NVIDIA GeForce GTX 1650 SUPER vs NVIDIA GeForce GTX 660 Comparison

The NVIDIA GeForce GTX 1650 SUPER and the NVIDIA GeForce GTX 660 represent two distinct eras of NVIDIA’s GPU design, separated by seven years of architectural evolution. The recorded data shows a decisive performance gap in favor of the newer card, but the comparison also highlights how far the GTX 660’s Kepler design has fallen relative to modern workloads. Below, the benchmark results, architectural differences, and specification changes are broken down using only the database measurements.

Head-to-Head Benchmarks

The database contains two direct head-to-head comparisons between these cards, and the GTX 1650 SUPER wins both by substantial margins. In Geekbench OpenCL, the GTX 1650 SUPER scores 43875 points against 11347 points for the GTX 660, a delta of 286.7%. That is nearly four times the raw compute output in a compute-oriented workload, reflecting the newer card’s higher shader throughput and memory bandwidth.

In Geekbench Vulkan, the gap widens further. The GTX 1650 SUPER records 50519 points, while the GTX 660 manages only 11415 points, a delta of 342.6%. Vulkan is a low-level API that benefits from modern hardware scheduling and asynchronous compute features, both of which favor the Turing architecture. The GTX 660, built for DirectX 11-era games, struggles to leverage the API’s parallelism, and the score difference reflects that.

Beyond the head-to-head tests, the average benchmark scores in the database reinforce the same conclusion. The GTX 1650 SUPER has an average benchmark score of 11047, placing it at the 50th percentile of all GPUs. The GTX 660 averages 9022 points and sits at the 45th percentile. The 2025-point gap between their averages is consistent with the direct test results, showing that the GTX 1650 SUPER is not just faster in isolated workloads but also more consistent across a broader range of benchmarks.

The GTX 1650 SUPER’s nearest rivals in the database include the AMD Radeon RX 550 with an average score of 11075 and a delta of -0.2%, meaning the GTX 1650 SUPER is essentially tied with that card. It also trails the NVIDIA RTX PRO 6000D Blackwell Max-Q and RTX PRO 6000 Blackwell Max-Q, both at 11088 with a -0.4% delta, by a negligible margin. Against the NVIDIA GeForce MX350, which averages 10883, the GTX 1650 SUPER leads by 1.5%. These tight deltas indicate that the GTX 1650 SUPER sits in a crowded performance band, but it is clearly above the GTX 660’s league.

For the GTX 660, its nearest rivals include the NVIDIA TITAN V CEO Edition at 9037 points (-0.2% delta) and the NVIDIA GeForce GTX 560 at 9058 points (-0.4% delta). The AMD Radeon 550X at 8918 points trails the GTX 660 by 1.2%, and the AMD Radeon Pro WX 5100 at 8863 points trails by 1.8%. The GTX 660’s average score places it among older or entry-level cards, confirming that its performance is now closer to low-end parts than to any modern midrange offering.

The wins tally in the head-to-head data is 2 for the GTX 1650 SUPER and 0 for the GTX 660. There is no benchmark in the database where the older card comes out ahead. The margin of victory is not close in either test, and the delta percentages show a performance advantage that grows with API modernity. In short, the data shows a one-sided contest.

Architecture Differences

The two cards are built on fundamentally different architectures. The GTX 1650 SUPER uses the TU116 chip, based on the Turing architecture, and is manufactured on a 12 nm process at TSMC. The GTX 660 uses the GK106 chip, based on the Kepler architecture, and is manufactured on a 28 nm process, also at TSMC. The process shrink from 28 nm to 12 nm is the largest single factor in the performance and efficiency differences between the two.

Transistor counts reflect the generational leap. The TU116 packs 6,600 million transistors on a 284 mm² die, yielding a transistor density of 23.2M per mm². The GK106 contains 2,540 million transistors on a 221 mm² die, with a density of 11.5M per mm². The newer chip has more than two and a half times the transistor count on a die that is only 63 mm² larger. That density increase enables the GTX 1650 SUPER to deliver far higher compute throughput while consuming less power.

