 in Passmark GPU Compute, and 561 in Passmark G2D. Its DirectX performance varies, with scores of 124 in DirectX 9, 58 in DirectX 11, 39 in DirectX 10, and 35 in DirectX 12. It also achieves 33,042 in Geekbench Vulkan and 305 in 3DMark Steel Nomad DX12. The 945M has no corresponding data for these tests, so a direct comparison is impossible, but the sheer number of tests the GTX 1650 can run successfully highlights its modern feature set.

The Verdict

The data is unambiguous: the GeForce GTX 1650 is the superior product. If you are choosing between these two, the GTX 1650 wins on every measurable front. Its OpenCL score is nearly four times higher, it has four times the memory, and it supports a newer DirectX feature level.

The GeForce 945M is an end-of-life product from 2015, built on a 28 nm process. Its 2 GB DDR3 memory and 28.80 GB/s bandwidth are severely limiting by modern standards. The GTX 1650, released in 2019, offers 4 GB of GDDR5 memory with 128.1 GB/s of bandwidth, a 4.4x increase in memory throughput. This alone makes the GTX 1650 viable for modern games and applications, while the 945M would struggle with anything beyond basic tasks.

However, the GTX 1650 is not without its own context. Its 40th percentile ranking and average benchmark score of 7,472, which places it near integrated graphics like the Intel UHD 750, suggest it is an entry-level part. It is not a high-performance card; it is a capable 1080p gaming solution for esports titles and older games. The 945M, with its 42nd percentile, is even lower, but its percentile is based on a single compute test. The GTX 1650’s percentile is more representative of its overall gaming capability.

For a user with a legacy laptop that only supports MXM modules, the 945M is what it is: a low-power, low-performance part. For anyone building a new system or upgrading a desktop, the GTX 1650 is the only logical choice, provided its launch MSRP of 149 USD fits the build. The performance difference justifies the architectural leap.

FAQ

Q: How much faster is the GTX 1650 than the 945M in OpenCL?

A: The GTX 1650 scores 29,629 in Geekbench OpenCL, which is 266% higher than the 945M’s score of 8,099. The delta between the two is -72.7% in favor of the GTX 1650.

Q: Which GPU has better memory bandwidth?

A: The GTX 1650 has significantly better memory bandwidth at 128.1 GB/s compared to the 945M’s 28.80 GB/s. The GTX 1650 uses 4 GB of GDDR5 memory, while the 945M uses 2 GB of DDR3.

Q: Are these GPUs comparable in terms of TDP?

A: Yes, both have a TDP of 75 W. However, the GTX 1650 is a dual-slot desktop card with a 250 W suggested PSU, while the 945M is an MXM module for portable devices and requires no power connectors.

Q: What is the DirectX support difference?

A: The GTX 1650 supports DirectX 12 (12_1), while the 945M supports DirectX 12 (11_0). This means the GTX 1650 supports the full DirectX 12 feature set, while the 945M only supports the 11_0 subset.

Q: Which GPU has higher clock speeds?

A: The GTX 1650 has a base clock of 1485 MHz and a boost clock of 1665 MHz. The 945M has a base clock of 928 MHz and a boost clock of 1020 MHz.

Q: How much memory does each GPU have?

A: The GTX 1650 has 4 GB of GDDR5 memory, while the 945M has 2 GB of DDR3 memory. The bus width is 128-bit for both.

Specification Differences

The two GPUs differ in almost every key specification. The GeForce GTX 1650 uses the TU117 chip built on a 12 nm process, while the GeForce 945M uses the GM107 chip on a 28 nm process. This process shrink allows the GTX 1650 to pack 4,700 million transistors into a 200 mm² die, compared to the 945M’s 1,870 million transistors on a 148 mm² die. The transistor density increases from 12.6M / mm² to 23.5M / mm².

Clock speeds are substantially higher on the GTX 1650, with a base of 1485 MHz and a boost of 1665 MHz, compared to the 945M’s 928 MHz base and 1020 MHz boost. Memory is another major divide: the GTX 1650 offers 4 GB of GDDR5 at 2001 MHz (8 Gbps effective), while the 945M offers 2 GB of DDR3 at 900 MHz (1800 Mbps effective). This results in bandwidth of 128.1 GB/s versus 28.80 GB/s.

