AMD Radeon RX 6600S vs NVIDIA GeForce GTX 1660 Comparison

The Verdict

The benchmark data splits these two end-of-life GPUs into distinct roles. The AMD Radeon RX 6600S is the faster raw performer in most modern workloads, winning 5 of the 8 head-to-head tests. Its PassMark G3D score of 12,649 beats the NVIDIA GeForce GTX 1660’s 11,646 by 7.9%, and its Geekbench OpenCL result of 66,435 is 28% higher than the GTX 1660’s 47,850. The RX 6600S also leads in every DirectX 10, 11, and 12 PassMark test, with margins ranging from 12.5% to 24.8%.

However, the GTX 1660 is not obsolete. It wins the legacy DirectX 9 test (177 vs. 176, a 0.6% edge), the 2D graphics test (776 vs. 647, a 19.9% lead), and the GPU compute test (4,963 vs. 4,879, a 1.7% advantage). Its average benchmark score of 11,680 is actually higher than the RX 6600S’s 10,629, despite the RX 6600S winning more individual tests. This discrepancy reflects the GTX 1660’s stronger performance in older API and compute workloads, while the RX 6600S dominates in modern DirectX and OpenCL scenarios.

For buyers prioritizing current-generation gaming APIs and raw compute throughput, the RX 6600S is the clear choice. For those running legacy DirectX 9 titles, 2D applications, or OpenCL-agnostic compute tasks, the GTX 1660 remains competitive. The GTX 1660’s percentile ranking of 51 versus the RX 6600S’s 49 indicates both sit near the middle of all GPUs, but the GTX 1660 has a slight edge in overall standing.

Where Each One Wins

The RX 6600S wins decisively in modern graphics APIs. Its PassMark DirectX 11 score of 105 is 24.8% higher than the GTX 1660’s 79, and its DirectX 10 score of 81 beats the GTX 1660’s 61 by 24.7%. The DirectX 12 gap narrows to 12.5% (56 vs. 49), but still favors AMD. The Geekbench OpenCL result of 66,435 represents the largest single victory, a 28% margin over the GTX 1660’s 47,850. This suggests the RX 6600S is substantially stronger in compute-oriented workloads that leverage modern instruction sets.

The GTX 1660’s wins are narrower but meaningful. Its PassMark G2D score of 776 outperforms the RX 6600S’s 647 by 19.9%, indicating better 2D rasterization and desktop composition performance. The GPU compute win is slim at 1.7% (4,963 vs. 4,879), but it shows the GTX 1660’s Turing architecture handles certain compute patterns more efficiently. The DirectX 9 result is essentially a tie at 177 vs. 176, a 0.6% margin that favors NVIDIA.

In terms of API coverage, the RX 6600S supports DirectX 12 Ultimate (12_2) whereas the GTX 1660 is limited to DirectX 12 (12_1). The RX 6600S also includes 28 ray tracing cores, which the GTX 1660 lacks entirely. This architectural difference explains why the RX 6600S pulls ahead in DirectX 12 and OpenCL tests, while the GTX 1660’s older-but-optimized design retains an edge in DirectX 9 and 2D workloads.

Architecture Differences

The two GPUs come from different design philosophies. The GTX 1660 uses NVIDIA’s Turing architecture on a 12 nm TSMC process, packing 6,600 million transistors into a 284 mm² die with a density of 23.2M per mm². The RX 6600S employs AMD’s RDNA 2.0 architecture on a 7 nm TSMC node, fitting 11,060 million transistors into a smaller 237 mm² die with a much higher density of 46.7M per mm². The 7 nm process gives AMD a clear transistor density advantage.

Clock speeds favor AMD as well. The RX 6600S has a base clock of 1700 MHz and a boost clock of 2000 MHz, with a game clock of 1881 MHz. The GTX 1660 runs at 1530 MHz base and 1785 MHz boost. Memory configurations differ: the GTX 1660 uses 6 GB of GDDR5 on a 192-bit bus, delivering 192.1 GB/s bandwidth, while the RX 6600S uses 4 GB of GDDR6 on a 128-bit bus but achieves higher bandwidth at 224.0 GB/s. The RX 6600S’s faster memory compensates for its narrower bus.

Compute resources favor AMD in raw counts. The RX 6600S has 1,792 shading units, 112 TMUs, and 64 ROPs, versus the GTX 1660’s 1,408 shading units, 88 TMUs, and 48 ROPs. This translates to higher theoretical throughput: the RX 6600S delivers 7.168 TFLOPS FP32 and 14.34 TFLOPS FP16 (2:1), while the GTX 1660 manages 5.027 TFLOPS FP32 and 10.05 TFLOPS FP16 (2:1). Pixel rate and texture rate also favor AMD: 128.0 GPixel/s and 224.0 GTexel/s versus 85.68 GPixel/s and 157.1 GTexel/s.

