AMD Radeon 550X vs NVIDIA GeForce GTX 1660 Ti Comparison

# AMD Radeon 550X vs NVIDIA GeForce GTX 1660 Ti

The AMD Radeon 550X and the NVIDIA GeForce GTX 1660 Ti occupy different tiers of the graphics card market, and the benchmark data confirms a clear performance hierarchy. The GTX 1660 Ti holds a decisive lead in raw compute across the board, while the 550X counters in a single workload. The recorded database results, drawn from Geekbench OpenCL and Vulkan tests, show a 53.9% win rate in favor of the NVIDIA card in one test, but a substantial 82.6% win for NVIDIA in the other. The data tells a story of a modern, high-throughput Turing design against a compact, power-efficient Polaris part.

Head-to-Head Benchmarks

Looking at the head-to-head benchmark results, the differences are stark. In the Geekbench OpenCL test, the AMD Radeon 550X scored 8,866 while the NVIDIA GeForce GTX 1660 Ti scored 51,029. That is an 82.6% advantage for the NVIDIA card, a decisive margin that reflects the generational gap in architecture and memory bandwidth. The AMD card cannot compete in raw compute throughput in this workload.

The Vulkan test tells a different story. Here the AMD Radeon 550X scored 8,970 against 5,827 for the GTX 1660 Ti. This is a 53.9% win for the AMD part. This result suggests that in Vulkan, the Radeon 550X is able to execute the workload more efficiently, likely due to driver scheduling and work-group distribution. The database shows the Radeon 550X winning one of the two head-to-head tests, with the GeForce GTX 1660 Ti winning the other.

The aggregate benchmark score for the Radeon 550X is 8,918, which places it in the 45th percentile of all GPUs in the database. The GeForce GTX 1660 Ti, by comparison, has an aggregate score of 7,723, sitting at the 41st percentile. The delta of 15.5% in the average score favors the AMD card. However, these aggregate numbers obscure the workload-specific results. The AMD card is a competent generalist, yet it is not a top-tier performer.

Architecture Differences

The architectural divide is fundamental to the performance gap. The AMD Radeon 550X is built on the Lexa chip, using the GCN 4.0 architecture, part of the Polaris generation. It is manufactured on a 14nm process at GlobalFoundries, with 2,200 million transistors on a 103 mm² die. That yields a transistor density of 21.4 million per square millimeter. The NVIDIA GeForce GTX 1660 Ti is a Turing design, chip name TU116, built by TSMC on a 12nm node. It packs 6,600 million transistors into a 284 mm² die, for a density of 23.2 million per square millimeter.

The Radeon 550X is configured with 2GB of GDDR5 memory on a 128-bit bus. This delivers a bandwidth of 112.0 GB/s. The GTX 1660 Ti uses 6GB of GDDR6 on a 192-bit bus, which brings bandwidth to 288.0 GB/s. A 2.57x bandwidth advantage is a fundamental factor in the NVIDIA card's favor.

The shading resources are also lopsided. The AMD chip has 512 shading units, 32 texture units, 16 ROPs, and a pixel rate of 19.49 GPixel/s. The NVIDIA chip has 1,536 shading units, 96 texture units, and 48 ROPs, with a pixel rate of 84.96 GPixel/s. The NVIDIA card's texture fill rate of 169.9 GTexel/s is far ahead. The AMD card's FP32 throughput is 1,247.2 GFLOPS; the NVIDIA card's is 5,437 GFLOPS. In FP16 (1:1) the AMD card delivers 1,247.2 GFLOPS, while the NVIDIA card delivers 10.87 TFLOPS, or 2:1 ratio.

The Verdict

Based on the recorded data, the choice is clear for workloads that stress raw compute, memory bandwidth, and newer API features. The GeForce GTX 1660 Ti is the superior card. It has 4.36x the FP32 throughput, 2.57x the memory bandwidth, a 4.36x higher pixel rate, and 4.59x the texture rate. The Radeon 550X wins the Vulkan test by a wide margin, which may benefit specific drivers or compute workloads that are Vulkan-heavy. But for overall compute, the GTX 1660 Ti is the stronger part. The Radeon's strength is its efficiency at 50W and 250W PSU while the NVIDIA card requires 120W and a 300W PSU. The AMD card is End-of-Life, while the NVIDIA card is End-of-Life as well.

Specification Differences

The two cards differ in several key specifications.

Memory:

The AMD Radeon 550X has 2 GB GDDR5 on a 128-bit bus, giving 112.0 GB/s bandwidth. The NVIDIA GeForce GTX 1660 Ti has 6 GB GDDR6 on a 192-bit bus, for 288.0 GB/s.

Processing and API:

AMD: 512 shading units, 32 TMUs, 16 ROPs, 1,247.2 GFLOPS FP32. APIs: DirectX 12 (12_0), OpenGL 4.6, Vulkan 1.3.

NVIDIA: 1,536 shading units, 96 TMUs, 48 ROPs, 5,437 GFLOPS FP32. APIs: DirectX 12 (12_1), OpenGL 4.6, Vulkan 1.4.

Dimensions and power:

AMD: 145 mm (5.7 inches) long, dual-slot, no power connector, 50 W TDP, 250 W PSU.

NVIDIA: 229 mm (9 inches) long, dual-slot, 1x 8-pin power connector, 120 W TDP, 300 W PSU.

FAQ

Q: Which card wins in Geekbench OpenCL?

A: The NVIDIA GeForce GTX 1660 Ti wins by 82.6% (11,029 vs 8,866). The AMD card is behind by a wide margin.

Q: Does the AMD card win in any benchmark?

A: Yes, in the Geekbench Vulkan test the AMD Radeon 550X wins by 53.9% (8,970 vs 5,827). This is the only head-to-head win for the AMD card.

Q: What is the memory capacity difference?

A: The AMD Radeon 550X has 2 GB GDDR5 with 112.0 GB/s bandwidth. The NVIDIA GeForce GTX 1660 Ti has 6 GB GDDR6 with 288.0 GB/s bandwidth.

Q: What is the architecture of each GPU?

A: The AMD Radeon 550X uses GCN 4.0 (Lexa chip) on a 14nm process. The NVIDIA GeForce GTX 1660 Ti uses the Turing architecture (TU116) on a 12nm process.

Q: Which card has higher FP32 performance?

A: The NVIDIA GeForce GTX 1660 Ti has 5,437 GFLOPS FP32, which is 4.36x higher than the AMD Radeon 550X's 1,247.2 GFLOPS.

Q: What are the production statuses?

A: Both cards are End-of-life. AMD released the Radeon 550X in March 2019, and NVIDIA released the GTX 1660 Ti in February 2019.