AMD Radeon 740M vs NVIDIA GeForce GTX 1660 SUPER Comparison

The NVIDIA GeForce GTX 1660 SUPER and AMD Radeon 740M represent two fundamentally different approaches to graphics processing: a dedicated, end-of-life desktop card from 2019 versus an active, power-sipping integrated GPU from 2024. While their average benchmark scores are nearly identical—the GTX 1660 SUPER sits at 12986 against the 740M's 12870—the data reveals that this parity is a statistical mirage, masking vast differences in architecture, performance characteristics, and use cases. The comparison hinges on understanding that the GTX 1660 SUPER delivers its performance through sheer dedicated hardware, while the Radeon 740M achieves comparable average scores through a completely different, system-dependent paradigm.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA GeForce GTX 1660 SUPER has a marginally higher average benchmark score of 12986 compared to the AMD Radeon 740M's 12870, a difference of roughly 0.9%. Both cards occupy the 53rd percentile among all GPUs.

Q: How do the two GPUs compare in the Geekbench OpenCL test?

A: The NVIDIA GeForce GTX 1660 SUPER delivers a score of 52490, which is 416% higher than the AMD Radeon 740M's 10172. This represents the largest performance gap in the head-to-head data.

Q: What is the most significant architectural difference between the two?

A: The GTX 1660 SUPER uses the Turing architecture on a 12 nm process with 6,600 million transistors, while the Radeon 740M uses RDNA 3.0 on a 4 nm process with 20,900 million transistors. The 740M also integrates ray tracing cores, which the GTX 1660 SUPER lacks entirely.

Q: Are these GPUs comparable in terms of power consumption?

A: No. The GTX 1660 SUPER has a 125 W TDP and requires a dual-slot cooler with a 1x 8-pin power connector, while the Radeon 740M has a 45 W TDP, is an IGP (integrated graphics processor), and requires no power connectors.

Q: What does the Radeon 740M's "System Shared" memory mean for performance?

A: The Radeon 740M uses System Shared memory, meaning its memory size, type, bus width, and bandwidth are all system-dependent. In contrast, the GTX 1660 SUPER has dedicated 6 GB of GDDR6 memory on a 192-bit bus with 336.0 GB/s of bandwidth.

Q: Which GPU has a higher boost clock speed?

A: The AMD Radeon 740M has a significantly higher boost clock of 2800 MHz compared to the GTX 1660 SUPER's 1785 MHz, despite the NVIDIA card having a higher base clock of 1530 MHz versus 800 MHz.

Architecture Differences

The architectural chasm between these two GPUs is stark, starting with their fundamental designs. The GTX 1660 SUPER is built on the Turing architecture using the TU116 chip, manufactured on TSMC's 12 nm process. This dedicated card packs 6,600 million transistors into a 284 mm² die, yielding a transistor density of 23.2 million per mm². In contrast, the Radeon 740M uses the Phoenix2 chip with RDNA 3.0 architecture, fabricated on a much more advanced 4 nm process. It integrates 20,900 million transistors into a smaller 137 mm² die, achieving an impressive 152.6 million transistors per mm².

The compute resources diverge dramatically. The GTX 1660 SUPER features 1408 shading units, 88 texture mapping units, and 48 raster output pipelines. The Radeon 740M, despite its newer architecture, has only 256 shading units, 16 TMUs, and 8 ROPs. However, the 740M includes 4 ray tracing cores, a feature completely absent from the GTX 1660 SUPER. This difference in raw hardware counts explains the massive gap in pixel and texture rates: the NVIDIA card produces 85.68 GPixel/s and 157.1 GTexel/s, while the AMD card manages only 22.40 GPixel/s and 44.80 GTexel/s.

The FP32 compute performance also favors the dedicated card, with the GTX 1660 SUPER delivering 5.027 TFLOPS versus the 740M's 2.867 TFLOPS. Interestingly, the FP16 performance tells a different story: the GTX 1660 SUPER achieves 10.05 TFLOPS using a 2:1 ratio, while the Radeon 740M delivers 2.867 TFLOPS at a 1:1 ratio, meaning the NVIDIA card's FP16 advantage is even larger than its FP32 lead. The memory subsystems are incomparable—the GTX 1660 SUPER has dedicated 6 GB GDDR6 on a 192-bit bus with 336.0 GB/s bandwidth, while the 740M relies entirely on system-shared memory with system-dependent bandwidth.

Head-to-Head Benchmarks

The head-to-head benchmark data reveals a decisive victory for the NVIDIA GeForce GTX 1660 SUPER, winning both available tests. In Geekbench OpenCL, the GTX 1660 SUPER scores 52490 against the Radeon 740M's 10172, representing a 416% advantage. This is not a marginal lead but a generational gap in raw compute throughput, reflecting the NVIDIA card's dedicated memory bandwidth and vastly higher shading unit count.

The Vulkan results paint a similar picture, though with a slightly smaller margin. The GTX 1660 SUPER scores 57102 in Geekbench Vulkan, while the Radeon 740M manages 15568, giving the NVIDIA card a 266.8% lead. This suggests that even in a modern, low-level graphics API, the dedicated card's hardware advantages—more TMUs, ROPs, and dedicated GDDR6 memory—overwhelm the integrated GPU's architectural efficiency and higher clock speeds.

