NVIDIA GeForce GTX 660 vs NVIDIA GeForce RTX 3050 A Mobile Comparison

The NVIDIA GeForce GTX 660 and the NVIDIA GeForce RTX 3050 A Mobile represent two distinct eras of GPU design, separated by over a decade of architectural evolution. The data reveals a stark contrast in compute capability, yet the story is more nuanced than a simple generational leap. The GTX 660, a desktop card from 2012, and the RTX 3050 A Mobile, a laptop part from late 2023, offer different strengths that defy a single winner across all use cases.

Head-to-Head Benchmarks

The only direct benchmark comparison available in the data is the Geekbench OpenCL test, and the results are decisive. The RTX 3050 A Mobile scores 52,998, a massive jump from the GTX 660’s 11,347. The deltaPct of -78.6% (from the perspective of the GTX 660) indicates that the GTX 660’s score is a fraction of the newer card’s output. This translates to the RTX 3050 A Mobile delivering approximately 4.7 times the raw compute performance in this specific test.

This single data point is powerful, but it represents just one workload. The GTX 660 does not win any of the head-to-head benchmarks, with winsA at 0, while the RTX 3050 A Mobile takes a single victory, with winsB at 1. However, the broader benchmark suites available paint a more complex picture. The RTX 3050 A Mobile’s average benchmark score is 8,746, which places it in the 44th percentile of all GPUs. Its nearest rival is the NVIDIA GeForce GTX 460 v2, with an average score of 8,743 and a deltaPct of 0, meaning they are statistically identical in overall performance.

The GTX 660, with an average benchmark score of 9,022, sits in the 45th percentile. Its closest competitor is the NVIDIA TITAN V CEO Edition, which scores 9,037, a deltaPct of -0.2%. This means the GTX 660 is essentially on par with a TITAN V CEO Edition in these aggregate benchmarks, a surprising result given the TITAN’s reputation. The GTX 660 also edges out the AMD Radeon 550X (8,918, deltaPct 1.2%) and the AMD Radeon Pro WX 5100 (8,863, deltaPct 1.8%). While the RTX 3050 A Mobile wins the OpenCL test hands down, its overall average score is actually 3.2% lower than the GTX 660’s average, a discrepancy driven by its specific benchmark results across different test suites.

Where Each One Wins

The data suggests a clear split in workloads. The RTX 3050 A Mobile is the undisputed champion in compute-heavy tasks. Its Geekbench OpenCL score of 52,998 dwarfs the GTX 660’s 11,347, indicating a substantial advantage in applications that leverage general-purpose GPU computing, such as rendering, physics simulations, or machine learning inference. The RTX 3050 A Mobile’s Passmark G3D score of 11,664 further supports its strength in modern 3D graphics rendering.

Conversely, the GTX 660 demonstrates resilience in legacy and specific workloads. Its Geekbench Vulkan score of 11,415 and Geekbench Metal score of 4,305 are not directly comparable to the RTX 3050 A Mobile’s data, but they show the older card is not a slouch. The GTX 660’s higher aggregate average score (9,022 vs 8,746) suggests it may perform relatively better in certain older DirectX or OpenGL scenarios where its architecture is more efficient. The GTX 660 also wins on the basis of its transistor density advantage, but that’s a spec detail, not a benchmark. The RTX 3050 A Mobile’s low Passmark DirectX 10 and DirectX 9 scores (61 and 152, respectively) are curious; they are far lower than its DirectX 11 score of 94 and DirectX 12 score of 55, hinting at driver optimizations favoring newer APIs. The GTX 660’s architecture, while older, may have more mature drivers for those legacy APIs.

Architecture Differences

The two GPUs are built on fundamentally different architectures. The GTX 660 is based on the Kepler architecture, using the GK106 chip, while the RTX 3050 A Mobile uses the Ampere architecture with the GA106 chip. This generational gap is evident in nearly every design aspect.

The manufacturing process is a key differentiator. The GTX 660 is fabricated on a 28 nm process at TSMC, while the RTX 3050 A Mobile uses an 8 nm process at Samsung. This node shrink allows the newer chip to pack 12,000 million transistors into a 276 mm² die, achieving a transistor density of 43.5M / mm². In contrast, the GTX 660’s GK106 chip contains 2,540 million transistors on a 221 mm² die, with a density of just 11.5M / mm². The RTX 3050 A Mobile is nearly four times denser, a clear sign of architectural and process advancements.