The execution units also scale up. The GTX 1650 has 896 shading units, 56 TMUs, and 32 ROPs, while the 945M has 640 shading units, 40 TMUs, and 16 ROPs. The pixel rate increases from 16.32 GPixel/s to 53.28 GPixel/s, and the texture rate jumps from 40.80 GTexel/s to 93.24 GTexel/s. FP32 compute goes from 1,305.6 GFLOPS to 2.984 TFLOPS, and the GTX 1650 also supports FP16 at 5.967 TFLOPS (2:1), a feature the 945M lacks.

Physically, the GTX 1650 is a dual-slot card measuring 229 mm x 111 mm x 35 mm, while the 945M is an MXM module. The GTX 1650 interfaces via PCIe 3.0 x16 and has display outputs of 1x DVI, 1x HDMI 2.0, and 1x DisplayPort 1.4a. The 945M’s display outputs are listed as "Portable Device Dependent." The GTX 1650 also has a suggested PSU of 250 W, while the 945M requires none. The GTX 1650 was released on 2019-04-22, while the 945M was released on 2015-10-26.

Architecture Differences

The architectural divide is fundamental. The 945M is based on the Maxwell architecture, a design that predates modern features like asynchronous compute optimizations and variable rate shading. The GTX 1650 is based on Turing, which introduces a substantially newer feature set, including support for DirectX 12 (12_1) compared to the 945M’s DirectX 12 (11_0). This means the GTX 1650 can handle more advanced rendering techniques and is more future-proof for modern game APIs.

The process technology also differs significantly. The 945M is fabricated on TSMC’s 28 nm node, a mature but power-hungry process. The GTX 1650 uses TSMC’s 12 nm node, which is significantly more efficient and allows for higher clock speeds and transistor density. The 945M’s GM107 chip is a modest design, while the GTX 1650’s TU117 is a more complex processor with nearly 2.5 times the transistors.

Memory architecture is another major difference. The 945M uses DDR3 memory, which is slow and energy-inefficient for graphics tasks. The GTX 1650 uses GDDR5, which provides a massive bandwidth advantage. This is not just a speed difference; it changes the nature of the GPU’s performance profile, allowing the GTX 1650 to handle larger textures and higher resolutions. The GTX 1650 also has a newer bus interface (PCIe 3.0 x16) compared to the 945M’s MXM-B (3.0), which is a mobile-specific form factor.

The GTX 1650 also benefits from a more modern API stack, supporting Vulkan 1.4 and OpenGL 4.6, same as the 945M, but the underlying Turing architecture provides better driver optimization and feature support for newer games. The 945M’s Maxwell architecture is end-of-life, meaning it receives no new feature updates.

Where Each One Wins

The GeForce GTX 1650 wins in every single benchmark category where both have data. It is the clear victor in compute performance, with an OpenCL score of 29,629 versus the 945M’s 8,099. It also wins in memory capacity and bandwidth, offering 4 GB of GDDR5 at 128.1 GB/s versus 2 GB of DDR3 at 28.80 GB/s. This makes the GTX 1650 suitable for modern gaming at 1080p with medium to high settings, depending on the title.

The GTX 1650 is also the only one of the two with Vulkan benchmark data, scoring 33,042 in Geekbench Vulkan, and it has multiple Passmark scores, including a G3D score of 7,880 and a GPU Compute score of 3,048. These results indicate a well-rounded GPU capable of handling both graphics and compute tasks. Its TDP of 75 W is remarkably low for its performance level, making it an efficient choice for a desktop build.

The GeForce 945M has no benchmark wins. Its only advantage is its form factor. As an MXM module with no power connectors, it is designed for portable devices, specifically laptops. If you have a laptop that supports MXM upgrades, the 945M might be the only option available to you. In that context, it offers a modest improvement over older integrated graphics, but it is not competitive with any modern discrete GPU.

In terms of use cases, the GTX 1650 is for anyone building a budget desktop PC who wants a dedicated GPU for light gaming, media creation, or general compute acceleration. The 945M is for someone repairing or upgrading a legacy laptop, where its low power draw and MXM form factor are the only reasons to consider it. The data shows no scenario where the 945M outperforms the GTX 1650.