AMD Radeon Pro 555 vs NVIDIA GeForce GTX 1660 SUPER Comparison

The AMD Radeon Pro 555 and the NVIDIA GeForce GTX 1660 SUPER occupy entirely different tiers of the GPU landscape, and benchmark data confirms this is not a close contest. The GTX 1660 SUPER dominates in every shared metric, delivering roughly 4.5 times the compute performance in OpenCL and Vulkan workloads, positioning it as a discrete desktop solution, while the Radeon Pro 555 serves as a legacy integrated mobile part. The data shows a clear hierarchy based on architectural generation, memory configuration, and raw throughput.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The AMD Radeon Pro 555 has a higher average benchmark score of 13407 compared to the NVIDIA GeForce GTX 1660 SUPER’s 12986. However, this average is derived from different test sets; the GTX 1660 SUPER’s score is pulled down by legacy DirectX 9 and DirectX 10 tests where it scores 189 and 65 respectively, while the Radeon Pro 555’s average is based solely on modern Geekbench compute tests.

Q: How do the two cards compare in raw FP32 floating-point performance?

A: The NVIDIA GeForce GTX 1660 SUPER delivers 5.027 TFLOPS of FP32 performance, which is nearly four times the AMD Radeon Pro 555’s 1,305.6 GFLOPS. This massive gap in shading throughput directly explains the NVIDIA card’s superior performance in modern compute and graphics workloads.

Q: What are the memory specifications of each card, and why does it matter?

A: The Radeon Pro 555 features 2 GB of GDDR5 memory on a 128-bit bus with 81.60 GB/s bandwidth, while the GTX 1660 SUPER has 6 GB of GDDR6 on a 192-bit bus with 336.0 GB/s bandwidth. The NVIDIA card offers three times the capacity and over four times the bandwidth, which is critical for higher-resolution textures and data-intensive compute tasks.

Q: Which card has a higher transistor count and die size?

A: The GTX 1660 SUPER uses the TU116 chip with 6,600 million transistors on a 284 mm² die, while the Radeon Pro 555 uses the Polaris 21 chip with 3,000 million transistors on a 123 mm² die. Despite being on a larger 12 nm process versus the Radeon’s 14 nm node, the NVIDIA chip packs over twice the transistor count.

Q: Do these cards support modern graphics APIs differently?

A: Yes. The GTX 1660 SUPER supports DirectX 12_1 and Vulkan 1.4, while the Radeon Pro 555 supports DirectX 12_0 and Vulkan 1.3. Both support OpenGL 4.6, but the NVIDIA card’s higher DirectX feature level and newer Vulkan version indicate better support for recent rendering techniques.

Q: What is the form factor difference between the two?

A: The Radeon Pro 555 is an integrated graphics processor (IGP) with no power connectors and a portable-device-dependent display output, while the GTX 1660 SUPER is a dual-slot discrete card measuring 229 mm in length, requiring a single 8-pin power connector and a 300 W suggested power supply.

Where Each One Wins

The Radeon Pro 555 wins in exactly zero benchmark comparisons against the GTX 1660 SUPER in the provided head-to-head data. Its only statistical advantage is a marginally higher average benchmark score, which is misleading due to differing test suites. The Radeon’s strength lies in its power efficiency and integration: at 75 W TDP with no power connectors, it fits into portable devices where the 125 W, dual-slot GTX 1660 SUPER with an 8-pin connector cannot physically exist.

The GTX 1660 SUPER wins decisively in every compute benchmark. In Geekbench OpenCL, it scores 52490 versus 11682, a 77.7% advantage. In Geekbench Vulkan, it scores 57102 versus 12303, a 78.5% advantage. The NVIDIA card also wins on architectural features: it has nearly double the shading units (1408 vs 768), almost double the texture mapping units (88 vs 48), and triple the ROPs (48 vs 16). For any workload that leverages these resources—gaming, rendering, or GPGPU compute—the GTX 1660 SUPER is the only viable choice.

Architecture Differences

The two GPUs represent fundamentally different design philosophies separated by two generations of GPU architecture. The Radeon Pro 555 uses AMD’s GCN 4.0 architecture on the Polaris 21 chip, manufactured on a 14 nm process at GlobalFoundries. GCN 4.0 was designed for efficiency and balanced compute, but it lacks the specialized throughput enhancements of newer designs. The chip integrates 3,000 million transistors on a 123 mm² die, yielding a transistor density of 24.4M per mm².

The GTX 1660 SUPER uses NVIDIA’s Turing architecture on the TU116 chip, manufactured on a 12 nm process at TSMC. Turing introduced significant changes to the shader pipeline, including unified caching and improved integer/floating-point concurrency, even without the RT and tensor cores found in higher-end Turing parts. The TU116 packs 6,600 million transistors on a 284 mm² die, with a slightly lower transistor density of 23.2M per mm² due to the larger, more complex design. This architectural leap explains why the NVIDIA card achieves 5.027 TFLOPS FP32 from 1408 cores, while the AMD card manages only 1,305.6 GFLOPS from 768 cores—the per-core efficiency gain in Turing is substantial.