Power and form factor differ dramatically. The GTX 1660 is a dual-slot card with a 120 W TDP, requiring a 300 W PSU and a single 8-pin connector. It measures 229 mm in length, 111 mm in height, and 35 mm in width. The RX 6600S is an IGP (integrated graphics processor) with an 80 W TDP and no power connectors, designed for portable devices with device-dependent display outputs. The RX 6600S uses PCIe 4.0 x8, while the GTX 1660 uses PCIe 3.0 x16.

Feature support also separates them. The RX 6600S supports DirectX 12 Ultimate (12_2) and includes 28 ray tracing cores, enabling hardware-accelerated ray tracing. The GTX 1660 has no ray tracing or tensor cores, and its DirectX 12 (12_1) support is a generation behind. Both support OpenGL 4.6 and Vulkan 1.4. The GTX 1660 offers a DVI, HDMI 2.0, and DisplayPort 1.4a output set, whereas the RX 6600S’s outputs are portable-device dependent.

FAQ

Q: Which GPU has more raw compute power?

A: The RX 6600S delivers 7.168 TFLOPS FP32 versus the GTX 1660’s 5.027 TFLOPS, a 42.6% higher theoretical ceiling. This aligns with its 28% lead in Geekbench OpenCL.

Q: Why does the GTX 1660 have a higher average benchmark score despite losing more head-to-head tests?

A: The GTX 1660’s average score is 11,680 versus 10,629 for the RX 6600S. This reflects the GTX 1660’s large 19.9% win in PassMark G2D and its smaller wins in DirectX 9 and GPU compute, which offset the RX 6600S’s larger margins in other tests.

Q: Can the GTX 1660 do ray tracing?

A: No. The GTX 1660 has no ray tracing cores. The RX 6600S includes 28 ray tracing cores and supports DirectX 12 Ultimate (12_2), enabling hardware ray tracing.

Q: Which GPU is more power-efficient?

A: The RX 6600S has an 80 W TDP versus the GTX 1660’s 120 W TDP, making AMD’s part 33% more power-efficient on paper while delivering higher performance.

Q: How do memory bandwidth and capacity compare?

A: The GTX 1660 has 6 GB of GDDR5 on a 192-bit bus providing 192.1 GB/s. The RX 6600S has 4 GB of GDDR6 on a 128-bit bus providing 224.0 GB/s. AMD wins bandwidth; NVIDIA wins capacity.

Q: Which GPU is better for older DirectX 9 games?

A: The GTX 1660 barely edges out the RX 6600S in PassMark DirectX 9, scoring 177 vs. 176, a 0.6% difference. The margin is negligible, so either card handles legacy titles similarly.

Head-to-Head Benchmarks

The largest single victory belongs to the RX 6600S in Geekbench OpenCL, where its 66,435 score crushes the GTX 1660’s 47,850 by 28%. This is a compute-heavy workload that leverages the RX 6600S’s higher shading unit count (1,792 vs. 1,408) and faster FP32 throughput (7.168 TFLOPS vs. 5.027 TFLOPS). The margin is so large that it dominates the overall performance picture.

In PassMark DirectX 11, the RX 6600S scores 105 against the GTX 1660’s 79, a 24.8% lead. The DirectX 10 result is nearly identical in percentage terms: 81 vs. 61, a 24.7% margin. These are the two most significant modern-API wins for AMD. The DirectX 12 gap narrows to 12.5% (56 vs. 49), still favoring the RX 6600S but suggesting diminishing returns as the API matures.

The PassMark G3D test, which aggregates overall 3D performance, shows the RX 6600S ahead by 7.9% (12,649 vs. 11,646). This is the most representative single number for general gaming performance, and it confirms the RX 6600S as the faster card despite the GTX 1660’s higher average composite score.

The GTX 1660’s biggest win is in PassMark G2D, where its 776 score beats the RX 6600S’s 647 by 19.9%. This is a 2D rasterization test, and the GTX 1660’s 48 ROPs and Turing-era 2D pipeline clearly outperform the RX 6600S’s 64 ROPs in this specific workload. The GPU compute test is nearly a tie: the GTX 1660 scores 4,963 against the RX 6600S’s 4,879, a 1.7% edge. The DirectX 9 test is even closer at 177 vs. 176, a 0.6% margin.

Notably, the RX 6600S’s nearest rivals include the AMD Radeon RX 7800 XT (0.5% higher average score) and the NVIDIA Tesla K20c (1.8% higher), while the GTX 1660’s nearest rivals include the AMD Radeon RX 6500 XT (1.4% lower) and the Radeon Pro 5500M (1.3% higher). These proximity values show both cards sit in a tight performance cluster, with the GTX 1660’s average score of 11,680 slightly above the RX 6600S’s 10,629 despite the latter winning more individual tests.