AMD Radeon Pro 555 vs NVIDIA GeForce GTX 1660 Comparison

Head-to-Head Benchmarks

The recorded benchmark data shows a decisive victory for the NVIDIA GeForce GTX 1660 in every shared test. In Geekbench OpenCL, the GTX 1660 scores 47,850 against the AMD Radeon Pro 555's 11,682, a delta of 75.6% in favor of the NVIDIA card. The Vulkan results tell a similar story: the GTX 1660 posts 50,137 while the Radeon Pro 555 manages 12,303, putting the NVIDIA card 75.5% ahead. These are not marginal gains; they represent a generational leap in compute throughput, with the GTX 1660 delivering roughly four times the raw performance in both APIs.

The AMD Radeon Pro 555's average benchmark score across its tested workloads is 13,407, placing it at the 54th percentile of all GPUs in the database. Its nearest rivals include the AMD Radeon HD 8950M (13,376, 0.2% slower), the AMD Radeon RX 5500M (13,356, 0.4% slower), and the AMD FirePro M6100 (13,354, 0.4% slower). The NVIDIA P106-090 sits slightly ahead at 13,470, a 0.5% margin over the Radeon Pro 555. These figures indicate that the Radeon Pro 555 is competitive within its own performance tier, but that tier is far below what the GTX 1660 occupies.

The NVIDIA GeForce GTX 1660, by contrast, averages 11,680 across its benchmark suite, which includes DirectX 9 through 12 tests, compute workloads, and 3DMark Steel Nomad. That average places it at the 51st percentile, with rivals such as the AMD Radeon RX 7800 XT (11,627, 0.5% slower), the AMD Radeon Pro 5500M (11,528, 1.3% slower), and the AMD Radeon RX 6500 XT (11,842, 1.4% faster). The NVIDIA Tesla K20c trails by 1.8% with a score of 11,479. While the average scores look similar on paper, the GTX 1660's individual test results, particularly its OpenCL and Vulkan scores, are dramatically higher than anything the Radeon Pro 555 can produce.

The head-to-head comparison includes only two tests, both of which the GTX 1660 wins outright. The Radeon Pro 555 has zero wins in this matchup. The delta percentages of 75.6% and 75.5% are among the largest margins recorded for any GPU pairing in this database segment, underscoring the architectural and generational gap between the two products.

The Verdict

The data points to a clear conclusion: the NVIDIA GeForce GTX 1660 is the superior performer in every measurable benchmark shared between the two cards. Its OpenCL score of 47,850 is more than four times the Radeon Pro 555's 11,682, and its Vulkan result of 50,137 dwarfs the AMD card's 12,303. For any workload that leverages compute APIs, the GTX 1660 is the obvious choice.

The AMD Radeon Pro 555, however, has its own niche. It is an integrated graphics processor (IGP) with a 75 W thermal design power, no power connectors, and a slot width listed as IGP, meaning it is designed for portable, power-constrained systems. Its 2 GB of GDDR5 memory on a 128-bit bus delivers 81.60 GB/s of bandwidth, which is modest but sufficient for light duties. The GTX 1660, in contrast, is a dual-slot, 120 W card requiring a 1x 8-pin power connector and a 300 W power supply. Its 6 GB of GDDR5 memory on a 192-bit bus provides 192.1 GB/s of bandwidth, more than double the AMD card's throughput.

If the use case is a compact, low-power laptop or portable workstation where the GPU must share space with other components and draw minimal power, the Radeon Pro 555 is the only viable option among these two. Its end-of-life status and 2017 release date do not diminish its usefulness in that specific context. But for any desktop system with a power supply capable of 300 W, the GTX 1660 is the clear winner: it offers four times the compute performance, triple the memory capacity, and nearly 2.4 times the memory bandwidth. The GTX 1660's launch MSRP is 219 USD, which the database records as its only pricing reference.

The verdict, strictly from the data: pick the GTX 1660 unless the physical constraints of the chassis and power budget make the Radeon Pro 555 the only feasible installation. The GTX 1660 outperforms the Radeon Pro 555 in every shared benchmark by a margin exceeding 75%, and that gap is not recoverable by any driver or software optimization.

FAQ

Q: Which GPU has the higher OpenCL score?

A: The NVIDIA GeForce GTX 1660 scores 47,850 in Geekbench OpenCL, while the AMD Radeon Pro 555 scores 11,682, a 75.6% difference in favor of the NVIDIA card.

Q: Does the AMD Radeon Pro 555 win any benchmark against the GTX 1660?

A: No. The head-to-head database records two tests (Geekbench OpenCL and Geekbench Vulkan), and the GTX 1660 wins both. The Radeon Pro 555 has zero wins in this comparison.

Q: What is the memory bandwidth difference between the two cards?

A: The GTX 1660 has 192.1 GB/s of bandwidth from 6 GB of GDDR5 on a 192-bit bus. The Radeon Pro 555 has 81.60 GB/s from 2 GB of GDDR5 on a 128-bit bus. The GTX 1660 provides more than double the bandwidth.

Q: Which card has a higher transistor density?

A: The Radeon Pro 555, built on a 14 nm process, packs 3,000 million transistors into a 123 mm² die, yielding 24.4M transistors per mm². The GTX 1660, on a 12 nm process, has 6,600 million transistors on a 284 mm² die, yielding 23.2M per mm². The AMD chip is denser despite the older process node.

Q: What are the power requirements for each card?