The average benchmark scores, however, tell a more nuanced story. The GTX 1660 SUPER's average of 12986 is only 0.9% higher than the 740M's 12870, placing them in the same percentile (53rd) of all GPUs. This discrepancy between the head-to-head tests and the average scores arises because the average includes a broad range of benchmarks—like PassMark DirectX 9, 10, 11, and 12 tests, G2D, G3D, and compute—where the Radeon 740M's data is not available. The head-to-head tests, which are the only direct comparisons, show the GTX 1660 SUPER as overwhelmingly superior.

Specification Differences

The specification table highlights where these two GPUs diverge most sharply. The most obvious differences are in process node (12 nm vs 4 nm), transistor count (6,600 million vs 20,900 million), and die size (284 mm² vs 137 mm²). The GTX 1660 SUPER has a much higher base clock (1530 MHz vs 800 MHz), but the Radeon 740M counters with a much higher boost clock (2800 MHz vs 1785 MHz).

Memory is a categorical difference: the NVIDIA card has 6 GB of GDDR6 on a 192-bit bus with 336.0 GB/s bandwidth, while the AMD card uses System Shared memory with system-dependent bandwidth. The shading units (1408 vs 256), TMUs (88 vs 16), and ROPs (48 vs 8) all massively favor the dedicated card. The GTX 1660 SUPER also has higher pixel rate (85.68 GPixel/s vs 22.40 GPixel/s), texture rate (157.1 GTexel/s vs 44.80 GTexel/s), and FP32 compute (5.027 TFLOPS vs 2.867 TFLOPS).

Power and physical specifications are equally divergent. The GTX 1660 SUPER has a 125 W TDP, requires a dual-slot design with a 1x 8-pin power connector, and a 300 W suggested PSU. The Radeon 740M has a 45 W TDP, is an IGP with no power connectors, and no suggested PSU. The NVIDIA card is 229 mm long, 111 mm tall, and 35 mm wide, while the AMD IGP has no listed dimensions. The GTX 1660 SUPER supports PCIe 3.0 x16, while the 740M uses PCIe 4.0 x8. Finally, the GTX 1660 SUPER supports DirectX 12 (12_1), while the Radeon 740M supports DirectX 12 Ultimate (12_2), reflecting its newer architecture.

The Verdict

The data is unambiguous for direct performance: the NVIDIA GeForce GTX 1660 SUPER is the superior GPU in every head-to-head benchmark. Its 416% lead in OpenCL and 266.8% lead in Vulkan are not statistical noise but reflect fundamental hardware advantages—more shading units, dedicated memory, and higher texture and pixel rates. For any workload that these benchmarks represent, the GTX 1660 SUPER is the clear choice.

However, the Radeon 740M's value proposition lies elsewhere. Its 45 W TDP, IGP form factor, and lack of power connectors make it suitable for systems where a dedicated card is impossible or undesirable. The fact that its average benchmark score is within 0.9% of the GTX 1660 SUPER, despite having a fraction of the compute resources, suggests that its 4 nm process and RDNA 3.0 architecture are highly efficient. The 740M also supports DirectX 12 Ultimate and includes ray tracing cores, features the GTX 1660 SUPER lacks.

The GTX 1660 SUPER, as an end-of-life product, offers raw performance at the cost of power and space. The Radeon 740M, as an active product, offers efficiency and modern features at the cost of raw throughput. The choice depends entirely on the use case: the GTX 1660 SUPER for dedicated graphics workloads, the Radeon 740M for integrated, power-constrained systems.

Where Each One Wins

The NVIDIA GeForce GTX 1660 SUPER wins decisively in raw compute benchmarks. In Geekbench OpenCL, it achieves 52490 points, which is 416% higher than the Radeon 740M's 10172. This indicates a massive advantage in general-purpose GPU compute tasks, likely benefiting applications that leverage OpenCL for acceleration. Similarly, in Geekbench Vulkan, the GTX 1660 SUPER scores 57102 versus 15568, a 266.8% lead, suggesting it is substantially better for modern graphics APIs and gaming workloads that use Vulkan.

The Radeon 740M's wins are not in benchmark scores but in system integration and efficiency. Its 45 W TDP versus the GTX 1660 SUPER's 125 W means it can operate in systems without dedicated power connectors or additional cooling. Its IGP form factor eliminates the need for a dual-slot card, making it suitable for compact or laptop designs. The 740M's support for DirectX 12 Ultimate and its 4 ray tracing cores provide modern feature support that the GTX 1660 SUPER cannot offer, despite the NVIDIA card's higher raw performance.

For users with a desktop system that can accommodate a 229 mm dual-slot card with an 8-pin power connector, the GTX 1660 SUPER is the clear winner for any performance-sensitive task. For users building a power-efficient, compact system where integrated graphics are the only option, the Radeon 740M offers respectable average performance (12870) that rivals the GTX 1660 SUPER's average (12986) in overall percentile ranking, albeit through a completely different set of benchmarks. The data suggests that the 740M's performance is highly dependent on the system's memory configuration, given its system-shared memory design, making it a variable but potentially adequate solution for light gaming and general use.