Memory architecture also differs fundamentally. The Radeon Pro 555 uses GDDR5 with a 128-bit interface and 5.1 Gbps effective speed, while the GTX 1660 SUPER uses GDDR6 with a 192-bit interface and 14 Gbps effective speed. The newer GDDR6 standard and wider bus give the NVIDIA card a 336.0 GB/s bandwidth ceiling versus 81.60 GB/s, a 4.1x advantage that becomes the bottleneck in many modern workloads.

Specification Differences

The most striking difference is in the compute core configuration. The GTX 1660 SUPER has 1408 shading units, 88 TMUs, and 48 ROPs, versus the Radeon Pro 555’s 768 shading units, 48 TMUs, and 16 ROPs. This translates to a pixel rate of 85.68 GPixel/s and texture rate of 157.1 GTexel/s for the NVIDIA card, compared to 13.60 GPixel/s and 40.80 GTexel/s for the AMD card.

Clock speeds vary significantly. The GTX 1660 SUPER has a base clock of 1530 MHz and boost clock of 1785 MHz, while the Radeon Pro 555 has no listed base or boost clocks, only a memory clock of 1275 MHz (5.1 Gbps effective). The NVIDIA card’s memory runs at 1750 MHz (14 Gbps effective). FP16 performance also differs: the GTX 1660 SUPER delivers 10.05 TFLOPS at a 2:1 ratio, while the Radeon Pro 555 offers 1,305.6 GFLOPS at a 1:1 ratio.

Physical and power specifications diverge completely. The Radeon Pro 555 is an IGP with 75 W TDP and no power connectors, while the GTX 1660 SUPER is a dual-slot card with 125 W TDP, a single 8-pin connector, and a 300 W suggested PSU. The NVIDIA card measures 229 mm by 111 mm by 35 mm, while the AMD card has no listed dimensions. The bus interface also differs: PCIe 3.0 x8 for the Radeon versus PCIe 3.0 x16 for the GTX 1660 SUPER. The NVIDIA card offers 1x DVI, 1x HDMI 2.0, and 1x DisplayPort 1.4a outputs, while the Radeon’s outputs are portable-device dependent.

Head-to-Head Benchmarks

The head-to-head data contains only two shared benchmarks, and the GTX 1660 SUPER wins both by overwhelming margins. In Geekbench OpenCL, the NVIDIA card scores 52490 against the Radeon’s 11682, a delta of -77.7% for the AMD card. This means the GTX 1660 SUPER is approximately 4.5 times faster in OpenCL compute. The gap is nearly identical in Geekbench Vulkan: 57102 versus 12303, a -78.5% delta, again roughly 4.6 times faster.

These results align with the theoretical specifications. The GTX 1660 SUPER’s 5.027 TFLOPS FP32 is 3.85 times the Radeon’s 1,305.6 GFLOPS, and the NVIDIA card’s 336.0 GB/s bandwidth is 4.12 times higher. The benchmark deltas of 77-78% slightly exceed what raw specs suggest, indicating that Turing’s architectural efficiency provides an additional advantage beyond just core count and clock speed. The GTX 1660 SUPER’s nearest rivals include the AMD Radeon RX 580 (deltaPct 0.4) and NVIDIA RTX 3050 Ti Mobile (deltaPct 0.4), placing it in a competitive mid-range tier. The Radeon Pro 555’s nearest rival is the NVIDIA P106-090 (deltaPct -0.5), a much lower-performance part.

Notably, the Radeon Pro 555’s average benchmark score of 13407 is higher than the GTX 1660 SUPER’s 12986, but this is an artifact of test selection. The NVIDIA card’s average includes legacy DirectX 9 and DirectX 10 tests (scores of 189 and 65) that drag down its average, while the AMD card’s average comes only from Geekbench compute tests. In any modern compute or graphics workload, the GTX 1660 SUPER is categorically superior.

The Verdict

The data is unambiguous: the NVIDIA GeForce GTX 1660 SUPER is the superior GPU by every meaningful performance metric. It wins both head-to-head benchmarks with deltas exceeding 77%, offers 4.1x the memory bandwidth, 3.85x the FP32 throughput, and supports a newer DirectX feature level and Vulkan version. Its nearest rivals include the AMD Radeon RX 580 and RTX 3050 Ti Mobile, placing it in a performance class that the Radeon Pro 555 cannot approach.

The AMD Radeon Pro 555’s only advantages are its 75 W TDP, integrated form factor, and lack of external power requirements. It is designed for portable devices where the GTX 1660 SUPER physically cannot fit. For any user building or buying a desktop system, the GTX 1660 SUPER is the only rational choice, offering a launch MSRP of 229 USD for roughly 450% more compute performance. For users constrained to a laptop or all-in-one with a soldered GPU, the Radeon Pro 555 is a legacy part (end-of-life, released in 2017) that should be avoided if any alternative exists, given its 2 GB memory limit and 81.60 GB/s bandwidth bottleneck. The verdict is simple: choose the GTX 1660 SUPER for performance, choose the Radeon Pro 555 only if the form factor forces it.