A: The Radeon Pro 555 has a 75 W TDP and requires no power connectors, making it suitable for integrated systems. The GTX 1660 has a 120 W TDP, needs a 1x 8-pin power connector, and the database suggests a 300 W power supply.

Q: How do the two cards compare in terms of pixel and texture rates?

A: The GTX 1660 achieves 85.68 GPixel/s and 157.1 GTexel/s, while the Radeon Pro 555 reaches 13.60 GPixel/s and 40.80 GTexel/s. The NVIDIA card is over six times faster in pixel throughput and nearly four times faster in texture throughput.

Specification Differences

The two cards differ in nearly every measurable specification. The AMD Radeon Pro 555 uses a Polaris 21 chip on a 14 nm GlobalFoundries process, while the NVIDIA GeForce GTX 1660 uses a TU116 chip on a 12 nm TSMC process. Transistor counts are 3,000 million versus 6,600 million, and die sizes are 123 mm² versus 284 mm². The Radeon Pro 555 has a higher transistor density at 24.4M per mm² compared to the GTX 1660's 23.2M per mm².

Clock speeds: the Radeon Pro 555 lists no base or boost clock in the database, only a memory clock of 1275 MHz (5.1 Gbps effective). The GTX 1660 has a base clock of 1530 MHz, a boost clock of 1785 MHz, and a memory clock of 2001 MHz (8 Gbps effective). Memory capacity is 2 GB versus 6 GB of GDDR5, with bus widths of 128 bit versus 192 bit, and bandwidth of 81.60 GB/s versus 192.1 GB/s.

Compute resources differ substantially: the Radeon Pro 555 has 768 shading units, 48 TMUs, and 16 ROPs, while the GTX 1660 has 1,408 shading units, 88 TMUs, and 48 ROPs. Pixel rates are 13.60 GPixel/s versus 85.68 GPixel/s, and texture rates are 40.80 GTexel/s versus 157.1 GTexel/s. FP32 performance is 1,305.6 GFLOPS versus 5.027 TFLOPS. FP16 is 1,305.6 GFLOPS (1:1 ratio) for AMD and 10.05 TFLOPS (2:1 ratio) for NVIDIA.

Power and physical specifications: the Radeon Pro 555 is an IGP with 75 W TDP and no power connectors; the GTX 1660 is dual-slot, 120 W, with a 1x 8-pin connector and a suggested 300 W PSU. The GTX 1660 measures 229 mm by 111 mm by 35 mm. Bus interfaces are PCIe 3.0 x8 for AMD versus PCIe 3.0 x16 for NVIDIA. Display outputs: the Radeon Pro 555 is portable-device dependent, while the GTX 1660 offers 1x DVI, 1x HDMI 2.0, and 1x DisplayPort 1.4a.

API support: both support DirectX 12 and OpenGL 4.6, but AMD supports Vulkan 1.3 and DirectX 12_0, while NVIDIA supports Vulkan 1.4 and DirectX 12_1. Release dates are June 2017 for the Radeon Pro 555 and March 2019 for the GTX 1660, both now end-of-life.

Architecture Differences

The architectural divide is fundamental. The AMD Radeon Pro 555 uses GCN 4.0 architecture on a Polaris 21 chip, fabricated by GlobalFoundries at 14 nm. This is a mature design from the Radeon Pro Mac 500 Series generation, optimized for compatibility and power efficiency rather than raw throughput. Its FP16 performance runs at a 1:1 ratio with FP32, meaning no dedicated half-precision acceleration. The GPU relies on 768 shading units arranged in a configuration that delivers 1,305.6 GFLOPS of FP32 compute.

The NVIDIA GeForce GTX 1660 uses Turing architecture on the TU116 chip, built by TSMC at 12 nm. Turing is a newer design that introduces substantial changes in compute efficiency. The GTX 1660 has 1,408 shading units, nearly double the AMD card's count, and its FP32 throughput of 5.027 TFLOPS is almost four times higher. Notably, the GTX 1660's FP16 performance is 10.05 TFLOPS at a 2:1 ratio, indicating the presence of accelerated half-precision paths that the Radeon Pro 555 lacks entirely. This makes the NVIDIA card significantly more capable for workloads that can leverage FP16 arithmetic, such as certain machine learning inference tasks.

Cache and memory architecture also differ. The Radeon Pro 555 uses a 128-bit memory interface with 2 GB GDDR5, while the GTX 1660 uses a 192-bit interface with 6 GB GDDR5. The larger bus width and higher memory clock yield 192.1 GB/s versus 81.60 GB/s. The GTX 1660's 48 ROPs compared to the Radeon Pro 555's 16 ROPs directly impacts pixel fill rate, which the data shows at 85.68 GPixel/s versus 13.60 GPixel/s. TMU counts of 88 versus 48 similarly affect texture filtering throughput, measured at 157.1 GTexel/s versus 40.80 GTexel/s.

The Radeon Pro 555's IGP form factor and lack of power connectors reflect its design as an integrated solution for portable Mac systems, with no discrete display outputs of its own. The GTX 1660, in contrast, is a full add-in board with dual-slot cooling, a single 8-pin power connector, and multiple display outputs. The transistor density figures are close (24.4M versus 23.2M per mm²), but the absolute transistor count of 6,600 million versus 3,000 million illustrates the GTX 1660's larger, more complex die. Both cards are end-of-life, but the GTX 1660's later release date and Turing architecture give it a clear structural advantage in every computational category measured.