Functionally, the RTX 3050 A Mobile introduces hardware features that the GTX 660 completely lacks. It includes 14 RT cores and 56 tensor cores, enabling hardware-accelerated ray tracing and AI-based features like DLSS. The GTX 660 has neither of these, relying purely on traditional rasterization. The RTX 3050 A Mobile also supports DirectX 12 Ultimate (12_2), while the GTX 660 is limited to DirectX 12 (11_0). Both support OpenGL 4.6, but the RTX 3050 A Mobile has a newer Vulkan version (1.4) compared to the GTX 660’s 1.2.175.

Specification Differences

Beyond architecture, the core specifications differ significantly. The RTX 3050 A Mobile has 1,792 shading units, nearly double the GTX 660’s 960. It also has more ROPs (32 vs 24) but fewer TMUs (56 vs 80). Clock speeds are higher on the newer part, with a base of 1065 MHz and boost of 1343 MHz, compared to the GTX 660’s 980 MHz base and 1032 MHz boost.

Memory is another major point of divergence. The RTX 3050 A Mobile comes with 4 GB of GDDR6 memory on a 128-bit bus, delivering 192.0 GB/s of bandwidth. The GTX 660 has 2 GB of GDDR5 on a wider 192-bit bus, but its bandwidth is lower at 144.2 GB/s. The memory clocks are also different: the RTX 3050 A Mobile runs at 1500 MHz (12 Gbps effective), while the GTX 660 runs at 1502 MHz (6 Gbps effective). The RTX 3050 A Mobile’s FP32 compute is rated at 4.813 TFLOPS, more than double the GTX 660’s 1.981 TFLOPS. It also supports FP16 at a 1:1 ratio, which the GTX 660 does not.

Power and physical characteristics are starkly different. The GTX 660 is a dual-slot desktop card with a 140 W TDP, requiring a 1x 6-pin power connector and a 300 W suggested PSU. It measures 241 mm in length. The RTX 3050 A Mobile is an IGP (integrated graphics processor) with a 45 W TDP and no power connectors, designed for portable devices. The GTX 660 uses a PCIe 3.0 x16 interface, while the RTX 3050 A Mobile uses a PCIe 4.0 x8 interface.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The NVIDIA GeForce GTX 660 has a higher average benchmark score of 9,022, compared to the NVIDIA GeForce RTX 3050 A Mobile’s 8,746.

Q: How much faster is the RTX 3050 A Mobile in the Geekbench OpenCL test?

A: The RTX 3050 A Mobile scores 52,998, while the GTX 660 scores 11,347, a deltaPct of -78.6% from the GTX 660’s perspective.

Q: Does the RTX 3050 A Mobile support ray tracing?

A: Yes, it has 14 RT cores and 56 tensor cores, enabling hardware-accelerated ray tracing and AI features. The GTX 660 has no such cores.

Q: What are the memory capacities of each card?

A: The GTX 660 has 2 GB of GDDR5 memory, while the RTX 3050 A Mobile has 4 GB of GDDR6 memory.

Q: Which card has a higher transistor density?

A: The RTX 3050 A Mobile has a significantly higher transistor density of 43.5M / mm², compared to the GTX 660’s 11.5M / mm².

Q: What is the TDP difference between the two?

A: The GTX 660 has a TDP of 140 W, while the RTX 3050 A Mobile has a much lower TDP of 45 W.

The Verdict

The data points to a clear, but nuanced, recommendation. For users seeking maximum raw compute performance, particularly in OpenCL workloads, the NVIDIA GeForce RTX 3050 A Mobile is the obvious choice. Its 52,998 OpenCL score versus the GTX 660’s 11,347 is a decisive victory, and its support for hardware ray tracing and tensor cores makes it future-proof for modern applications. Its 4 GB of GDDR6 memory also offers more headroom for larger textures and datasets.

However, the GTX 660 cannot be dismissed. Its higher average benchmark score of 9,022 vs 8,746 suggests it holds its own in a variety of other tests, and its lower power consumption (140 W vs 45 W is not lower, but its desktop form factor allows for more robust cooling) and mature drivers may make it a more stable choice for legacy software. It is also a desktop card with a 241 mm length, offering upgradeability and connectivity options like 2x DVI, HDMI 1.4a, and DisplayPort 1.2, which the mobile part lacks.

Ultimately, the choice depends on the user’s priorities. If the goal is to run modern, compute-intensive applications, the RTX 3050 A Mobile is superior. If the need is for a reliable, legacy-compatible desktop GPU with a solid aggregate performance profile, the GTX 660 remains a viable option. The RTX 3050 A Mobile is the forward-looking choice, while the GTX 660 confirms the enduring relevance of older architectures in specific niches.