The GTX 1650 SUPER has 1280 shading units, 80 texture mapping units, and 32 ROPs. The GTX 660 has 960 shading units, 80 texture mapping units, and 24 ROPs. The shading unit advantage of 320 units, combined with the higher clock speeds, explains the large FP32 performance gap. The GTX 1650 SUPER reaches 4.416 TFLOPS of FP32 compute, while the GTX 660 peaks at 1.981 TFLOPS. That is more than double the raw floating-point throughput. The GTX 1650 SUPER also supports FP16 at 8.832 TFLOPS with a 2:1 ratio, while the GTX 660 has no FP16 capability listed in the database.

Pixel and texture rates follow the same pattern. The GTX 1650 SUPER delivers 55.20 GPixel/s and 138.0 GTexel/s. The GTX 660 manages 20.64 GPixel/s and 82.56 GTexel/s. The pixel rate advantage is especially large, nearly threefold, which matters for high-resolution rendering and multi-sample anti-aliasing. The texture rate advantage is more moderate but still substantial at over 67%.

Memory architecture also differs significantly. The GTX 1650 SUPER uses 4 GB of GDDR6 memory on a 128-bit bus, with a bandwidth of 192.0 GB/s. The GTX 660 uses 2 GB of GDDR5 memory on a 192-bit bus, with a bandwidth of 144.2 GB/s. Although the GTX 660 has a wider bus, the GDDR6 memory on the newer card operates at a much higher effective speed, 12 Gbps versus 6 Gbps, resulting in higher total bandwidth. The doubling of memory capacity from 2 GB to 4 GB is also critical for modern games and applications that exceed 2 GB of video memory usage.

API support marks another clear separation. The GTX 1650 SUPER supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. The GTX 660 supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. The DirectX 12 feature level difference, 12_1 versus 11_0, means the older card lacks some modern rendering features such as conservative rasterization and certain shading models. The Vulkan version difference also indicates less mature driver and hardware support for the newer API on Kepler.

Display outputs differ as well. The GTX 1650 SUPER offers 1x DVI, 1x HDMI 2.0, and 1x DisplayPort 1.4a. The GTX 660 offers 2x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2. The newer card supports higher display bandwidth and newer connection standards, which matters for high refresh rate monitors and VR headsets. The GTX 660’s older outputs limit its usefulness with modern displays.

The Verdict

The data points to a clear choice for anyone comparing these two cards: the GTX 1650 SUPER is the superior product in every measurable way. It wins both head-to-head benchmarks, has a higher average benchmark score, sits at a higher percentile rank, and offers more memory, higher bandwidth, and newer API support. The GTX 660’s only advantages are its wider memory bus and lower launch MSRP, but the latter is a historical figure and the former does not overcome the bandwidth deficit.

For users considering a GTX 660 for any modern workload, the benchmark results indicate severe limitations. Its 2 GB of memory and 144.2 GB/s bandwidth will bottleneck in many current games, and its DirectX 12 (11_0) feature level means it cannot take full advantage of modern rendering APIs. The 45th percentile ranking places it among entry-level and legacy hardware.

The GTX 1650 SUPER, by contrast, sits at the 50th percentile and competes closely with cards like the AMD Radeon RX 550 and the NVIDIA MX350. It offers enough compute and memory for 1080p gaming at medium settings, and its Vulkan 1.4 support ensures better forward compatibility. The 100 W TDP and 300 W suggested PSU also make it an easier fit in existing systems than the 140 W GTX 660.

The GTX 660 should only be considered for legacy systems where the 2 GB memory and older API support are sufficient, or for collectors interested in Kepler-era hardware. For any practical purpose, the GTX 1650 SUPER is the recommended choice, strictly based on the recorded data.

Specification Differences

The two cards differ in nearly every core specification. The GTX 1650 SUPER uses the TU116 chip with 6,600 million transistors on a 284 mm² die, while the GTX 660 uses the GK106 chip with 2,540 million transistors on a 221 mm² die. The process node is 12 nm for the newer card and 28 nm for the older one, with transistor densities of 23.2M and 11.5M per mm², respectively.

Clock speeds favor the GTX 1650 SUPER. Its base clock is 1530 MHz and boost clock is 1725 MHz, versus 980 MHz base and 1032 MHz boost for the GTX 660. The memory clock is 1500 MHz with 12 Gbps effective speed on the newer card, compared to 1502 MHz with 6 Gbps effective speed on the older card.

Memory configuration differs in size, type, bus width, and bandwidth. The GTX 1650 SUPER has 4 GB GDDR6 on a 128-bit bus with 192.0 GB/s bandwidth. The GTX 660 has 2 GB GDDR5 on a 192-bit bus with 144.2 GB/s bandwidth.

Compute resources are higher on the newer card. The GTX 1650 SUPER has 1280 shading units, 80 TMUs, and 32 ROPs. The GTX 660 has 960 shading units, 80 TMUs, and 24 ROPs. Pixel rate, texture rate, and FP32 figures all favor the GTX 1650 SUPER.

Power and physical specs show some similarities. Both are dual-slot cards with a single 6-pin power connector and a 300 W suggested PSU. The TDP differs: 100 W for the GTX 1650 SUPER and 140 W for the GTX 660. The GTX 1650 SUPER is 229 mm long, 111 mm tall, and 35 mm wide. The GTX 660 is 241 mm long with no recorded height or width.

API support and display outputs are newer on the GTX 1650 SUPER. The newer card supports DirectX 12 (12_1), OpenGL 4.6, Vulkan 1.4, and outputs 1x DVI, 1x HDMI 2.0, and 1x DisplayPort 1.4a. The older card supports DirectX 12 (11_0), OpenGL 4.6, Vulkan 1.2.175, and outputs 2x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2.

Release dates and launch MSRPs are recorded for both. The GTX 1650 SUPER launched on 2019-11-21 with a launch MSRP of 159 USD. The GTX 660 launched on 2012-09-05 with a launch MSRP of 229 USD.

FAQ

Q: How much faster is the GTX 1650 SUPER than the GTX 660 in Vulkan?

A: In the Geekbench Vulkan test, the GTX 1650 SUPER scores 50519 points versus 11415 points for the GTX 660, a delta of 342.6%.

Q: Which card has more memory?

A: The GTX 1650 SUPER has 4 GB of GDDR6 memory, while the GTX 660 has 2 GB of GDDR5 memory. The newer card also has higher bandwidth at 192.0 GB/s versus 144.2 GB/s.

Q: Do both cards support the same DirectX version?

A: No. The GTX 1650 SUPER supports DirectX 12 (12_1), while the GTX 660 supports DirectX 12 (11_0). The feature level difference matters for modern rendering features.

Q: What is the average benchmark score for each card?

A: The GTX 1650 SUPER has an average benchmark score of 11047, placing it at the 50th percentile. The GTX 660 averages 9022 points, placing it at the 45th percentile.

Q: How does the GTX 1650 SUPER compare to its nearest rivals?

A: The GTX 1650 SUPER trails the AMD Radeon RX 550 by 0.2% (11075 versus 11047) and leads the NVIDIA GeForce MX350 by 1.5% (10883 versus 11047). It is also within 0.4% of the NVIDIA RTX PRO 6000D Blackwell Max-Q and RTX PRO 6000 Blackwell Max-Q, both at 11088.

Q: What is the TDP difference between the two cards?

A: The GTX 1650 SUPER has a TDP of 100 W, while the GTX 660 has a TDP of 140 W. Both cards use a single 6-pin power connector and a 300 W suggested PSU.